Antenna engineering handbook

Source: ANTENNA ENGINEERING HANDBOOK P ● A ● R ● T ● 1 Introduction and Fundamentals Downloaded from Digital...

Author: John Volakis

2160 downloads 7339 Views 62MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Source: ANTENNA ENGINEERING HANDBOOK

P

●

A

●

R

●

T

●

1

Introduction and Fundamentals

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.

Introduction and Fundamentals



Chapter 1

Fundamentals of Antennas, Arrays, and Mobile Communications Thomas F. Eibert Universität Stuttgart

John L. Volakis The Ohio State University CONTENTS 1.1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-4 1.2 HUYGENS’ AND EQUIVALENCE PRINCIPLES . . . . . . . . . . . . . . . . . . 1-5 1.3 HERTZIAN AND FITZGERALD ELEMENTARY RADIATORS . . . . . . . 1-7 1.4 FAR-FIELD ANTENNA PROPERTIES, POWER TRANSFER, AND RECIPROCITY. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-8 1.5 ANTENNAS AS ELECTROMAGNETIC CIRCUITS. . . . . . . . . . . . . . . 1-11 1.6 POLARIZATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-14 1.7 DIRECTIVITY PATTERNS FROM CONTINUOUS LINE SOURCES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-17 1.8 DIRECTIVITY PATTERNS FROM AREA SOURCE DISTRIBUTIONS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-21 1.9 FUNDAMENTALS OF ANTENNA ARRAYS . . . . . . . . . . . . . . . . . . . . 1-27 1.10 BASIC CONCEPTS IN MOBILE COMMUNICATIONS . . . . . . . . . . . 1-32 1-3 Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.

Fundamentals of Antennas, Arrays, and Mobile Communications 1-4

CHAPTER ONE

1.1 INTRODUCTION * Antennas are key components of any wireless communication system.1,2 They are the devices that allow for the transfer of a signal (in a wired system) to waves that, in turn, propagate through space and can be received by another antenna. The receiving antenna is responsible for the reciprocal process, i.e., that of turning an electromagnetic wave into a signal or voltage at its terminals that can subsequently be processed by the receiver. The receiving and transmitting functionalities of the antenna structure itself are fully characterized by Maxwell’s equations and are fairly well understood.3 The dipole antenna (a straight wire, fed at the center by a two-wire transmission line) was the first antenna ever used and is also one of the best understood.1,2 For effective reception and transmission, it must be approximately l/2 long (l = wavelength) at the frequency of operation (or multiples of this length). Thus, it must be fairly long (or high) when used at low frequencies (l = 1 m at 300 MHz), and even at higher frequencies (UHF and greater), its protruding nature makes it quite undesirable. Further, its low gain (2.15 dB), lack of directionality, and extremely narrow bandwidth make it even less attractive. Not surprisingly, the Yagi-Uda antenna (typically seen on the roof of most houses for television reception) was considered a breakthrough in antenna technology when introduced in the early 1920s because of its much higher gain of 8–14 dB. Log-periodic wire antennas introduced in the late 1950s and 1960s and wire spirals allowed for both gain and bandwidth increases. On the other hand, even today high gain antennas rely on large reflectors (dish antennas) and waveguide arrays (used for many radar systems). Until the late 1970s, antenna design was based primarily on practical approaches using off-the-shelf antennas such as various wire geometries (dipoles, Yagi-Uda, log-periodics, spirals), horns, reflectors, and slots/apertures as well as arrays of some of these. The antenna engineer could choose or modify one of them based on design requirements that characterize antennas, such as gain, input impedance, bandwidth, pattern beamwidth, and sidelobe levels (see References 4 and 5 for a description of these quantities). Antenna development required extensive testing and experimentation and was, therefore, funded primarily by governments. However, in recent years, dramatic growth in computing speed and development of effective computational techniques for realistic antenna geometries has allowed for low-cost virtual antenna design. Undoubtedly, the explosive growth of wireless communications and microwave sensors, microwave imaging needs, and radars has been the catalyst for introducing a multitude of new antenna designs over the past decade and an insatiable desire for using modern computational techniques for low-cost designs. Requirements for ∗

Heinrich R. Hertz was the first to demonstrate the generation of radio waves at UHF using a gap dipole in 1885– 1886 at Karlsruhe University (Germany). Hertz was able to detect radio waves 20 m away using a high-voltage electrical spark discharge to excite the dipole gap. From recorded conversations, Hertz did not seem to understand the impact of his experiments, but instead did it as a validation of the newly developed Maxwell’s equations. Within ten years, Tesla at the Franklin Institute in the U.S., Marconi in Bologna, Italy, Popov in Russia, and Bose in India, demonstrated wireless telegraphy. In 1892, Tesla delivered a widely distributed presentation at the IRE of London about “transmitting intelligence without wires,” and in 1895, he transmitted signals detected 50 miles (80 km) away. Concurrently, in 1894 Bose used wireless signals to ring a bell in Calcutta, and Popov presented his radio receiver to the Russian Physical & Chemical Society on May 7, 1895. Marconi is certainly considered the key individual for his contributions to the commercialization of radio waves, and he received the Nobel prize for his work in 1909. Nevertheless, Marconi’s widely advertised first radio wave transmission experiment was in 1895, and his British patent application in 1897 was preceded by that of Tesla. A culmination of Marconi’s experiments was the December 12, 1901, trans-Atlantic radio wave transmission of the Morse code for the letter S. The success of this experiment is often disputed, possibly due to strong atmospheric noise during the time of the experiment, but by the 1920s the U.S. had hundreds of radio stations, and in 1922, the BBC began transmitting in England. Subsequent development of radio detectors, vacuum tubes, and the tiny transistor in 1947 played a critical role in the practical everyday use of radio waves for communication and wireless transmission of information.


Fundamentals of Antennas, Arrays, and Mobile Communications FUNDAMENTALS OF ANTENNAS, ARRAYS, AND MOBILE COMMUNICATIONS

1-5

conformal antennas (non-protruding) for airborne systems, increased bandwidth requirements, and multifunctionality have led to heavy exploitation of printed (patch) or other slot-type antennas4 and the use of powerful computational tools (commercial and noncommercial) for designing such antennas. Needless to say, the commercial mobile communications industry has been the catalyst for the recent explosive growth in antenna design needs. Certainly, the past decade has seen an extensive use of antennas by the public for cellular, GPS, satellite, wireless LAN for computers (WiFi), Bluetooth technology, Radio Frequency ID (RFID) devices, WiMAX, and so on. However, future needs will be even greater when a multitude of antennas are integrated into say automobiles for all sorts of communication needs and into a variety of portable devices and sensors for monitoring and information gathering. Certainly, future RFID devices will most likely replace the bar codes on all products while concurrently allowing for instantaneous inventorying. For military applications, there is an increasing need for small and conformal multifunctional antennas that can satisfy a plethora of communications needs using as little space as possible. In this first chapter of the handbook, we provide a summary of antenna fundamentals and introduce antenna parameters typically used for characterizing antenna properties often employed to evaluate the entire radio system. We start with the radiation of an ideal (Hertzian) or infinitesimal dipole and proceed to the resonant l/2 dipole, antenna arrays, and mobile communication concepts.

1.2 HUYGENS’ AND EQUIVALENCE PRINCIPLES The electromagnetic behavior and thus the functioning of antennas is governed by Maxwell’s equations,3 which must be solved for a particular antenna and a given excitation. Typically, exact solutions of Maxwell’s equations are not available and thus numerical modeling is often used to compute approximate solutions for practical configurations. A formal simplification of electromagnetic antenna problems can be achieved by employing the equivalence principle.3 If interest is restricted to the field solution in a limited region of space, the antenna configuration can be replaced by the equivalent electromagnetic sources located on the surface of a volume enclosing the antenna configuration (see Figure 1-1). Because the antenna materials are no longer there, these sources are usually radiating in a homogeneous solution space (such as free-space), and the corresponding fields can thus be calculated by evaluating the radiation integrals. The equivalent sources are not uniquely defined, and there are many different ways of constructing them. In general, the equivalent sources are a composition of electric and magnetic surface current densities representing the excitation terms in Maxwell’s equations. A straightforward way of constructing equivalent sources is provided by Huygens’ principle.3 Huygens’ principle states that the field solution in a region V is completely determined by the tangential fields over the surface S enclosing V. The corresponding electric and magnetic equivalent surface current densities are given by Electric current density: J = nˆ × H

(1-1)

M = − nˆ × E

(1-2)

Magnetic current density:



CHAPTER ONE

S

n

R= r−r M J

r

r

z y

x FIGURE 1-1 Replacement of an antenna by equivalent electric and magnetic surface current densities

where both J and H are expressed in amperes per meter (A/m) and M and E are expressed in volts per meter (V/m). For the problem of a radiating antenna, as illustrated in Figure 1-1, the outer boundary of V is assumed to be located at infinity, where the fields radiated by the corresponding equivalent sources can be neglected. As shown in the figure, the antenna can be replaced by equivalent sources on an arbitrary surface S enclosing it. As already mentioned, these equivalent sources reproduce the radiated fields of the antenna, and they can be assumed as radiating in homogeneous space. For a particular antenna configuration, the exact determination of J and M requires knowledge of the true field distribution on S. However, for many practical antennas, an approximate determination of J and M is possible. For instance, placing S to coincide with a metallic section of the antenna structure causes M to vanish on these portions of S. The radiated fields from any antenna can be obtained by integrating the field contributions of the equivalent electric and magnetic current densities using the well-known radiation integral:3  e − jk0 |r-r '| 1 e − jk0 |r-r '|  E = − jωµ0  ∫∫ J(r ') 4π | r - r ' | + k02 (∇ ' ⋅ J(r '))∇ 4π | r - r ' | ds ' S + ∫∫ M(r ') × ∇

e − jk0 |r-r '| ds ' 4π | r - r ' |

which for the far-field ( r → ∞) reduces to (see Figure 1-1) E = − jωµ0 where

e − jk0r 4π r



 ∫∫ ( I − rrˆˆ) ⋅ J(r ') − 

 ε0 rˆ × M( r ') e jkrˆ⋅r ' ds ' µ0 

I = unit dyad r = defines location of observation point (see Figure 1-1) r = distance (in m) to observation point r' = defines location of the integrated surface current densities rˆ = unit vector in radial direction e0 = free-space permittivity m0 = free-space permeability Z0 =

µ0 = free-space impedance ε0



1-7

k0 Z 0 = ωµ0 k0 = b = 2p /l l = wavelength (in meters, m) j = −1 E is given in volts per meter (V/m) H is given in amperes per meter (A/m) For the ideal (delta) or infinitesimal electric (Hertzian) or magnetic (Fitzgerald) dipole sources, the radiation integrals are eliminated and the fields can be given in closed forms. The resulting field expressions can then be used to extract and study the usual antenna parameters.

1.3 HERTZIAN AND FITZGERALD ELEMENTARY RADIATORS Considering the infinitesimal electric dipole J = zˆ Idz δ (z), as illustrated in Figure 1-2, the resulting rms (root mean square) electric and magnetic field components are given by Er = k02 Eθ = jk02 Hφ = jk02

µ0 Idz  1 j  − jk r −   cosθ e 0 , 2 ε 0 2π  ( k0 r ) ( k0 r )3  µ0 Idz  1 j 1  − jk r − −   sinθ e 0 , 2 ε 0 4π  k0 r ( k0 r ) ( k0 r )3 

(1-3)

Idz  1 j  jk r −   sinθ e 0 , 4π  k0 r ( k0 r )2 

Eφ = H r = Hθ = 0 where Idz = moment of the differential current element ( I is given in rms amperes, and dz is given in meters)

FIGURE 1-2 Coordinate system for an electric dipole



CHAPTER ONE

A time factor of e jw t has been suppressed since a sinusoidally time-varying current excitation of constant frequency is assumed. These are the exact fields, but antenna parameter evaluation is usually carried out using simplified far-fields, i.e., when r is much greater than the wavelength l. Under these conditions, terms of order 1/r2 and greater are neglected to obtain Eθ = jk0

µ0 Idz e − jk0r sinθ = r ε 0 4π

Hφ = jk0

Idz e − jk0r sinθ = 4π r

µ0 H , ε0 φ

(1-4)

ε0 E . µ0 θ

The complex Poynting vector S3 in the far-field is given by

S = E × H* = k02

µ0 Idz sin 2θ rˆ ε 0 16π 2 r 2

(1-5)

showing that it is a purely real quantity and indicating that power transport is in the r direction (away from the elementary current) without any reactive energy. Also, it is seen that the radiated power density (power flow per unit area) for any r = const. depends on sin2q (independent of f ). This is referred to as the radiation pattern plotted in dB. For an elementary (ideal or infinitesimal) magnetic dipole M = zˆ I m dz δ ( z ), the radiated fields can be found by duality3 (replacing E by H, H by –E, and I by Im).

1.4 FAR-FIELD ANTENNA PROPERTIES, POWER TRANSFER, AND RECIPROCITY Because antenna radiation can be represented by radiation integrals over equivalent current sources, as considered in the previous paragraph, no reactive field components will be found in the far-field of any antenna. Further, far-field antenna characterization can be performed by considering power flow under the constraint of energy conservation. However, the distance from an antenna to its far-field depends on the antenna, and it is commonly accepted that the far-field region starts after the distance R=r=

2D2 λ

(1-6)

where D is the largest dimension of the antenna. This is due to the varying propagation distances of field contributions from different parts of the antenna to an observation point P, as illustrated in Figure 1-3. In the far-field, every antenna is considered a point source, and the far-field criterion in Eq. 1-6 is derived under the assumption that the phase errors due to the varying propagation distances are less than p/8. Consider an antenna located at the origin of a spherical coordinate system, as illustrated for the electric current element in Figure 1-2. Assume that the antenna is transmitting and let ●

Pt = power accepted by the antenna (in Watts)



1-9

FIGURE 1-3 Schematic representation of an antenna aperture showing the observation point P and the distances to the observation point from two points on the antenna

●

●

Prad = power radiated by the antenna (in Watts) h = radiation efficiency (unitless) These quantities are related by

η=

Prad Pt

(1-7)

Let ●

St(q,f) = power density (in Watts/square meter, W/m2)

and remark that St (q,f) is independent of the distance from the antenna r, a characteristic of the far-field region. The total radiated power can be obtained by integrating the power density over a surface enclosing the antenna. Such a surface can be of any shape and can be as close to the antenna as desired. However, for simplicity, it is convenient to choose the surface to be a sphere, giving Prad =

2π

π

∫0 ∫0 St (θ ,φ )r 2 sinθ dθ dφ

(1-8)

with the average power density being Pavg = Now, let ●

Prad

(1-9)

4π r 2

Dt(q,f ) = directivity (unitless)

Directivity is a measure of the antenna to concentrate radiated power in a particular direction, and it is related to power density as Dt (θ , φ ) =

St (θ , φ ) Pavg

=

St (θ , φ ) Prad / ( 4π r 2 )

(1-10)



CHAPTER ONE

The directivity of an antenna is the ratio of the achieved power density in a particular direction to that of an isotropic antenna. Let Gt(q,f) = gain (unitless) The gain of an antenna is related to the directivity and radiated power density via the relation Gt (θ , φ ) = η Dt (θ , φ ) =

ηSt (θ , φ ) Prad / ( 4π r 2 )

(1-11)

and from Eq. 1-7 Gt (θ , φ ) =

St (θ , φ ) Pt / ( 4π r 2 )

(1-12)

If you have a lossless antenna (i.e., h = 1), the directivity and the gain will be identical. Now consider a receiving antenna exposed to a power density radiated from some transmitting antenna. The ability of the antenna to receive energy is quantified through ●

Ae,r(q,f) = effective area (in square meters)

where the antenna’s location is assumed to be at the origin of the coordinate system. Under the assumption of reciprocity,5 the effective area of an antenna is related to the gain via the relation Ae,r (θ , φ ) =

λ2 G (θ , φ ) 4π r

(1-13)

where l = wavelength. Note that reciprocity holds only for lossless antennas. Also, Eq. 1-13 depends on wavelength and therefore on frequency. Under these circumstances, antenna characterization can be performed either for the transmitting or receiving case with behavior for the other being immediately known. Suppressing the angular dependencies of the transmitting and receiving antennas in their local coordinate systems, the received power is equal to the product of the power density of the incident wave and the effective aperture of the receiving antenna. That is Pr = St Ae,r Substituting from Eq. 1-12 and Eq. 1-13 yields Pr =

Gt Pt λ 2 Gr 4π r 2 4π

or 2

 λ  Pr =   Gt Gr Pt  4π r 

(1-14)



−3 dB

1-11

0 dB

q qfirst null

q3dB

Backlobe Sidelobes

FIGURE 1-4 Antenna pattern in plane f = const

Note that Gt is the gain of the transmitting antenna in the direction of the receiving one and Gr is the gain of the receiving antenna in the direction of the transmitting one. A form of this equation was presented first by Friis6 and is usually called the Friis transmission formula. The angular dependence of the radiating and receiving properties of an antenna in the far-field is often referred to as the antenna radiation pattern. Thus, a pattern is a normalized plot of the directivity, gain, or effective aperture as a function of angle and is often given in dB scale. Typically, the radiated normalized radiated power density or radiated field is plotted in dB (for the infinitesimal or ideal dipole, the power density sin 2 θ is plotted in dB). A typical antenna pattern has a main lobe, sidelobes, minor lobes, a backlobe, and several nulls, as illustrated in Figure 1-4, in a f = const. plane. The half-power or 3 dB beamwidth of the main lobe (or main beam) is indicated in the drawing. If the pattern of an antenna is given in a plane parallel to the E field vector, the corresponding pattern is referred to as an E plane pattern. Alternatively, if the pattern cut is in a plane parallel to the H field polarization, it is called an H plane pattern. There are many types of antenna radiation patterns, but the most common are ●

●

●

●

Omnidirectional (azimuthal-plane) Pencil beam Fan beam Shaped beam

The omnidirectional pattern is most popular in communication and broadcast applications. The azimuthal pattern is circular, but the elevation pattern has some directivity to increase the gain in the horizontal direction. The term pencil beam is applied to a highly directive antenna pattern consisting of a major lobe contained within a cone of a small solid angle. Highly directive antenna patterns can be employed for point-to-point communication links and help reduce the required transmitter power. A fan beam is narrow in one direction and wide in the other. A fan beam is typically used in search or surveillance radars. Shaped beam patterns are adapted to the requirements of particular applications.

1.5 ANTENNAS AS ELECTROMAGNETIC CIRCUITS A symbolic transition between a waveguide (transmission line) and an antenna is shown in Figure 1-5. In the case of a radiating antenna, a guided wave with amplitude a is traveling



CHAPTER ONE

toward the antenna and is more or less radiTransmitting a ated into the surrounding medium. In the case of a receiving antenna, the received energy is transferred into a guided wave with amplitude b traveling down the waveguide away from the b antenna. Consequently, from a circuit’s viewReceiving point, the antenna can be considered as a one1-5 Transition between waveguide port element provided only one guided mode FIGURE (transmission line) and antenna exists at the port, as is normally the case; higher order modes can be considered as additional ports. In the radiating mode, the one-port element is characterized by its complex reflection coefficient, Γ=

b a

(1-15)

In the receiving mode, the received signal is converted into a traveling wave of amplitude b0. In general, the guided wave amplitude b at the antenna port is given by b = Γa + b0

(1-16)

If the antenna is a two-terminal circuit, equivalent network parameters are employed to describe the antenna behavior within the context of an electromagnetic circuit. Accordingly, the antenna may be replaced by equivalent circuit models such as those given in Figure 1-6, where the antenna is characterized by a) its input impedance Z and a source voltage V0 or b) its input admittance Y and a source current I0. In the transmitting mode, V0 or I0 is zero, and Vin or Iin is the input voltage or current that drives the antenna. Part of the source power is dissipated in the loss resistance Rloss or loss conductance Gloss. The power associated with the radiation resistance Rrad or radiation conductance Grad is responsible for the radiated wave that escapes from the antenna to the surrounding space. The quantities X and B refer to the equivalent reactance and susceptance of the antenna. When the antenna is receiving, the terminals of the equivalent circuit are connected to a load (amplifier, speaker, etc.), and V0 or I0 are non-zero as dictated by the external wave that impinges on the antenna. Maximum transmission to the load occurs when the load impedance is the conjugate of the input antenna impedance. Some of the received energy usually remains within the radiation resistance or conductance and is re-radiated to contribute to the radar cross section (RCS) of the antenna. Mismatches between the load and the antenna input impedance can lead to higher or lower RCS. However, care must be exercised when

Rloss V0 + −

+ Z

Vin

Rrad

−

+ Y −

Iin Gloss B

I0

Grad

X (a)

(b)

FIGURE 1-6 (a) Impedance (Thevenin equivalent) and (b) admittance (Norton equivalent) representations of an antenna



1-13

Rloss 4kTN ∆fGrad

+ −

4kTN ∆fGrad

+ V0 −

Rrad

Y

Gloss

B

I0

Grad

X (a)

(b)

FIGURE 1-7 (a) Impedance (Thevenin equivalent) and (b) admittance (Norton equivalent) representations of an antenna including noise sources

the re-radiated energy is determined using equivalent circuit representations such as those in Figure 1-6. Reasonable results can only be expected if the employed equivalent circuit is an appropriate representation of the antenna’s internal structure. Also, the equivalent circuits of one antenna can be different for radiation and reception. Each antenna receives noise radiation related to the brightness temperature of its environment. Usually, the received noise power restricts the performance of an antenna and subsequent processing devices. However, in remote sensing or astronomy applications, the noise power can even be the intended signal. To formally account for the received noise power, the equivalent circuits in Figure 1-6 must be used with V0 and I0 representing sources dependent on the equivalent noise temperature TN of the antenna. The dependent source voltages and source currents are given in Figure 1-7, where k is the Boltzmann constant and BW = ∆f is the receiver or antenna bandwidth. TN is dependent on the radiation (or rather receiving) pattern and orientation of the antenna as well as the distribution of the brightness temperature in the environment of the antenna.† Usually, antennas work at certain frequencies. The bandwidth BW = fU – fL

(1-17)

is the operational frequency range of the antenna around some center frequency f0 = 1/2(fU – fL)

(1-18)

(e.g., the resonance frequency of the antenna). In this equation, the upper operational frequency limit is fU, and the lower operational frequency is fL. Often, the bandwidth is quoted with respect to f0 percent, and in that case, it is given by BW = (fU – fL)/f0 100%

(1-19)

From a circuit point of view, the matching of an antenna to the generator is very important. Therefore, bandwidth definitions with respect to the reflection coefficients or input impedances are common, for example, |G | = |(Zin – Z0)/(Zin + Z0)| < 0.5

†

(1-20)

Further noise considerations with respect to antennas can be found in the literature.1,2,11,16



CHAPTER ONE

Typically, the voltage standing wave ratio (VSWR) VSWR = (|G | + 1)/(|G | – 1) < 2

(1-21)

is used to define the operational bandwidth. Bandwidth definitions with respect to gain or other antenna pattern characteristics (for example, G0/G0,max < 0.5) can be used to characterize the antenna radiating and receiving properties.

1.6 POLARIZATION Polarization is a property of a single-frequency electromagnetic wave; it describes the shape and orientation of the locus of the extremity of the field vectors as a function of time.7,8 In antenna engineering, the polarization properties of plane waves or waves that can be considered to be planar over the local region of observation are of interest. For plane waves, it is sufficient to specify the polarization properties of the electric field vector since the magnetic field vector is simply related to the electric field vector. The plane containing the electric and magnetic fields is called the plane of polarization and is orthogonal to the direction of propagation. Generally, the tip of the electric field vector moves along an elliptical path in the plane of polarization. The polarization of the wave is specified by the shape and orientation of the ellipse and the direction in which the electric field vector traverses the ellipse. The shape of the ellipse is specified by its axial ratio—the ratio of the major axis to the minor axis. Its orientation is specified by the tilt angle—the angle between the major axis and a reference direction when viewed looking in the direction of propagation. The direction in which the electric field vector traverses the ellipse is the sense of polarization— right-handed or left-handed when viewed looking in the direction of propagation. The polarization of an antenna in a specific direction is defined to be the polarization of the far-field radiated in that direction. Usually, the polarization of an antenna remains relatively constant throughout the main lobe, but varies considerably in the minor lobes. It is convenient to define a spherical coordinate system associated with an antenna as illustrated in Figure 1-8. The polarization ellipse for the direction (q,f) is shown inscribed on

FIGURE 1-8 Polarization ellipse in relation to the antenna coordinate system (after7 © IEEE 1979)



1-15

the spherical shell surrounding the antenna. It is common practice to choose uq (the unit vector in the q direction) as the reference direction. The tilt angle is then measured from uq toward uf. The sense of polarization is clockwise if the electric field vector traverses the ellipse from uq toward uf, as viewed in the direction of propagation and counterclockwise if the reverse is true. In many practical situations, such as antenna measurements, it is convenient to establish a local coordinate system. Usually, the u3 axis is the direction of propagation; the u1 axis is horizontal; and the u2 axis is orthogonal to the other two so the unit vectors are related by u1 × u2 = u3. The tilt angle is measured from u1. When an antenna receives a wave from a particular direction, the response will be greatest if the polarization of the incident wave has the same axial ratio, the same sense of polarization, and the same spatial orientation as the polarization of the antenna in that direction. This situation is depicted in Figure 1-9, where Et represents a transmitted wave (antenna polarization) and Em represents a matched incident wave. Note that the sense of polarization for Et and Em are the same when viewed in their local coordinate system. Also, note that the tilt angles are different because the directions of propagation are opposite. As depicted in Figure 1-9, tt is the tilt angle of the transmitted wave and tm is the tilt angle of the polarization-matched received wave; they are related by

τ m = 180° − τ t

(1-22)

The polarization of the matched incident wave, as just described, is called the receiving polarization of the antenna. When the polarization of the incident wave is different from the receiving polarization of the antenna, then a loss due to polarization mismatch occurs. Let hp = polarization efficiency (unitless) where hp is the ratio of the power received by the antenna to the power received when polarization of the incident wave is matched to the receiving antenna polarization. The Poincaré sphere, shown in Figure 1-10, is a convenient representation of polarization states. Each possible polarization state is represented by a unique point on the unit sphere. Latitude represents axial ratio, with the poles being circular polarizations; the upper hemisphere is for left-handed sense, and the lower hemisphere is for right-handed sense. Longitude represents tilt angles from 0 to 180°. An interesting feature of the Poincaré sphere is that diametrically opposite points represent orthogonal polarizations.

FIGURE 1-9 Relation between polarization properties of an antenna when transmitting and receiving (after7 © IEEE 1979)



CHAPTER ONE

FIGURE 1-10 Polarization states on the Poincaré sphere (after7 © IEEE 1979)

The Poincaré sphere is also convenient for representing polarization efficiency. In Figure 1-11, W represents the polarization of an incident wave, and Ar is the receiving antenna polarization. If the angular distance between the points is 2x, then the polarization efficiency is

η p = cos 2 ξ

(1-23)

FIGURE 1-11 Receiving polarization of an antenna Ar for an incident wave polarization W



1-17

1.7 DIRECTIVITY PATTERNS FROM CONTINUOUS LINE SOURCES According to Huygens’ and the equivalence principle (see Section 1.2), the radiation of arbitrary antennas can be characterized by considering equivalent source current distributions. The simplest source distributions are the electric and magnetic point sources considered in Section 1.3. More degrees of freedom toward the realization of particular directivity properties are provided by continuous line sources that can, for instance, be used to characterize the radiation behavior of linear wire antennas. For line sources, the current distribution (electric and/or magnetic) is considered to be a function of only a single coordinate. The directivity pattern E(u) resulting from a given distribution is simply related to the distribution by a finite Fourier transform,5,9,10 given by E (u ) =

 +1 f ( x ) e jux dx , 2 ∫ −1

(1-24)

where f(x) = relative shape of source distribution over aperture as a function of x u = (pl) sin f  = overall length of aperture f = angle measured from normal to aperture x = normalized distance from normal to aperture –1 ≤ x ≤ 1 The simplest type of line source distribution is the uniform distribution where f(x) = 1 along the aperture and is zero elsewhere. The directivity pattern for such an antenna is (see Chapter 3)  π       sin φ λ    sin u = E (u ) =  u π     sin φ λ 

(1-25)

This type of directivity pattern 20 log10 |E(u)| is of interest because all field contributions add in phase, giving the highest gain without cancellation effects.5 However, sidelobe levels are high, and the intensity of the first sidelobe is only 13.2 dB less than the maximum. The intensity of the sidelobe levels can be reduced considerably by tapering the aperture distribution so the amplitude drops off smoothly from the center of the aperture to its edges. There are an unlimited number of possible distributions. However, a few simple types of distributions are typical and illustrate how the beamwidth, sidelobe level, and relative gain vary as a function of the distribution. Table 1-1 gives the important characteristics of several distributions having a simple mathematical form. Of considerable interest is the manner in which the sidelobes fall off as the angle from the main beam increases or as u increases. For the uniform distribution that has a discontinuity in both the function and its derivatives at the edge of the aperture, the sidelobes decrease as u–1. For the gable distribution or cosine distribution, both of which are continuous at the edge of the aperture but that have discontinuous first derivatives, the far-out sidelobes fall off as u–2. For the cosine-squared distribution that has a discontinuous second derivative, the far-out sidelobes fall off as u–3. Many distributions (obtained in practice) can be approximated by one of simpler forms or by a combination of simple forms.


Fundamentals of Antennas, Arrays, and Mobile Communications 1-18 TABLE 1-1

CHAPTER ONE

Line-Source Distributions

For instance, suppose you wanted to find the directivity pattern of a cosine-squared distribution on a pedestal, i.e., a combination of a uniform distribution and a cosine-squared distribution given by f ( x ) = C + cos 2

πx 2

(1-26)

The resulting directivity pattern is then obtained directly by superposition to obtain the direction pattern: E (u ) = C 

sin u  sin u π 2 + . u 2 u π 2 − u2

(1-27)

Note that the sidelobes and other characteristics of the pattern must be obtained from the new directivity pattern and cannot be interpolated from Table 1-1. By choosing the proper relative intensities of a uniform distribution and a cosine-squared distribution, it is possible to obtain a theoretical sidelobe level that is very low. For instance, if C = 0.071, then the intensity of the largest sidelobe will be 43 dB less than the maximum of the main beam with a half-power beamwidth given by 76.5l/, a value that is somewhat lower than that of the cosine-squared distribution by itself. In practice, it is not easy to synthesize prespecified continuous line-source distributions. Consider, for instance, a linear wire antenna; the electric current distribution along the wire is determined to fulfill Maxwell’s equations under the constraint of the given boundary Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.


z q

l = 0.25 l

2l

y

l = 0.5 l

x f

z z

z

l = 0.75 l

2a (a)

1-19

(b)

(c)

FIGURE 1-12 (a) Linear wire antenna geometry, (b) Current distributions according to Eq. 1-28, and (c) corresponding field radiation patterns in arbitrary f = const. planes (linear scale) for the first three resonances

conditions and cannot, therefore, be shaped arbitrarily. Good approximation of the current can be done, however, to study radiation from specific antennas. In the following considerations, focus will be on a linear center-fed dipole wire antenna of length 2l aligned along the z axis, as illustrated in Figure 1-12a. The characteristics of the corresponding monopole antenna of length l over a perfectly conducting plane can be obtained by applying image theory.1,3 As a first approximation, the thin wire antenna can be replaced by the z directed current distribution I ( z ) = I 0 sin(k (l − | z |))

(1-28)

defined along the wire length. The corresponding electric field intensity in the far-field is3 E(r ) =

jI 0 µ0 − jk0r  cos(k0 l cosθ ) − cos(k0 l )  ˆ e  θ 2π r ε 0 sinθ  

(1-29)

and can be used to derive an approximate far-field pattern. Obviously, the pattern is omnidirectional, and its q dependence varies with antenna length. The q dependence (in linear scale) of the field radiation pattern is illustrated in Figure 1-12c for three different wire lengths, with the corresponding current distributions according to Eq. 1-28 as depicted in Figure 1-12b. The radiation resistance can be calculated by integrating the total radiated power and relating it to the appropriate input current at the antenna port.1,3 Evaluating this expression gives the resistance as depicted in Figure 1-13. Wire antennas are usually operated close to their resonance lengths 2l/l = 0.5, 1, 1.5, 2, ..., where 2l/l = 0.5, 1.5, ... corresponds to the current resonance with low radiation resistance. Lengths 2l/l = 1, 2, ... result in voltage resonance with high radiation resistance. The lowest-order resonance 2l/l = 0.5 is associated with a radiation resistance of approximately 73 Ω. Difficulties due to nulls in the current distributions for voltage resonances can be overcome by a modified higher-order current distribution.‡1 Figure 1-14 ‡

Extensive studies of linear wire antennas, even with realistic geometric dimensions, are presented in R. W. P. King’s The Theory of Linear Antennas.12 Approximate integral equation solutions are the basis for characterizing the various antenna parameters and results for a large variety of parameters are given.



CHAPTER ONE

Rrad

350 300 250

Ω

200 150 100 73 50 0

0

0.5

1

1.5

2

2.5

3

3.5

2l l

4

FIGURE 1-13 Wire antenna radiation resistance obtained from the current distribution of Eq. 1-28 (normalized to I0)

Ω

Im (Z )

1500 l = 904 a

0.41 1000

0.37 0.43

0.32

500

0.33

0.24

0.33

0

0.75

0.38 0.41

0.51 −500

0.44

0.43

0.45

0.46

0.54 0.56

−1000 −1500

l = 122.4 a

l = 259 a

0.49 0.48

0.51 0

500

1000

1500

2000

Re (Z ) 2500 Ω

FIGURE 1-14 King–Middleton12 second-order impedances for different values of l/a (see Figure 1-12). The numbers given next to the dots are the l/l values.



1-21

illustrates the complex input impedance of the wire antenna for different l/a ratios and l/l values.§ You can see that thicker antennas have smaller impedance oscillations resulting in larger bandwidths. Also, the impedance curves show a slight shift of the reactance to the capacitive range. A large collection of numerical computer codes, often based on the method of moments1,11 (see Chapter 59 on numerical techniques), is now available for analysis, design, and optimization of many wire antenna types.

1.8 DIRECTIVITY PATTERNS FROM AREA SOURCE DISTRIBUTIONS Area source distributions serve as a radiation model for many antenna types, especially for those with radiation in a particular direction. Insight into various antenna characteristics can be gained from the consideration of simple aperture shapes such as rectangular, circular, or elliptical apertures. Rectangular Apertures The directivity pattern of an area distribution is found in a similar manner to that used for line-source distributions, except the aperture field is integrated over two dimensions instead of one dimension. If the aperture distribution is given by f(x, y), where x and y are the two coordinates, then the directivity pattern is given by E (θ , φ ) = ∫∫ f ( x, y)e jk0 sin θ ( x cosθ + y sin φ ) dx dy

(1-30)

The difficulty of evaluating this expression depends on the form of the distribution function. For many antenna types, such as the rectangular horn, for example, the distribution function is separable, that is f (x, y) = f (x)f (y) The directivity patterns in the principal planes are readily determined for the separable case because the pattern in the xz plane is identical to the pattern produced by a line-source distribution f (x), whereas the pattern in the yz plane is identical to the pattern produced by a line-source distribution f (y). If the distribution function is not separable, the integral must be evaluated either analytically, graphically, or numerically. Circular Apertures An antenna that is used frequently in microwave applications is a paraboloid having circular symmetry. The radiation pattern can be computed by projecting the field distribution §

The curves in Figure 1-14 represent the King-Middleton second-order impedances and are drawn from the numerical data in The Theory of Linear Antennas.12



CHAPTER ONE

on the paraboloid to a plane at the opening of the paraboloid and computing the directivity pattern due to the plane aperture: E (u, φ ′ ) = a 2

2π

1

∫0 ∫0

f (r , φ ′ )e jur cos(φ −φ ′ ) r dr dφ ′

(1-31)

where a = radius at outside of aperture p = radius at any point inside aperture r = p/a u = (2pa/l) sin q = (p D/l) sin q D = 2a = aperture diameter f (r, f ′ ) = the normalized aperture distribution function The coordinates are shown in Figure 1-15. The simplest forms of aperture distributions to evaluate are those in which the distribution is independent of the angular coordinate f ′ but depends only on the radial coordinate r. The integral for the directivity pattern then becomes 1

E (u ) = 2π a 2 ∫ f (r ) J 0 (ur )r dr 0

(1-32)

When the distribution is constant, the integral is evaluated to give E (u ) = 2π a 2

J1 (u ) u

(1-33)

Antenna engineers frequently need to evaluate the directivity pattern for an illumination that tapers down toward the edge of the aperture. One function, which is convenient for representing the aperture distribution, is f (r) = (1 – r 2) p

(1-34)

This function behaves in a similar fashion to the nth-power distributions, as discussed for the line-source case (Section 1.7). When the exponent increases, the distribution becomes more tapered and more concentrated to the center of the aperture. When the exponent decreases and approaches zero, the distribution approaches uniform illumination.

FIGURE 1-15 Coordinates for a circular aperture



1-23

Evaluating the directivity pattern, gives E (u ) = 2π a 2 = π a2

1

∫0 (1 − r 2 ) p J 0 (ur ) dr

2 p p! J p +1 (u ) u

p +1

=

a2 ∆ (u ), p + 1 p +1

(1-35)

where the Bessel functions Jp+1(u) and the lambda function Dp+1(u) are available in tabular form.13 The principal characteristics of the directivity patterns are given in Table 1-2 for the cases p = 0, 1, 2. Comparison of the patterns for the uniformly illuminated circular aperture (i.e., when p = 0) with results for the uniformly illuminated line source (Section 1.7) shows that the circular aperture has a lower sidelobe level and a broader beamwidth. This is expected because projections of the circular-aperture illumination onto a line produce an equivalent line source that is no longer uniform but has some degree of tapering. Elliptical Apertures In some applications, an elliptically shaped reflector is used to permit control of the relative beamwidth in the two principal planes and to control the sidelobes by shaping the reflector outline. Computation of the directivity patterns for the aperture shape can be carried out by knowing the Fourier components of the illumination function over the aperture. Realization of Continuous Aperture Distributions In practice, continuous aperture distributions of large extent are realized by horn, reflector, or lens antennas. TABLE 1-2

Circular-Aperture Distributions



CHAPTER ONE

Horn antennas (see Figure 1-16) are among the oldest and most popular microwave antennas. They can deliver 10 to 30 dB directivity, and their robust metallic constructions and waveguide feeds permit high-power handling. Horns are, therefore, often used as feeds for reflector antennas realizing even larger apertures. Horns are extensions of the widely used microwave guiding devices, such as rectangular and circular hollow waveguides in a natural manner (see Figure 1-16). To achieve matching from the waveguide to free-space, the hollow waveguide is tapered to a larger opening called the aperture. The effective aperture of the antenna (Section 1.4) is almost identical to the geometric aperture of the horn. The equivalent aperture source distributions are typically used for extracting the radiation properties of horn antennas. These sources are an approximation found from the guided wave field arriving at the horn’s aperture. Smaller beamwidths and larger directivities can be expected for horns with larger cross-section apertures, a consequence of the Fourier transformation relation between aperture distributions and far-field patterns. However, the maximum achievable directivity is limited due to phase errors of the aperture source distributions caused by the curved phase fronts in the tapered part of the horn antenna. Phase error corrections can be realized by using lenses in the horn aperture or by connecting the horn to a parabolic reflector (see Figures 1-17, 1-18, 1-19). Another important concept in designing horn antennas is corrugation. Corrugations are grooves on the waveguide walls and are equivalent to introducing anisotropic boundary conditions on these walls. These anistropic conditions offer additional degrees of freedom for controlling the radiation pattern and have been shown to lead to lower cross polarizations, higher aperture efficiencies, and more symmetrical far-field patterns. The E plane sectoral horn is tapered only in the plane containing the E field (of the lowest-order rectangular waveguide mode), and the H plane sectoral horn is tapered in the corresponding plane containing the H field. Correspondingly, the pyramidal horn is tapered in both planes, and horns that are extensions of a circular waveguide are referred to as conical horn antennas. Further details on the analysis and practical realization of horn antennas can be found in Chapter 14 and the literature.1,2,14

(a)

(b)

(c)

(d)

FIGURE 1-16 Common electromagnetic horn antennas: (a) E plane horn, (b) H plane horn, (c) pyramidal horn, (d) conical horn



Focal Line

1-25

Focal Point

Aperture

Aperture

(a)

(b)

FIGURE 1-17 (a) Cylindrical parabolic reflector with focal line and (b) paraboloidal reflector with focal point

(a)

(b)

FIGURE 1-18 (a) Cassegrain and (b) Gregorian double-reflector systems

(a)

(b)

FIGURE 1-19 (a) Delay lens with index of refraction > 1 and (b) fast lens with index of refraction < 1



CHAPTER ONE

From the preceding paragraphs and especially with knowledge of the Fourier transform relation between far-field patterns and aperture distributions, you can observe that high-gain antennas with small beamwidths typically require large (equivalent) aperture source distributions. Even though large apertures can be realized, however, unavoidable phase errors, as in the case of horn antennas, restrict achievable gain and correspondingly the effective aperture. A way to overcome this issue is to use a secondary device for shaping the phase fronts of the radiated or received waves. Metallic reflectors (curved metallic surfaces) serve this purpose and have been used since the early days of antennas to realize high gains. There are many different reflector types2 ranging from linear reflector elements over flat metallic sheets to more complicated shapes such as corner reflectors, circular reflectors, hyperbolic reflectors, elliptic reflectors, and the more commonly used parabolic reflectors. Next, we briefly discuss parabolic reflectors. For other reflector types, refer to Chapter 15 on reflector antennas and the literature.1,2,15 Figure 1-17 depicts the cylindrical and spherical parabolic reflectors. Both utilize the unique feature of the parabola to adjust the ray path lengths so that spherical waves emanating from the focal point generate fields of the same phase on the reflector aperture after reflection from the reflector surface. As a result, the rotationally symmetric paraboloid gives rise to very narrow patterns. Highly directive antennas can be realized by placing a primary line source in the focal line of a cylindrical paraboloidal reflector or a point source (a horn) at the focus of a paraboloid. Of course, these primary sources should be directed toward the reflector. Because the feed causes undesirable blockage and distortions to the main beam (a situation that worsens due to collateral equipment around the feed), subreflectors are introduced to place the feed at different locations. Shown in Figure 1-18 are classical Cassegrain and Gregorian double-reflector arrangements illustrating this concept. Cassegrain antennas use a hyperbolic subreflector whereas Gregorian reflectors employ an elliptical subreflector. Both subreflector types are usually designed so the feed or new focus can be located close to the primary reflector permitting feeding of the source at the focus through a hole in the primary reflector. Electromagnetic analysis of reflector antennas usually follows the principles of geometrical and physical optics, or geometrical theory of diffraction1,2,11,16 to account for edge diffraction. Two different analysis techniques are often employed for reflector analysis: the aperture distribution and current distribution methods. For the latter, the electric currents are found directly on the reflector dishes, and the radiated fields are calculated by integrating these currents (see Section 1.3). In the aperture distribution method, equivalent currents are derived and placed on a planar aperture in front of the reflector, as illustrated in Figure 1-17. The radiated fields are then found by integrating these equivalent currents. As noted earlier, aperture blocking by the primary feed or the subreflector reduces antenna efficiency. Also, the blockage gives rise to higher sidelobe levels and cross-polarization effects. These blockages can be circumvented by reshaping the reflector so the feed is offset from the center of the aperture,17 often below or outside of the aperture view. Similar to reflector antennas, lens antennas provide a means of shaping phasefronts of electromagnetic waves and influencing wave propagation in certain directions. Lens antennas can be divided into delay lenses and fast lenses (see Figure 1-19).2 In a delay lens medium, the electromagnetic path length is increased (refractive index n > 1), whereas in a fast lens medium, the electromagnetic path length is decreased inside the lens (refractive index n < 1). Lenses may also be divided among dielectric and metal-plate lenses. Dielectric lenses consist of natural dielectrics or artificial dielectrics. Metal-plate lenses consist of parallel metallic plates realizing parallel-plate waveguides between the individual plates. H plane metal-plate lenses have metallic plates parallel to the H field of the electromagnetic wave and lead to n > 1. In contrast, the E field metal-plate lenses have the plates parallel to the E field resulting in n < 1. As noted earlier, lens antennas are often used to form collimating beams and thus increase the gain of microwave antennas. In this context, their purpose is to transform spherical or cylindrical wavefronts into planar wavefronts, and this is the reverse of focusing plane waves into a focal line or point (see Figure 1-19). Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.


1-27

In contrast to reflector antennas, lens antennas do not have difficulties with aperture blocking and allow for beam scanning over relatively large angles. They also provide for additional design freedom since wave propagation is influenced by refraction at both the surfaces and the refractive index inside the medium, which can be adjusted as required (for instance by using layered designs). However, the corresponding design process is more complex when compared to metallic reflector antennas. Difficulties with lenses are caused by reflections at the two surfaces and losses in the lens material. Nevertheless, surface reflections can be reduced by matching layers or surface roughness. Analysis and design procedures for lens antennas are typically based on geometrical optics (Fermat’s principle), physical optics, and related approaches such as the geometrical theory of diffraction. Details on the analysis of lens antennas and on the various classes of such antennas can be found in the literature1,2,11,16 and in Chapter 18 on lens antennas. Recently, ideal lenses with arbitrarily good focusing properties have been discussed. Such lenses should be realizable by the use of so-called metamaterials with a refractive index of –1.18

1.9 FUNDAMENTALS OF ANTENNA ARRAYS The radiation and receiving characteristics of antennas can be shaped by synthesizing certain equivalent source distributions. Because it is difficult to control continuous aperture currents or fields, discrete configurations are often used, leading to the concept of antenna arrays. Putting the elements of an antenna array in a certain pattern and adjusting the amplitude and phase of the individual antenna elements appropriately allows for the synthesis of arbitrary aperture sources. These behave quite similarly to continuous aperture distributions (provided certain rules with respect to element spacing are followed). Design difficulties often arise due to coupling among array elements. Consequently, the amplitude and phase of the individual array elements cannot be adjusted independently from one another. The driving-point impedance of an individual element might differ considerably from its selfimpedance because of the mutual coupling with other array elements. In a multi-element array, a way to relate the terminal voltages and element currents is V1 = I1Z11 + I2Z12 + ⋅ ⋅ ⋅ + InZ1n V2 = I1Z12 + I2Z12 + ⋅ ⋅ ⋅ + InZ2n ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ Vn = I1Z1n + I2Z2n + ⋅ ⋅ ⋅ + InZnn

(1-36)

where Vn = impressed voltage at the nth element In = current flowing in the nth element Znn = self-impedance of the nth element Zmn = Znm = mutual impedance between mth and nth elements The driving-point impedance for element 1 is found from the ratio of the impressed voltage to the current and is obtained from the previous equation as follows: Z1 input =

V1 I I = Z11 + 2 Z12 + ⋅ ⋅ ⋅ + n Z1n I1 I1 I1

(1-37)

The reader can see that the input or driving-point impedance of a particular element is not only a function of its own self-impedance but also a function of the relative currents Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.


CHAPTER ONE

z

z

aN−1

a0N −1

aN−2

a0N−2 q

a2

a02 d

a0

dy

y d cosq

x

dz

a01

a1

a00

(a)

a10

aN −20

aN −10

y

(b)

FIGURE 1-20 (a) Equally spaced linear and (b) two-dimensional array configurations

flowing on the other elements and of mutual impedance between elements. In an array in which the current distribution is critical, it is necessary to determine the input impedance from the previously described relationship and to design the feed system to match the input impedance rather than the self-impedance.∗∗ To account for element coupling in the design, rigorous numerical methods1,4,11 must be employed. However, for simplicity, in the following array, coupling will be neglected because the main focus is the introduction of some basic concepts in array design. A very basic and important array configuration is that of a equally spaced linear array of N identical elements on a straight line, as illustrated in Figure 1-20a). The element spacing is d, and a linear phase progression is assumed for the element excitation currents. The total electric field intensity Etot in the far-field is given by N −1

E tot (θ , φ ) = E el (θ , φ ) ∑ an e jn ( k0 d cosθ −α ) n=0

N −1

= E el (θ , φ ) ∑ an e jnψ = E el (θ , φ ) f (ψ )

(1-38)

n=0

where ψ = k0 d cosθ − α and E el refers to the array element pattern. Also, an is the amplitude of the individual array elements; a is the phase progression from one element to the next; and k0 is the wavenumber of free-space. If all excitation currents are equal in amplitude (a0 = a1 = a2 = ... = aN-1), the array factor y becomes N −1

f (ψ ) = a0 ∑ e jnψ = a0 n=0

1 − e jNψ 1 − e jψ

(1-39)

**

Some examples of this are given in “The Effect of a Periodic Variation in the Field Intensity Across a Radiating Aperture.”19 See Chapter 3 for more information.



FUNDAMENTALS OF ANTENNAS, ARRAYS, AND MOBILE COMMUNICATIONS

This can be simplified to obtain the normalized form f (ψ ) f0 (ψ ) = = Na0

Nψ 2 ψ N sin 2 sin

(1-40)

where f0 (y ) is maximum when y = 0. For broadside radiation, a = 0 must hold, implying that q = p / 2. To scan the array beam toward other directions, a must be selected so that k0d cosq = a, resulting in α  θ = θ max = cos −1    k0 d 

(1-41)

as the angle of the maximum array radiation. The direction of maximum radiation can be adjusted by controlling a, and this is a concept characteristic to scanning arrays. In practical phased arrays, the phase of the individual array elements is controlled electronically to allow for much more flexible, faster, and reliable array scanning as compared to traditional mechanical steering approaches. Modern mobile communication systems are increasingly employing phased arrays at base stations (in conjunction with sophisticated signal processing algorithms) to expand the base station customer capacity and reduce interference among adjacent stations. In the wireless industry, such antennas are typically referred to as smart or adaptive antennas. If the spacing d between the array elements becomes greater than half a wavelength l0, the denominator of f0 (y ) can have further zeros, resulting in additional array radiation beams. These beams are referred to as grating lobes. In practice, suppressing such parasitic lobes is required. Figure 1-21 illustrates several array patterns derived from f0(y ). These show the characteristic narrow main beam and a larger number of sidelobes as the number of array

0° 30°

30°

60°

90°

0 dB

60°

−10

−20

−30

−30

−20

−10

N=5

N = 19

N = 19

a=0

a=0

a = 2.0

90° 0 dB

FIGURE 1-21 Linear array pattern factors f0 (array element spacing d = 0.4l0)



CHAPTER ONE

elements increases. Lower sidelobe levels can be achieved by tapering the excitation of the array elements. Several techniques, such as the binomial or the popular Chebyshev methods (see Chapter 3 on array design) are used for controlling the sidelobe levels and the main beamwidth. Next, we consider the 2D linear array in Figure 1-20b. This array has M elements along the z-axis and N elements along the y-axis. The array elements are equally spaced with corresponding spacings of dz and dy and associated linear phase progressions of az and ay. For the computation of the array factor f (q,y ), the 2D array can be viewed as M equally spaced linear arrays along the z-direction. It is given by N −1 M −1

f (θ , φ ) = ∑ ∑ amn e jn ( k0 dz cosθ −α z ) n=0 m=0

N −1 M −1

= ∑ ∑ amn e jnψ z e jmψ y n=0 m=0

(1-42)

where yz = k0 dz cosq − az and yy = k0 dy cosq − ay. Assuming the amplitudes of all array elements to be identical (a00 = a 0 = ... = aM-1N-1) and applying the same manipulations as in the case of the linear array, the normalized array factor f0 (q,y ) = f0 (q,y )/(MNa00) is found to be  Mk0 d y   Nk d  sinθ sinφ sin  0 z cosθ sin  2 2     f0 (θ , φ ) =  k0 d z   k0 d y  N sin  cosθ M sin  sinθ sinφ 2  2   

(1-43)

In general, the amplitude of every array element can be chosen independently in order to shape arbitrary array patterns. Conventional beam-shaping techniques try to approximate directivity patterns known from continuous aperture distributions (Sections 1.7 and 1.8). More recently, digital signal processing techniques are employed to realize a large variety of applications such as angle of arrival detection (see Chapter 47 on direction finding), tracking, interferer suppression, and adaptive signal to interference improvement. A severe disadvantage of linear array configurations is their restricted field of view. Reasonable beam-shaping and scanning can be achieved by modifying the array element amplitudes. However, many arrays have restricted angular range. This drawback can be overcome by array configurations conforming to curved surfaces. As such, every array element may have a different geometrical orientation, and suitable element amplitude directivity behavior can be a difficult task. Beam-shaping techniques have been developed incorporating digital signal-processing techniques for the necessary flexibility to achieve desirable designs (see Chapter 22). A disadvantage of an array conformed to a curved surface is that not all antenna elements contribute to the radiated fields in a particular direction. Thus, more array elements are required to achieve certain requirements with respect to beamwidths and sidelobe suppression. An example of a singly curved conformal array antenna is shown in Figure 1-22.



1-31

FIGURE 1-22 Singly curved array antenna realized with patch antenna elements (Courtesy of FGAN e.V., Wachtberg, Germany)

Another array configuration are volumetric arrays. To avoid shadowing of the individual array elements, the grid must be relatively thin, and reasonable beam-shaping is achieved only by randomly distributing array elements to suppress grating lobes due to large element spacing. In principal, all array elements can radiate in all directions. However, practical realization of volumetric arrays is not easy, and the choice of suitable array elements and feeding techniques is restricted. An example of a volumetric array is depicted in Figure 1-23.

FIGURE 1-23 Volumetric array antenna with randomly distributed loop elements: Crow’s nest antenna by FGAN e.V., Wachtberg, Germany20 (Courtesy of FGAN e.V., Wachtberg, Germany)



CHAPTER ONE

1.10 BASIC CONCEPTS IN MOBILE COMMUNICATIONS Mobile communications is an area where unguided electromagnetic waves and antennas play an important role. The goal is to offer communication links between any place on earth without limiting user mobility. Given the restricted spectrum availability, the need for frequency reuse, particularly in densely populated areas, creates special challenges that are most efficiently resolved within terrestrial mobile communication systems. However, the requirement to cover remote regions makes essential the use of satellite systems in the overall strategy. Terrestrial mobile communication systems are usually based on cellular principles, where the mobile terminal communicates with a fixed base station. Alternative strategies can, for instance, establish a connection via a network of mobile terminals only (ad-hoc networks); however, the focus here will be on cellular techniques. The fundamental issue in mobile communications is the restricted availability of frequency bands. Therefore, system design must aim at a high spectrum efficiency expressed in Erlang per square meter per Hertz and given by

ηs =

number of reuses number of channels time the channel is busy × × coverage area bandwidth total time of channel

(1-44)

where one busy communication channel is equivalent to the traffic of one Erlang. The basic behavior of cellular designs is often studied in the context of hexagonal cell coverage, as illustrated in Figure 1-24, where the assumption is that the base station is located at the center of a cell. The idea is to reuse a given set of communication channels or frequencies (in those cells) that are sufficiently apart from each other so that co-channel interference remains within acceptable limits. Code division multiple access (CDMA) techniques assign relatively broad frequency bands to individual cells. Those techniques, however, can often work with the same frequency in neighboring cells since channel separation is achieved on a code level and frequency planning may be replaced or supplemented by code planning. Therefore, discuss reuse considerations on the channel level. The group of cells not employing channel reuse is called a cluster. Such a cluster is depicted in Figure 1-24 as a grey-shaded area and consists of seven cells.

f5 f6

f7 f3

f2 f4

f6 f4

f1 f7

f5

f5

f3 f2

f5 f6

f1 f7

FIGURE 1-24 Hexagonal cellular pattern with a reuse factor of 7



1-33

The number of cells N comprising a cluster is, in general, given by N = i 2 + ij + j 2

(1-45)

where i and j are integers, with N = 7 being a popular number of cells per cluster. Obviously, you can increase the mobile system’s capacity (i.e., cellular phone users) by increasing the number of available channels (bandwidth) or the spectrum efficiency hs. Because bandwidth is restricted, the logical approach is to achieve better spectrum efficiency through channel reuse. This implies smaller cell sizes and consequently a greater number of cells in a given coverage area. For example, a single cell can be subdivided into several smaller cells, or a cell can be divided into sectors by using directional antennas to avoid installing additional base stations. Typical cell dimensions used to cover wide rural regions or suburban areas are referred to as macrocells. In densely populated urban regions, however, cells must often be installed at the street level with base stations located below building roofs. Such cells are often called microcells, and even smaller cells (restricted to a single site or the interior of a building) are referred to as picocells. The key to designing complex cellular mobile communication systems is electromagnetic wave propagation within individual cells. Propagation models are essential for evaluating modulation and coding schemes and their associated signal power within the service area or the interference that may be caused outside the service area. Even when adaptive power control is used at the transmitter, a thorough understanding of the relevant wave propagation mechanisms in the terrestrial environment is necessary. For propagation above a smooth ground, a simple ray-optical model including two rays is sufficient for modeling wave propagation. In general, ray approaches are based on high-frequency assumptions, typically fulfilled at mobile communication frequencies (greater than 400 MHz). The two-ray model includes a direct ray representing free-space propagation and a ray reflected at the earth’s surface. Assuming the earth’s surface to be planar, the two-ray model gives the following pathloss: r  λ  L p = 20 log  + 20 log 1 + Rb e − j2π / λ (rb −r ) rb  4π r 

(1-46)

Lp is the ratio of received to transmitted power in dB, where both antennas are assumed to be isotropic. The involved parameters are l, or the free-space wavelength; r, or the path length of the direct ray; rb, or the path length of the reflected ray; Rb, or the plane wave reflection coefficient at the reflection point. Figure 1-25 compares Lp as given to single/direct-ray (free-space) and four-ray models for different transmitter antenna heights. The typical two-ray interference pattern is clearly identified close to the transmitter. Also, at great distances from the transmitter, the two-ray model predicts a pathloss that has a constant slope of 40 dB/decade versus a pathloss with a slope of 20 dB/decade predicted by the single-ray or free-space model. Figure 1-25 also shows that changing the transmitter antenna height can control the coverage range of the base station. Microcells or picocells are often designed at the street level, and for these cases, the tworay model for flat earth can be improved by including ray reflections from the street side walls. For example, Figure 1-25 shows four-ray pathloss curves to model a 16 m–wide street canyon, where contributions from the two sidewall-reflected rays are added to the two-ray results. Again, the usual interference patterns are observed close to the transmitter. However, at great distances from the transmitter, the guiding effect of the street canyon results in a pathloss that is even less than that predicted by the single-ray or free-space model. Also, the height of the transmitter antenna no longer influences the pathloss at great distances.



CHAPTER ONE

−40

Lp

−60 −80

dB

−100 −120 −140 −160 −180 101

Free-space, ht = 30 m 2 rays, ht = 30 m 2 rays, ht = 5 m 4 rays, ht = 30 m 4 rays, ht = 5 m 102

103

104

105

x m

FIGURE 1-25 Pathlosses for different ray models over flat ground. Scalar computations with a receiver antenna height of 2.4 m, different transmitter antenna heights ht, and all reflection coefficients set to –1. The given four-ray models also include sidewall-reflected rays at each side of a 16 m broad street canyon.

Previously, we discussed simple propagation models. However, real-world terrestrial mobile communication channels are characterized by complicated wave propagation mechanisms. The received signal is composed of an extremely large number of different multiply reflected, diffracted, refracted, or scattered wave contributions, and a purely deterministic description of the radio channel might be impossible. Therefore, most receiver power or receiver field strength determination methods are designed to predict the corresponding median values as a first characterization of the channel properties. The fading behavior of the mobile radio channel is described by means of statistical methods where fast- and slowfading mechanisms are distinguished. Fast-fading can often be characterized by Rayleigh or Rice probability distributions, and slow-fading normally behaves as lognormal distributed with standard deviations of several dBs. Further insight into the behavior of a mobile radio channel can be gained from measured or predicted impulse responses, as shown in Figure 1-26, which typically consist of various signal contributions arriving at the receiver after different delay times and with different Doppler shifts due to moving transmitter and receiver antennas as well as scattering objects. Information that can be obtained from the impulse responses are the relative signal powers and delay spreads for the different signal contributions. Both the Doppler shifts and the delay spreads are essential in the design of mobile communication systems. Often, test sequences are included in the transmitted signals, which allow for the estimation of important channel parameters and can be used for channel equalization. Modern terrestrial communication systems often utilize several transmit and/or receive antennas (multiple input multiple output (MIMO)) to achieve improved signal to noise and/ or signal to interference ratios of the communication links. The different antennas can be arranged in closely spaced array configurations suitable for the realization of deterministic beam-forming strategies (see Section 1.8). Another strategy tries to arrange the antennas



FUNDAMENTALS OF ANTENNAS, ARRAYS, AND MOBILE COMMUNICATIONS

P/P0

f0 = 919 MHz

−100 −110 20

40

60 µ sec

f0 = 919 MHz

−110

Dt 20

40

60 µ sec

80

100

Dt 0

−100

0

−100

100

20

40

P/P0

−106 dB

dB

P/P0

80

f0 = 1873 MHz

−110

Dt 0

P/P0

−90 dB

dB

−90

60 µ sec

80

100

f0 = 1873 MHz

−110 −114

Dt 0

20

40

60 µ sec

80

100

FIGURE 1-26 Measured impulse response of different mobile radio channels in Garmisch Partenkirchen, Germany (hilly terrain). The measurement bandwidth was 6 MHz at the given carrier frequency f0. The measured data shown were provided by Deutsche Telekom, Darmstadt, Germany.

such that their transmitting or receiving characteristics become uncorrelated, leading to so-called diversity techniques. Further information on terrestrial mobile communication systems can be found in Chapter 22 on conformal antennas.

REFERENCES 1. C. A. Balanis, Antenna Theory: Analysis and Design, 2nd Ed. (New York: John Wiley & Sons, Inc., 1996). 2. J. D. Kraus, Antennas, 2nd Ed. (New York: McGraw-Hill, 1988). 3. J. A. Kong, Electromagnetic Wave Theory (New York: John Wiley & Sons, Inc., 1990). 4. D. Pozar and D. Schaubert, Microstrip Antennas (Piscataway: IEEE Press, 1995). 5. S. Silver, Microwave Antenna Theory and Design (New York: McGraw-Hill, 1949): sec. 2.14. 6. H. T. Friis, “A Note on a Simple Transmission Formula,” IRE Proc. (May 1946): 254–256. 7. IEEE Standard Test Procedures for Antennas, IEEE Std. 149-1979 (New York: Institute of Electrical and Electronics Engineers, 1979): sec. 11. 8. J. S. Hollis et al., Techniques of Microwave Antenna Measurements (New York: John Wiley & Sons, Inc., 1984). 9. R. C. Spencer and P. M. Austin, “Tables and Methods of Calculation for Line Sources,” MIT Rad. Lab Rep. 762-2 (March 1946); see also Rep. 762-1. 10. J. F. Ramsay, “Fourier Transform in Aerial Theory,” Marconi Rev., vol. 9 (1946): 139; vol. 10 (1947): 17, 41, 81, 157. 11. W. L. Stutzman and G. A. Thiele, Antenna Theory and Design, 2nd Ed. (New York: John Wiley & Sons, Inc., 1998).


Fundamentals of Antennas, Arrays, and Mobile Communications 1-36 12. 13. 14. 15. 16. 17. 18. 19. 20.

CHAPTER ONE

R. W. P. King, The Theory of Linear Antennas (Cambridge: Harvard University Press, 1956). E. Jahnke and F. Emde, Tables of Functions (New York: Dover Publications, Inc., 1943): 227. A. W. Love, Electromagnetic Horn Antennas (New York: IEEE Press, 1976). P. J. B. Clarricoats and G. T. Poulton, “High-Efficiency Microwave Reflector Antennas— A Review,” Proc. IEEE (1977): 1470–1502. Y. T. Lo and S. W. Lee, Antenna Handbook, Volumes I–III. (New York: Van Nostrand Reinhold, 1993). A. W. Rudge, “Offset-Parabolic Reflector Antennas: A Review,” Proc. IEEE 66 (1943): 1592–1618. J. B. Pendry, “Negative Refraction Makes a Perfect Lens,” Phys. Rev. Lett. vol. 85 (October 200): 3966–3969. J. Brown, “The Effect of a Periodic Variation in the Field Intensity Across a Radiating Aperture,” IEE Proc (London) part III, vol. 97 (November 1950): 419–424. J. Ender and H. Wilden, “The Crow’s Nest Antenna—a Spatial Array in Theory and Experiment,” Intern. Conf. on Antennas and Propagation (1981): 25–27.



Chapter 2

Frequency Bands for Military and Commercial Applications Derek M. K. Ah Yo Oceanit

Rudy Emrick Motorola

CONTENTS 2.1 INTRODUCTION TO FREQUENCY BANDS . . . . . . . . . . . . . . . . . . . .

2-2

2.2 FREQUENCIES AND TECHNOLOGIES OF INTEREST FOR MILITARY APPLICATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2-6

2.3 FREQUENCIES AND TECHNOLOGIES OF INTEREST FOR COMMERCIAL APPLICATIONS . . . . . . . . . . . . . . . . . . . . . . . . . .

2-9

2.4 EXAMPLES OF FUNDAMENTAL ANTENNA TYPES USED IN MILITARY AND COMMERCIAL APPLICATIONS. . . . . . . . . 2-15

2-1 Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.

Frequency Bands for Military and Commercial Applications 2-2

CHAPTER TWO

2.1 INTRODUCTION TO FREQUENCY BANDS The spectrum chosen for use in either military or commercial applications often depends on a number of factors, including regulatory requirements or licenses that designate bands for certain uses. Depending on the application, the frequency band selected may also depend on antenna size, propagation distance, or environment. Over time, a number of different designations for frequency bands have been developed. For comparison, each of these band designations is shown in Table 2-1. Historically, some of the band groupings have differed, in part based on the application(s) intended, such as radar, electric countermeasures, and so on. The most commonly used designations are also listed in Table 2-1. As mentioned, one of the factors to consider when selecting a band is the propagation characteristics of the band. Figure 2-1 shows attenuation as a function of wavelength or frequency. Attenuation is also a function of a number of factors that include both scattering losses and absorption losses. In general, lower frequencies tend to propagate farther and to transmit better under non-line-of-sight conditions. However, lower frequency antennas also tend to be larger because antenna size scales with frequency, creating a number of trade-offs that must be considered. For example, the military might select a lower frequency for operation so communication over mountain ranges is optimized; however, the lower frequency creates challenges in implementing an antenna that is a

TABLE 2-1

Frequency-Band Designations

IEEE Radar Bands*

Band

Frequency Range (in GHz)

HF

0.003–0.03

VHF UHF L

1–2

S

2–4

C

4–8

ITU Frequency Bands†

Band


HF

0.003–0.03

0.03–0.3

VHF

0.3–1

UHF

Common-usage Bands‡

Electriccountermeasure Bands§

Band


Band


HF

0.003–0.03

A

0–0.25

0.03–0.3

VHF

0.03–0.3

B

0.25–0.5

0.3–3

UHF

0.3–1

C

0.5–1

SHF

3–30

L

1–2

D

1–2

EHF

30–300

S

2–4

E

2–3

C

4–8

F

3–4 4–6

X

8–12

X

8–12.4

G

Ku

12–18

Ku

12.4–18

H

6–8

K

18–27

K

18–26.5

I

8–10

Ka

27–40

Ka

26.5–40

J

10–20

mm

40–300

Q

33–50

K

20–40

V

50–75

L

40–60

W

75–110

M

60–100

*

From Institute of Electrical and Electronic Engineers Standard 521-1976, Nov. 30, 1976. From International Telecommunications Union, Art. 2, Sec. 11, Geneva, 1959. ‡ No official international standing. § From AFR 55-44 (AR 105–86, OPNAVIST 3430.9B, MCO 3430.1), Oct. 27, 1964. †


Frequency Bands for Military and Commercial Applications FREQUENCY BANDS FOR MILITARY AND COMMERCIAL APPLICATIONS

2-3

FIGURE 2-1 Attenuation of electromagnetic waves as a function of frequency or wavelength (after L. Yujiri et al1 © IEEE 2003)

reasonable size. In other cases, the military might select a band where propagation is limited for security or frequency re-use opportunities. An example of such a band is the V-band around 60 GHz. This band is well suited to short-range applications, but the implementation of low-cost antennas that perform suitably can be difficult due to the high frequency of operation, though the antennas will be relatively small for a given antenna gain compared to the lower frequencies. The trade-offs associated with the use of various frequency bands have been studied extensively over time, and certain characteristics and applications are now commonly linked to specific frequency bands. Table 2-2 lists a number of frequency bands along with their common characteristics and uses. In general, the lower frequency bands tend to be used for longer range, lower bandwidth applications while higher bands tend to be used more often for shorter range applications that require higher bandwidth. In the case of radar, the higher frequency bands are often selected for higher resolution and the antennas take up less volume at these higher frequencies. In the case of television and radio broadcasts, the selected bands tend to be lower, such as VHF and UHF, because the broadcasting company will likely want to cover an entire city or an even greater area with a single transmit site. The specific frequency band used by a particular station also depends on its license through a regulatory agency such as the Federal Communications Commission (FCC), which designates specific channels or bands with well-defined separation between them to avoid interference with broadcast services. The bands and channels used for AM/FM radio and television broadcast in the U.S. are shown in Table 2-3. The separation between the bands, sometimes called guard bands, help prevent interference between adjacent bands, but these bands were established decades ago. With continuing advances in communications technology and equipment, and the move toward digital television, the FCC is looking at different ways of thinking about the bandwidth per channel and the guard bands. In transmitting digital television, less bandwidth is required than



CHAPTER TWO

TABLE 2-2

Common Frequency Bands: Characteristics and Typical Applications

Band

Characteristics

Applications

HF

Long-distance links possible with ionosphere reflections

Ocean vessel communications, telephone, telegraphy, long-range aeronautical communications, amateur radio communications, military communications

VHF

Ionosphere reflections possible at the lower end of the frequency band

Television and FM broadcasting, air traffic control, radio navigation, military communications

UHF

Tends to require more line-ofsight toward the high end of the frequency band

Television broadcasting, radar, mobile phones and radios, satellite communications, global positioning systems (GPSs), wireless LANs, wireless PANs, military communications

SHF

Atmospheric absorption at highest frequencies can be significant

Radar, microwave links, land mobile communications, satellite communications, direct broadcast satellite (DBS) television

EHF

Line-of-sight propagation only toward the high end of the band, subject to atmospheric absorption, and best suited to shorter range applications

Radar, secure and military communications, satellite links, gigabit per second backhaul (1–2 km), future wireless PANs

Optical / IR

Significant atmospheric absorption, typically will not penetrate fog, line-of-sight only

Optical communications, fiber-optical links, very short range wireless communications, building-to-building high-speed wireless connections

with analog, but the bandwidth does depend on the resolution or definition of the signal. In the future, digital television may evolve toward either one very high-definition digital signal or multiple lower definition signals for multiple television channels in what was one signal when transmitting analog. Advances in communications technology have also led to the discussion of the potential use of the guard bands for other applications since advances may allow for their use without interfering with the original broadcast channels. TABLE 2-3

Typical Broadcast Frequencies VHF TELEVISION FREQUENCIES

Band

Ch #

Frequency

Band

Ch #

Frequency

VHF LOW

02

54–60 MHz

VHF HIGH

07

174–180 MHz

VHF LOW

03

60–66 MHz

VHF HIGH

08

180–186 MHz

VHF LOW

04

66–72 MHz

VHF HIGH

09

186–192 MHz

VHF LOW

05

76–82 MHz

VHF HIGH

10

192–198 MHz

VHF LOW

06

82–88 MHz

VHF HIGH

11

198–204 MHz

VHF HIGH

12

204–210 MHz

VHF HIGH

13

210–216 MHz



TABLE 2-3

2-5

(continued) UHF TELEVISION FREQUENCIES

Ch #

Frequency

Ch #

Frequency

Ch #

Frequency

14

470–476 MHz

38

614–620 MHz

62

758–764 MHz

15

476–482 MHz

39

620–626 MHz

63

764–770 MHz

16

482–488 MHz

40

626–632 MHz

64

770–776 MHz

17

488–494 MHz

41

632–638 MHz

65

776–782 MHz

18

494–500 MHz

42

638–644 MHz

66

782–788 MHz

19

500–506 MHz

43

644–650 MHz

67

788–794 MHz

20

506–512 MHz

44

650–656 MHz

68

794–800 MHz

21

512–518 MHz

45

656–662 MHz

69

800–806 MHz

22

518–524 MHz

46

662–668 MHz

70

806–812 MHz

23

524–530 MHz

47

668–674 MHz

71

812–818 MHz

24

530–536 MHz

48

674–680 MHz

72

818–824 MHz

25

536–542 MHz

49

680–686 MHz

73

824–830 MHz

26

542–548 MHz

50

686–692 MHz

74

830–836 MHz

27

548–554 MHz

51

692–698 MHz

75

836–842 MHz

28

554–560 MHz

52

698–704 MHz

76

842–848 MHz

29

560–566 MHz

53

704–710 MHz

77

848–854 MHz

30

566–572 MHz

54

710–716 MHz

78

854–860 MHz

31

572–578 MHz

55

716–722 MHz

79

860–866 MHz

32

578–584 MHz

56

722–728 MHz

80

866–872 MHz

33

584–590 MHz

57

728–734 MHz

81

872–878 MHz

34

590–596 MHz

58

734–740 MHz

82

878–884 MHz

35

596–602 MHz

59

740–746 MHz

83

884–890 MHz

36

602–608 MHz

60

746–752 MHz

37

608–614 MHz

61

752–758 MHz

Broadcast Frequencies AM Radio = 535 kHz–1605 kHz (MF) 107 Channels each with 10 KHz separation TV Band I (Channels 2–6) = 54 MHz–88 MHz (VHF) FM Radio Band II = 88 MHz–108 MHz (VHF) 100 Channels each with 200 KHz separation TV Band III (Channels 7–13) = 174 MHz–216 MHz (VHF) TV Bands IV & V (Channels 14–69) = 470 MHz–806 MHz (UHF)

In addition to commercial broadcast channels, a number of amateur radio bands have been established. These bands also tend to be at lower frequencies since it is usually desirable to have propagation over long distances to communicate with other amateur radio sites, as shown in Table 2-4. The successful use of the higher bands depends on the proximity of other amateur sites because they tend to propagate over shorter distances.



CHAPTER TWO

TABLE 2-4

Amateur Radio Bands

Band

Frequency

160 m

1.8–2.0 MHz

80 m

3.5–4.0 MHz

40 m

7.0–7.3 MHz

20 m

14.0–14.35 MHz

15 m

21.0–21.45 MHz

10 m

28.0–29.7 MHz

6m

50.0–54.0 MHz

2m

144.0–148.0 MHz 220–225 MHz 420–450 MHz 1215–1300 MHz 2300–2450 MHz 3300–3500 MHz 5650–5925 MHz

2.2 FREQUENCIES AND TECHNOLOGIES OF INTEREST FOR MILITARY APPLICATIONS Military and commercial communication needs have similarities and differences, as shown in Figure 2-2 and Figure 2-4, later in the chapter in Section 2.3. One area of commonality is that both stem from the demand for quick and efficient delivery of information. However, the pace at which they develop their communication technologies and their dynamic motivations for doing so are where the two sectors diverge and are continuously evolving. In some cases, the commercial sector’s communication development outpaces the military’s because its driving force is economic and its market is vast. In other instances, military technology outpaces the commercial sector since performance may be more important than low cost for critical equipment. In some ways, the two sectors are merging with the military’s technology, research, and development organizations such as Defense Advanced Research Projects Agency (DARPA), Office of Naval Research (ONR), Naval Research Laboratory (NRL), Air Force Research Laboratory (AFRL), and Space and Naval Warfare (SPAWAR) Systems Command, moving toward leveraging more commercial technology as the military increases its demand for commercial sector innovations to bring its communication systems up to 21st-century standards. The military can optimize its communication systems by complementing available commercial technology with newly developed, customized technology providing a superior hybrid solution. Most terrestrial communication systems used by the military and the commercial market, including HF, VHF, and UHF radios, utilize conventional monopole and dipole antennas that are electrically small but long enough to provide adequate gain and the largest possible coverage area. Most satellite communication (satcom) systems depend on mechanically rotated reflector dish antennas that provide high gain and a narrow beam from a relatively simple parabolic antenna structure in order to illuminate the orbiting satellites with precision accuracy. These antenna structures are, however, still large and can be used as a visual cue in the identification of the communication system’s location. In the case of the military,


Satellite Comm

UHF

Ku-band

Ku-Satcom

Ka-band

243–318 MHz Uplink: 11.2–11.7 Ghz downlink: 14–14.5 GHz Uplink: 27.5, 31 GHz downlink: 18.3, 18.8, 19.7,20.2 GHz

UHF

Ka-Satcom

2.412–2.462 GHz

S-band

1–1.6 cm

2–2.7 cm

.94–1.23 m

12.5 cm

66–71 cm 15–23 cm

20–450 MHz

0.5–10 m

0.5–10 m

1308–1484 MHz 1700–2000 MHz

30–512 MHz

SecNet11 Secure WLAN UHF TacSat

L-band

RT-1720 EPLRS

VRC-99

VHF-UHF

PSC-5D

30–512 MHz

0.5–10 m

3.4–10 m

10–150 m

Free Space λ

FIGURE 2-2 Wireless technologies for military communication systems

Military

Terrestrial Comm

VHF-UHF

PRC-117

30–512 MHz

30–88 MHz

VHF

VHF-UHF

RT-1523

PRC-148

2–30 MHz

HF

PRC-150

Frequency

Frequency Band

Wireless Technology Range

Earth to GEO

Earth to GEO

Earth to LEO

120 m

150 miles w/ relay

6–60 miles

10–50 miles

10–50 miles

12 miles

10–100s miles

30+ miles

Data Rate

Uplink 2 Mbps downlink 30 Mbps

.5–5 Mbps

NA

1–11 Mbps

625 kbps; 10 Mbps bursts

486 kbps

76.8 kbps

NA

NA

9.6–14.4 kbps

9.6–14.4 kbps

April 2005

Late 1970s

NA

NA

NA

2000

1997

NA

2001

1994

2001

Deploy Date

Ka-band Sat Comm System

Ku-band Sat Comm System

UHF Tac Sat Comm

Secure Wireless LAN

Comm Devices/ Operation secure voice/data/ networking SINCGARS single ch gnd&air radio system Voice/low rate data secure voice/data LOS & UHF Satcom secure voice/data LOS & UHF Satcom secure voice/data/ networking Voice/data/ video/ network radio

Compact

Phased Array

Patch

Half-Disc Ant

PIFA

Sperical Helix

Sperical Helix

Sperical Helix

Sperical Helix

Sperical Helix

Electronic Mech Dish Ant Scan Phased Array

Electronic Mech Dish Ant Scan Phased Array

Dish Ant

Mono-Dipole

Mono-Dipole

Mono-Dipole

Mono-Dipole

Mono-Dipole

Mono-Dipole

Mono-Dipole

Mono-Dipole

Traditional

Antenna Technologies

Frequency Bands for Military and Commercial Applications

FREQUENCY BANDS FOR MILITARY AND COMMERCIAL APPLICATIONS

2-7



CHAPTER TWO

inconspicuous communication vehicles are essential to protecting communication capability and the personnel both operating the system and receiving its intelligence support. The commercial sector also benefits from reduced antenna architectures because the savings in real estate allow for additional communication systems to be installed. In the case of cellular infrastructure, reduced size systems can translate into reduced recurring costs because less space may need to be leased at the site’s location. Methods for reducing antenna size include using planar antenna geometries similar to those of microstrip antennas with parasitic shorting structures,2–5 high dielectric substrates,6 geometrical variations of the antenna architecture including slots,7–10 and active element integration into the antenna. Connectivity on today’s battlefield is of such great importance that communication development programs are being forced to develop rapidly. Multifunction communication platforms are being built to satisfy the need for multiple radio access and centralized intelligence distribution using a router methodology that consolidates the information from several different radio sources and redistributes it to personnel using the appropriate communication system. A multitude of processing and communication systems are utilized together to keep all mission participants apprised of tactical and strategic information as it becomes available, allowing command decisions to be handed down as required. The radios used by the military vary in frequency and constitute a range of frequencies that span from the HF to Ka-band, as shown in Figure 2-3. Tactical radios in the HF frequency, including the PRC-150 from 2 MHz to 30 MHz, are important in transmitting voice and limited data over long distances by pumping large amounts of power (20–100 W) through substantially large monopole antennas. These tactical radios also fall within the civilian Ham radio frequencies, which also require large amounts of power to operate. VHF radios operate at slightly higher frequencies from 30–88 MHz, providing for increased bandwidth to transmit more data over the same long distances. This is true for the SINCGARS RT-1523E radio system, which also pumps large amounts of power through a large monopole antenna to extend the range of its transmissions. Radios that operate at UHF, which also include the L-band and S-band from 300 MHz to 3 GHz, provide increased bandwidth capability and faster data rates than HF and VHF systems but at the expense of reduced transmission range. Operating at these higher frequencies also scales down the size of the antennas used to Tx/Rx signals, which reduces the visibility of mobile communication platforms. As presented in Figure 2-2, the PRC-148, PRC-117F, PSC-5D, RT-1720 EPLRS, and VRC-99 radios all operate at these frequencies. The range of the systems are reduced due to the limitations of shorter

SecNet11 Ku-Satcom Ka-Satcom

S-Band

Ku-

Ka-Band

27 GHz 31 GHz

1.7 GHz

1.6 GHz

512 MHz

1.2276 GHz 1.3 GHz

L-Band

318 MHz

292 MHz

270 MHz

88 MHz

UHF

243 MHz

VHF

30 MHz

2 MHz

HF

IRIDIUM

UHFTacSat

1.5 GHz 1.57542GHz

PRC-150 RT-1523E

GPSL2 GPSL1

18 GHz

PSC-5D

11 GHz 14 GHz

VRC-99

PRC-117

2GHz 2.4 GHz

PRC-148

FIGURE 2-3 Military communications frequency spectrum



2-9

wavelength transmissions, but this shortfall is more than compensated for because larger bandwidths are available and long-distance data transfers can be achieved by using relays to leapfrog signals between intermediate points and then on to remote positions. Situational awareness reports that include highly detailed maps with reconnaissance pictures and even video are transmitted by tactical satellite radios operating at the C-, Ku-, and Ka-band. These frequencies are capable of penetrating through the ionosphere to orbiting satellites and then back down to other positions on earth that are over-the-horizon from the position of the originating source transmission. This over-the-horizon capability and increased bandwidth are major goals of the military and also of the commercial market, which means greater amounts of data can be sent more quickly to more distant locations thus providing needed intelligence for informed decision making. This wide array of radio systems over the large frequency range increases military capability, especially when placed on a mobile platform, but at the cost of increased hardware including the radios themselves, networking equipment, power sources, and the antennas to perform the necessary signal propagation. This excessive amount of hardware is expensive and bulky, and in the case of the antennas, presents the problem of physically placing them on top of a small roof, deck, mast, or fuselage. Not only are there co-site signal propagation issues associated with multiple antennas for multiple communication systems, but also there is the burden of concealing the vehicle, ship, or aircraft’s purpose from unwanted spectators. In order to decrease the visibility of the mobile platform as a target and maximize its carrying capacity, there is a need to scale down the size, weight, and volume of all components including the antennas. This requires antenna designs that use planar antenna technologies and allow for dual-band, multiband, and broadband multifunction capability.

2.3 FREQUENCIES AND TECHNOLOGIES OF INTEREST FOR COMMERCIAL APPLICATIONS Commercial frequency bands and application technologies are driven by a number of factors but none more significant than cost. As previously discussed, this factor is more relevant for commercial applications than for military ones. The overall cost of a solution or service is a direct function of equipment cost but can also be affected by the spectrum selected for use. In some cases, licensing spectrums can be very expensive and that cost must eventually be passed on to the consumer by a service provider. One of the advantages of using the licensed spectrum is that there is recourse if another company or individual interferes with the spectrum. The unlicensed spectrum is typically open for use by anyone as long as the conditions and limits set by that country’s regulatory agency, such as the FCC in the U.S., are adhered to, but there is no recourse if another user creates interference when using the spectrum in the same vicinity. The backbone of communication is the allocated frequency bands that radio equipment use to transmit and receive its data, and there are various wireless networks and protocols used for management. Commercial wireless protocols are varied and continue to evolve from the present second-generation (2G) to third-generation (3G) and future fourth-generation (4G) network system protocols that provide expanded bandwidth and increased data-rate capability. The 2G protocols were based on low-band digital data signals that provided a vast improvement over the previous analog systems. The 3G protocols allow wideband digital data transfers for large video and data transmissions. The trend is toward multimedia services, and this trend has made possible podcasting and streaming video that were once tethered to Ethernet cable connections on desktop computers but are now accessible with wireless multimedia phones, personal digital assistants (PDA), and laptops.


Commercial

NMT-450 Rx: 463–468 Tx: 435–458 NMT-900 Rx: 925–960 Tx: 890–915

Rx: 869–894 Tx: 824–849

AMPS/NAMPS Narrow Band Advanced Mobile Phone System FDMA NMT-900 Nordic Mobile Telephone FDMA

Determined by host network

(3G) GPRS General Packet Radio Service

Rx: 810–826 Tx: 940–956 Rx: 1429–1453 Tx: 1477–1501 1920–1980 MHz 2110–2170 MHz Any existing band

UHF - Lband

Rx: 1805–1880 Tx: 1710–1785

Frequency Rx: 869–894 Tx: 824–849 Rx: 1930–1990 Tx: 1850–1910 Rx: 869–894 Tx: 824–849 Rx: 1930–1990 Tx: 1850–1910 Rx: 2110–2170 Tx: 1920–1980 Rx: 925–960 Tx: 880–915 Rx: 1805–1880 Tx: 1710–1785 Rx: 1930–1990 Tx: 1850–1910

CDMA2000 3G CDMA based network

3GSM 3G GSM network

(2G) PCS Personal Comm System TDMA/FDM

(2G) DCS1800/ DCS1900 TDMA/FDM

Frequency Band

31–69 cm

34–36 cm

NA

NA

14–16 cm

20–37 cm

16–18 cm

15–34 cm

48.6 Kbps

Data Rate

100–10,000 m

100–10,000 m

100–10,000 m

100–10,000 m

100–10,000 m

100–10,000 m

100–10,000 m

100–10,000 m

2001

40–50 Kbps; 171 Kbps limit

NA

1986

1978

Feb 2003 Korea

NA

?

2 Mbps

1994

mid 1990s

1987 Europe 1995 USA

1996

1994

Deploy Date

144 Kbps future capability 4.8 Mbps

42 Kbps

270.8 Kbps

270.8 Kbps

100–10,000 m 1.2288 Mbps

100–10,000 m

15–36 cm

14–36 cm

Range

Free Space λ

NMT enabled cellular phones

Mono-Dipole

Mono-Dipole

Mono-Dipole

GPRS enabled cell phones/ networks Interface overlaid on existing GSM networks allowing for internet access AMPS/NAMPS enabled cellular phones

Mono-Dipole

Mono-Dipole

Mono-Dipole

Mono-Dipole

Mono-Dipole

Mono-Dipole

Mono-Dipole

Patch Variant

Patch Variant

Patch Variant

Patch Variant

Patch Variant

Patch Variant

Patch Variant

Patch Variant

Patch Variant

Patch Variant

Antenna Technologies Traditional Compact

Mobile phones on CDMA2000 networks

3GSM-enabled cell phones, PDAs, pagers

PCS enabled cell phones

DCS enabled cell phones, PDAs, pagers

GSM enabled cell phones, PDAs, pagers

Mobile phones on CDMA networks

TDMA cellular phones

Comm Devices/Operation

2-10

Analog Wireless Protocols

Digital Wireless Protocols

(2G) GSM Global System for Mobile Comm TDMA/FDM

(2G)CDMA IS-95 Code Division Multiple Access FDM

Wireless Technology (2G) TDMA IS-54/IS-136 Time Division Multiple Access FDM

Frequency Bands for Military and Commercial Applications CHAPTER TWO


Satellite Comm

C-band

S-band

802.11a 802.11b Wi-Fi 802.11g 802.11n

C-band

C-band Satcom

Ka-band

L-band

Iridium

Ka-band Satcom

C-band W-band

(4G) Broadway HIPERLAN/2 HIPERSPOT OFDM

Ku-band

S-band C-band

(4G) 802.16 WiMax OFDM FDD/TDD

Ku-band Satcom

ISM band Industrial scientific and medical

802.15.4 ZigBee

802.15.1 Bluetooth

VHF UHF

VHF TV UHF TV

Uplink 5.925– 6.425 GHz downlink: 3.7– 4.2 GHz Uplink: 11.2– 11.7 Ghz downlink: 14–14.5 GHz Uplink: 27.5, 31 GHz downlink: 18.3, 18.8, 19.7,20.2 GHz

1.616–1.628 GHz

1–1.6 cm

2–2.7 cm

4.7–8.1 cm

18 cm

0.5 cm 6 cm

5–12 cm

2.5–2.69 GHz, 2.7–2.9 GHz, 3.4–3.6 GHz, 5.725–5.86 GHz 5 GHz 59–65 GHz

6 cm, 33 cm, 35 cm

868 MHz, 915 MHz, 2.4 GHz

12.5 cm

12.5 cm 12.5 cm

2.4 GHz

12.5 cm

2.4 GHz 2.4 GHz

6 cm

1.4–7 m 37–64 cm

32–35 cm

2.4 GHz

5 GHz

NTACS: Rx: 860–870 Tx: 915–925 ETACS Rx: 916–949 Tx: 871–904 44–216 MHz 470–806 MHz

Earth to GEO

Earth to GEO

Earth to GEO

Earth to LEO

10–100 m

1000–5000 m

< 50 m

100kW, 3 GHz) frequencies—the importance of low and medium frequencies has not diminished, because they offer unique properties such as very stable propagation conditions and the ability to penetrate the sea and earth. A fundamental difficulty in achieving efficient radiation at low frequencies is the necessity of having antenna dimensions comparable to radiation wavelength (l) in the air. Restricting the antenna geometry to a vertical monopole of height h, the radiation efficiency drops with the ratio (h / l)2 as the wavelength increases. Maximum efficiency is obtained if the antenna height is h = l /4, also providing that a good match to a low-impedance transmission line (50-Ω) is achieved. For medium frequencies (MF) at 600 kHz at the radiation wavelength l = 500 m, maximum efficiency is achieved if h = 125 m, which doesn’t pose a serious construction problem. However, at the low edge of low frequencies with a corresponding height of much more than 500 m, you have to restrict the antenna height to less than 400 m. The basic environmental condition restricting the antenna height is the wind load. In particular, wind loads exhibiting no homogeneity along the vertical axis could pose serious challenges to the mechanical stability of the mast. Earthquake-induced forces are not significant. Although the possible radiation mechanism at low frequencies is not restricted only to a vertical monopole, as will be discussed later in Section 27.5, vertical monopole antennas predominately are used in constructing very-low- (10–30-kHz) and lowfrequency (30–300-kHz) radiation systems. Another unique property of low-frequency antennas is related to the insignificance of the receiving antenna’s dimensions. This is because at low frequencies atmospheric noise exceeds the receiver internal noise, and therefore the receiver output signal-to-noise power ratio is independent of antenna efficiency and size.

27.2 FUNDAMENTALS OF VERTICAL MONOPOLE ANTENNA RADIATION PROPERTIES The radiation properties of a vertical monopole antenna, shown in Figure 27-1 and having height h, are easily computed by using the image theory, assuming the earth surface is a perfect conductor. For an earth medium with good conductivity s > 0.01 S/m (Siemens/meter),


Low- and Medium-Frequency Antennas 27-5

LOW- AND MEDIUM-FREQUENCY ANTENNAS

r

A

θ

h

ψ

Perfect Conductor Earth Surface

FIGURE 27-1 Vertical monopole antenna geometry

in practical terms, the ground surface can be assumed to have perfect conductivity. Usually wire grids (see “VLF and LF Antennas” in Section 27.3) are used to achieve good conductivity in the vicinity of the antenna. Therefore, to understand the fundamental radiation properties of low-frequency vertical monopole antennas, the earth surface could be assumed to be a perfect conductor. Then, based on fundamental concepts of antenna theory, the total electric current distribution on the vertical antenna tower is almost equal to a sinusoidal function as I ( z ) ≈ I 0 sin( k0 ( h − z )) /sin( k0 h )

(27-1)

where z is the distance measured from the base of the monopole, I0 is the base current, and k0 = 2p / l is the free space propagation constant. At very low frequencies (VLF), h  λ, Eq. 27-1 reduces to a linear relationship, I(z) ≈ I0 k0 (h − z)/h. The antenna characteristics are defined in terms of its radiation resistance Rr = 160p2 (he / l)2

(27-2)

where he is the equivalent antenna height computed by the equation he =

1 I0

h

∫0 dzI ( z)

(27-3)

The antenna radiation efficiency is computed as the ratio of the radiated power to total power fed to the antenna:

η=

Rr Rr + Rl

(27-4)

where Rl is the total loss antenna resistance representing the non-radiated power consumed by the antenna structure and the ground medium in the vicinity of the monopole. The radiation pattern that is identical to a Hertzian dipole is defined in terms of radiated power density (W / m2) in free space: P(θ ) = P(π / 2 )sin 2 (θ )

(27-5)



CHAPTER TWENTY-SEVEN

where the radiation power density at the horizontal plane is P(π / 2 ) =

22.75k02 he2 I 02 r2

where r  λ is the distance of field measurement point and q is the angle between the vertical axis and the position vector. From the early period of radio engineering (1910–1920), it was understood that the efficiency of a monopole antenna can be increased significantly by making the antenna current as constant as possible along the vertical axis. This is achieved by loading the antenna top with various conductor structures, as shown in Figure 27-2. Usually, a symmetric radial network of wire conductors as shown in Figure 27-2c is used. The loading increases the equivalent antenna height and therefore the radiation efficiency. Top-loading increases the equivalent antenna height, he, and thus the radiation resistance and radiation efficiency. The antenna loading is accounted for by the existing capacitance, C0, between the top-loading and ground. The equivalent antenna height, he, can be computed by using Eq. 27-3 and the current distribution: (27-6)

I ( z ) ≈ I 0 sin( k0 ( h + s − z ))/sin( k0 ( h + s ))

where s is the slant length of the top-loading wires. In modeling the low-frequency antenna, the equivalent circuit shown in Figure 27-3 is used for the input impedance of the monopole. Based on the equivalent circuit, the input impedance of the antenna can be computed as follows: Z in = RT (ω ) + jω L − 1 / jω C = RT (ω ) + jXT (ω )

(27-7)

where RT = Rr + Rl is the total antenna base resistance, L is the antenna mast inductance, and C = Ca + C0 is the total (mast plus the top-loading wires) antenna capacitance. The reactance (imaginary) part of the impedance can be written in terms of the antenna resonance frequency

ω 0 = 1 / LC

(27-8)

XT (ω ) = ω L (1 − (ω 0 / ω )2 )

(27-9)

as follows:

Insulator

Insulators

Wire h

(a) Inverted L

Insulators h

(b) T - Top Loading

h

(c) Radial Wire Top Loading

FIGURE 27-2 Three types of top-loading monopole antennas: (a) inverted L, (b) T top-loading, and (c) radial wire top-loading



LOW- AND MEDIUM-FREQUENCY ANTENNAS

L

Antenna Input

Rr

Rl

Cα

∆C

Co

FIGURE 27-3 Monopole antenna equivalent circuit: L = antenna inductance, C0 = top-loading capacitance, Ca = antenna mast capacitance, Rr = radiation resistance, Rl = total loss resistance, and ∆C = antenna feedpoint leak capacitance

Usually at VLF (3–30 kHz) the operation frequency is much less than the resonance frequency. A critical parameter, especially at frequencies less than 100 kHz, is the antenna bandwidth determining the available communication channel capacity or the maximum data bit rate of transmission.3 The intrinsic 3-dB antenna bandwidth is computed3 using the definition of the quality (Q) parameter of the equivalent circuit of Figure 27-3: BW0 = f / Q = 1.102458 * 10 −7 Che2 f 4

(27-10)

where f = w / (2p) is the operation frequency. The overall bandwidth of the system is determined taking into account the antenna losses and is BWl = BW0 / η

(27-11)

Above 150 kHz the antenna bandwidth restriction is limited, and traditionally, monopole antennas are used to broadcast double-side amplitude modulation signals. The most common broadcast antenna type used in the frequency region of VLF, LF, and MF is the vertical monopole. However, in theory other types of antenna, especially vertical loop antennas, have been suggested to cope with the difficulty of constructing very tall masts needed at the low edge of the spectrum. This topic is analyzed in Section 27.4.

27.3 PRACTICAL LOW- AND MEDIUM-ANTENNA STRUCTURES: DESIGN AND CONSTRUCTION BASICS Design and construction of low-frequency antennas poses formidable civil engineering tasks primarily having to do with the erection of the metallic tower structures, which could be as high as 400 m. Concerning the electromagnetic properties of the antenna, issues to be addressed are ●

●

●

The conductivity properties of the ground environment where the antenna is to be installed and, in case of poor conductivity (1000 km) because the poor radiation pattern at higher take-off angles limits the signal strength received. In Figure 27-12 the realization of a horizontal array is shown where two broadside arrays parallel to earth surface dipoles are used. No surfaceground wave is induced, and only sky mode is emitted. The take-off angle is a function of the height above the ground of the dipole array at a given frequency. Antenna Tuning In all VLF and LF antennas, the antenna height being much smaller than the wavelength of the antenna input presents an impedance Zin = R + jX with few ohms of the real part and a capacitive reactance up to a few thousand ohms. Taking into account the modern solid-state power amplifiers of 50-Ω output impedance, you need to use matching circuits. In Figure 27-13 a common antenna impedance matching circuit is shown. In all cases a mechanical switch should be placed at the antenna input to protect maintenance personnel during service operations with the antenna system to prevent possible high voltages induced on the tall antenna structure.

Variable Inductance Coil

To Antenna Input

Low Pass Filter

Transformer

Protection Switch

FIGURE 27-13 VLF / LF antenna input impedance matching circuit


Low- and Medium-Frequency Antennas LOW- AND MEDIUM-FREQUENCY ANTENNAS

27-19

As shown in Figure 27-13, a fixed and variable inductive serial network element is used to compensate the highly capacitive load of the antenna. A step-up transformer using low-loss core material is used to increase the low resistance impedance to a higher value near the commonly used 50-Ω transmission line wave impedance connecting the power amplifier with the antenna system. Considering that the distance of the antenna to a 10–100-kW amplifier could be up to 500 m, a matching of better than –20-dB return loss is very important. This requires a coarse- and fine-tuning variation of the matching circuit inductance. A common practice to achieve fine variation of the serial inductance element is to use an air core variable transformer, known as variometer, with a shortcircuited secondary element. By rotating the secondary element and varying the coupling coefficient of the transformer, a fine change of the inductance is obtained. An additional tuner circuit, such as shown in Figure 27-13, is used between the transmission line and the transformer unit to achieve proper impedance matching of the whole system. To achieve an overall low loss, the antenna tuning circuit elements should be designed and constructed of large-diameter hollow copper wires or even pipes. An important issue is the protection of the antenna and especially of the power transformer– power amplifier unit from direct lightning strikes on the antenna mast. Gas discharge spark gap protection could be used in regions with high thunderstorm activity. The Tee or Pi-type LC matching networks usually are used in tuning circuits for VLF/LF and MF antennas. In case of power exceeding 50–100 kW, gas-filled capacitors are used to provide protection from induced arcs. All the tuning circuit network elements should be installed into a shielded room that is grounded to the earth and grounding conductor grid. Environmental Considerations During the last decades the scientific community and the public have paid increasing attention to possible negative health effects of electromagnetic radiation. Regarding low frequencies and in particular because of the extensive use of industrial electrical energy of 50/60-Hz extremely low frequencies (ELF), there has been extensive study and argument about negative health effects primarily attributed to the magnetic field component of ELF fields. Considering that VLF, LF, and MF fields are quite close to 50/60 Hz, the issue of possible health effects near broadcasting antenna fields are of concern. Various international and government bodies during the last decades have regulated the exposure of the general public and employers to electromagnetic radiation. Presently the most stringent standards are those of European Commission Directive 519-12/7/1999, which sets reference levels concerning the rms electric field strength of 87 V/m and magnetic field strength of 5 A/m. Considering that the vertical monopole top-loaded antenna is the most used antenna and because of the very high capacitive input impedance in the vicinity of the antenna, under the top-loading structure the electric field could have significant values while the magnetic field is comparatively very weak. Therefore, care should be taken to check the electric field near the antenna. Both theoretical and experimental results show that outside of the top-loading region the antenna electric field doesn’t violate the 87-V/m standard.

27.4 NUMERICAL COMPUTATIONS OF LOW-FREQUENCY ANTENNAS The most suitable numerical computation method of analyzing and designing lowfrequency ntennas is the method of moments (MoM) using a Pocklington integral equation. The analysis of a finite length thin cylindrical dipole antennas has been a fundamental research topic in the history of applied electromagnetism.




In applying the method of moments the radiation problem is formulated by modeling the active parts of the conductor structure by linear overlapping cylindrical segments being smaller than l/20. If the number of linear segments is N, then each segment is characterized with its length li, (i = 1, 2, ...N) and radius ai, (i = 1, 2, ...N). For mast structures consisting of multisegment tubular cage structures, the radius ai is computed by Eq. 27-24, while for the top-loading or other wires the actual radii are used. Then the integral equation, taking into account the validity of the condition ai  l and ai  li, is written as follows: l    2 ∂2  N i j  k0 + ∂s 2  ∑ ∫ ds ' K (rj , ri ') Ii ( s ') = − Es 0 ( s ) i =1 s ' = 0

(27-29)

where the kernel function is the Green’s function of the infinite dimension and conductivity of the earth surface averaged over the wire conductor periphery. If the earth medium is assumed to be homogenous and of finite conductivity, the Green’s function still is available by taking into account the Sommerfeld integral as an additional term. The s and s' variables represent the observation and source points on the wire structures, respectively. The right-hand side represents the imposed electric fields on the antenna structures and more particularly the applied voltage on the antenna driving points. To determine the unknown current distributions, following the method of moments (MoM) developed during the 1960s–1990s, a set of linear “physical” basis functions are selected to describe the current distribution in terms of a linear sum with unknown weighting coefficients. Then a linear system of equation is obtained after an inner product by a set of “testing” functions of this integral equation of the equation on each segment. In case both “describing” and “testing” belong to the same family, a highly numerically stable procedure is developed, providing accuracy exceeding 1 percent, which is known as the Galerkin method. The selection of physical describing and testing functions require that the engineering inside be pertinent to the problem. An important issue is the stability of the Galerkin method because of the satisfaction of the variational condition, and the accuracy is valid for global quantities such as the radiation pattern and input impedance. Various commercial codes have been developed and are available based on the MoM.

27.5 VERTICAL MAGNETIC LOOP AND HORIZONTAL SLOT ELF AND VLF ANTENNAS USING NATURAL STRUCTURES Traditional antenna engineering at the low-frequency spectrum mostly uses vertical monopole antennas even as low as 16-kHz radiation frequencies. The limitation imposed by tower heights of 400–500 m makes it rather difficult to construct antennas radiating less than 10 kHz since the achievable efficiencies are very small and the required power levels are excessively high. During the last 50 years various alternatives to vertical monopole antennas have been tried experimentally. Considering all possible small electrical antennas on the perfect conductivity ground surface and derived from electrical and magnetic dipoles and the complementary slot antennas, as illustrated in Figure 27-14, we can state the following radiation modalities: a. Vertical monopole, with its image doubling the antenna electrical length, which has been the basis of low-frequency antenna technology the last 100 years b. Horizontal magnetic loop or electric dipole antennas, both placed near the earth surface and both being non-radiating structures because the image currents are opposite to the antenna currents c. Vertical magnetic loop antenna, which excites an image loop antenna inside the earth, enhancing the primary source d. Horizontal slot antenna, which, as in the preceding case (c), could provide radiating structure Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.

Low- and Medium-Frequency Antennas LOW- AND MEDIUM-FREQUENCY ANTENNAS

+ − Vertical loop

27-21

+ − Horizontal loop (b)

(a)

+ Sea Land

+

Mountain − Vertical loop Antenna (c)

FIGURE 27-14

Sea

Peninsula

(d)

Possible radiating mechanisms at very long wavelengths

Review of all possible radiating mechanisms apart from the traditional vertical monopole antenna, already analyzed in previous sections, and considering the possibility of building such antennas based on the previously mentioned principles (c) and (d) providing radiation capabilities, gives the following possibilities: ●

●

Vertical magnetic loop, which could be constructed by placing a conductor on a mountain or an island and grounding the ends of the wire. The loop is excited by a serial generator, as shown in Figure 27-14a. Horizontal slot antenna, which could be constructed using an island or a peninsula and exiting by a shunt voltage source, as shown in Figure 27-14b.

Both possibilities have been investigated by several researchers during the past 50 years. The best-known case has been the U.S. Navy Extremely Low Frequency (ELF) project, which was proposed in the 1950s by Christofilos, and aimed to emit as low as 70–80 Hz. The vertical magnetic loop was obtained by running parallel wires and grounding them deep inside the earth, which had low conductivity earth in the specific used area. The possibility of utilizing a tunnel and circling a wire around the above mountain has been investigated in New Zealand, and interesting results have been demonstrated in the frequency region of 1–10 kHz providing a proof of concept of a vertical loop antenna. The few reported measurements using vertical magnetic loop antennas confirm the viability of developing broadcasting antennas in the 1–10 kHz frequency region with sufficient efficiency, provided loops with effective area of 1–5 km2 are used. The possibility of developing a horizontal slot antenna using rocky islands and peninsulas also has been investigated by several researchers measuring the input impedance of the wire located in the throat of a narrow island. Measurements of radiated field strength have been carried out on a rocky, narrow (average width 200 m) peninsula of 2.5-km length in the frequency region of 3–10 KHz. Although, based on visual intuition, the possibility of achieving a large size slot radiator is attractive, in practice this is not the case since underground structures penetrated by seawater effectively short circuits the slot radiator and decreases the antenna efficiency to a very small value. Indeed the impossibility of operating the slot mode ELF antenna was predicted in the early 1960s.10 Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.



REFERENCES 1. H. Hertz, Electric Waves: Being Researches on the Propagation of Electric Action with Finite Velocity Through Space, Translated by D. E. Jones (London and New York: Macmillan & Co., 1893) and (New York: Dover Publications, 1962). 2. K. A. Norton, Nature-London, vol. 135 (1935): 954 and Proc. Institute of Radio Engineers, vol. 29 (1941): 623. 3. A. D. Watt, VLF Radio Engineering (Oxford: Pergamon Press, 1967). 4. H. R. Bhojwani and L. W. Zelby, “Spiral Top-Loaded Antenna: Characteristics and Design,” IEEE Trans. Antennas Propagat., vol. AP-21 (1973): 293. 5. C. E. Smith and E. R. Gnof, “Increased Capacitance for VLF Umbrella Antennas Using Multiple-Wire Rib Construction,” IEEE Trans. Antennas Propagat. (November 1968): 766–767. 6. H. A. Wheeler, “Fundamental Relations in the Design of a VLF Transmitting Antenna,” IRE Trans. Antennas Propagat., vol. AP-6 (1958): 120–122. 7. C. Balanis, Antenna Theory Analysis and Design, 2nd Ed., (New York: John Wiley and Sons, 1997). 8. J. R. Wait, “On the Calculation of Transverse Current Loss in Buried Wire Ground Systems,” Apl. Sc. Res., vols. B6: 259–275 and B7: 355–360. 9. G. Markov, Antennas (English Translation), (Moscow: Progress Publishers, 1965). 10. H. Staras, “Analysis of a Natural VLF Slot Antenna,” Electromagnetic Theory and Antennas, E. C Jordan (ed.) (Oxford: Pergamon Press, 1963).

BIBLIOGRAPHY “Millett G. Morgan, In Memoriam,” Radio Science Bulletin—URSI, no. 3000 (March 2002). Barr R. and W. Ireland, “Low-Frequency Input Impedance of a Very Large Loop Antenna with a Mountain Core,” IEE Proc.-H, vol. 140, no. 2 (1993). Barr, R. “A New Interpretation of the VLF Impedance Measurements of Island Slot Antennas,” Radio Science, vol. 15, no. 5 (1980): 959–964. Burton, R. W., R. W. P. King, and T. T. Wu, “The Loop Antenna with a Cylindrical Core: Theory and Experiment,” IEEE Trans. Antennas Propagat., vol. AP-31, no. 2 (March 1983). Gould, R. N., “Some Preliminary Experimental Tests of a Novel Method of Radiating at Very Low Frequencies,” Nature, No. 42773 (April 1961): 332–333. Hansen, R. C. “Slot Antenna in a Resistive Screen,” IEEE Trans. Antennas Propagat., vol. 46 (July 1998): 1028–1031. King, R. W. P., M. Owens, and T. T. Wu, Lateral Electromagnetic Waves (New York: Springer-Verlag N.Y. Inc., 1992). Kraus J. D., Antennas, Chap.13 (New York: McGraw-Hill, 1988). Morgan, M. G., “An Island as a Natural Very-Low Frequency Transmitting Antenna,” IRE Tran. PGAP, vol. AP-8 (1960): 528–530. Morgan, M. G., “Comment on a New Interpretation of the VLF Impedance Measurements of Island Slot Antennas by R. Barr,” Radio Science, vol. 15 (1980): 965–967. Uzunoglu, N. and S. Kouridakis, “Radiation of Very Low and Extremely Low Frequencies (VLF & ELF) by a Natural Antenna Based on Island or a Peninsula Structure,” Radio Science Bulletin, no. 308 (March 2004): 7–12. Watt, A. D. “VLF Radio Engineering” (Oxford: Pergamon Press, 1967): 94.



Chapter 28

HF Antennas Brian S. Collins BSC Associates Ltd. and Queen Mary,University of London

CONTENTS 28.1 INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

28-2

28.2 SPECIFYING THE RIGHT ANTENNA FOR THE JOB. . . . . . . . . . . .

28-2

28.3 ANTENNAS MOUNTED ABOVE GROUND . . . . . . . . . . . . . . . . . .

28-4

28.4 MAJOR FORMS OF HF COMMUNICATIONS ANTENNA . . . . . . .

28-7

28.5 OTHER TYPES OF HF COMMUNICATIONS ANTENNAS . . . . . . . 28-12 28.6 HF RECEIVING ANTENNAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-14 28.7 BROADBAND DIPOLE CURTAIN ARRAYS . . . . . . . . . . . . . . . . . . . 28-16 28.8 SITING HF ANTENNAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-20


HF Antennas 28-2

CHAPTER TWENTY-EIGHT

28.1 INTRODUCTION The high-frequency (HF) band (2–30 MHz) is widely used for communications and broadcasting. Long-distance communication relies on (virtual) reflection from the ionosphere, a region of ionized particles created by the interaction of the solar wind of high-energy charged particles with upper layers of the atmosphere. The ionosphere is a variable and turbulent medium, and successful communication relies on matching the frequency of transmission and antenna characteristics to its behavior. There is a substantial difference in practice between public broadcasting operations and point-to-point or point-to-multipoint operations. Broadcast transmissions intended for public reception operate at high power levels to deliver strong signals to distant receivers equipped with simple low-gain omnidirectional antennas. To ensure good audibility, high transmitter powers are used, and these are associated with high-gain directional transmitting antennas whose function is to optimize the signal strength laid down in the target area as well as to conserve spectrum resources by avoiding laying down signals outside the target area. It is common to transmit the same program material simultaneously on more than one frequency band, and the chosen bands may change over the course of a day to match changing propagation conditions. A typical broadcasting station is equipped with a suite of high-gain antennas firing in the directions of interest, often with facilities to modify the beam directions at will in azimuth (and sometimes elevation) direction. Major stations may also be equipped with a rotating high-gain antenna. All the antennas will be of broadband design, allowing them to be used on a number of different broadcast frequency bands, sometimes simultaneously. International broadcasting is a high-cost activity, with large investments in both transmitters and high-gain antenna arrays and with the high energy costs of operating transmitters with output powers up to 500 kW. The advent of digital transmission permits the reduction of transmitter powers, but this technical possibility will only be secured by international agreement as well as by a change in perception that the station with the strongest signal wins the listeners. Professional point-to-multipoint services make use of cooperation by the connected stations, and the advent of channel sounders and automatic link establishment (ALE) facilities has led to the availability of reliable communications with relatively low transmitted powers. The advent of digital modulation schemes, adaptive modems, and ALE has provided more reliable communications with increased data throughput, greatly increasing the potential applications of the HF band. Ionospheric Radio1 and the High-Frequency Radio Automatic Link Establishment (ALE) Application Handbook2 serve as excellent introductions to the subject of modern HF communications. The amateur radio service has used the HF band for many decades and has proved fertile ground for experimentation with antennas, both of conventional and unorthodox design, especially in the areas of compact antennas and multiband antennas with low to medium gain. The distinguishing features of HF antennas are their form of construction (usually from wires), their electrical proximity to ground (which determines many of their radiation characteristics), and the way in which antenna performance must be married to a knowledge of the ionospheric path that will support the intended transmission.

28.2 SPECIFYING THE RIGHT ANTENNA FOR THE JOB The ionosphere has an effective refractive index that depends on height, and an incident electromagnetic wave is progressively refracted. Depending on the ionization density, frequency, and angle of incidence, the signal may pass through the reflecting layer or may be


HF Antennas HF ANTENNAS

28-3

diffracted through a sufficient angle that it returns to earth. If the incident and emerging ray paths are projected back toward the ionosphere, we can define a “virtual height” that we can regard as being the height of an imperfect reflecting mirror. The ionization density and refractive index are a complex function of time of day, season of year, geographical position, solar activity, and frequency. Four layers are of importance: The D-layer (the lowest) absorbs transmissions at MF frequencies during the day, but is transparent when its density falls at night. The E, F1, and F2 layers are important for HF transmission and are further discussed below. The HF band has unique transmission properties and continues to be intensively used. The proper selection of antennas is an important means by which spectrum utilization is maximized. The use of directional transmitting antennas not only increases the signal-to-noise ratio on the wanted path, but it also reduces the signal strength laid down in regions where the signal is not needed and may interfere with other users; directional receiving antennas reduce the extent of interference suffered from other users. The first step in choosing an antenna is to determine the optimum working frequency (OWF) for the ionospheric path involved. Computer programs are readily available that estimate path losses at different frequencies and the reliability of service for the chosen transmitter power and specified antennas at both ends of the link.3,4 The antennas must support the whole range of frequencies that may be needed at any time of day, season of the year, and phase of the 11-year cycle of sunspot activity. The same computer program will identify the probable active ionospheric propagation modes and indicate the corresponding elevation takeoff angles (TOAs) that must be supported by the antennas. Figure 28-1 shows the path geometry for a one-hop path. The height of the E layer is about 100 km. The F-layer is more complex in structure and variable in height, particularly with respect to the local solar zenith angle. First-order estimations can be made on the rough assumption that the F1 and F2 layers have virtual heights of approximately 250 km and 400 km. Multiple-hop propagation can be assessed approximately by using Figure 28-1 for each hop. For distances of more than 4000 km, maximum signal results from low-angle transmission in the range from 2 to 15°, with the lower angles generally providing the highest and most stable received power. Propagation prediction programs3,4 should be used to establish a link budget and to define the antenna parameters. Paths of up to about 200 km require high takeoff angles, giving rise to the term near vertical incidence skywave (NVIS) paths. The different takeoff angles associated with various path lengths require that we use antennas with different elevation pattern characteristics to serve these paths.

FIGURE 28-1 Takeoff angles for one-hop ionospheric transmission as a function of virtual layer height. The example shows that the TOA for a 2000-km path with a virtual layer height of 300 km is around 12°.


HF Antennas 28-4


The minimum azimuth beamwidth needed for reliable point-to-point coverage depends on the effects of irregularities and magnetic storms in the ionosphere. Azimuth directions of arrival may typically vary by ±5° because of these effects.5 This makes the use of antennas with extremely narrow horizontal beamwidths undesirable, especially on circuits that skirt or traverse the auroral zone where much higher angles of off-axis propagation are experienced. For broadcasting or point-to-multipoint services, the beamwidth must cover the target area with allowance for these path-deviation effects.6 High antenna gain is required for transmission when large effective radiated power is needed to overcome ionospheric transmission losses. For receiving, high directivity is desirable in the presence of interference, but low radiation efficiencies can often be accepted because in many circumstances the sensitivity of a receiving system is limited by external noise and not by the noise of the receiving system—this is explored later in Section 28.6. For transmitting antennas, the acceptable input VSWR is usually determined by the characteristics of the transmitter. Modern solid-state transmitters will operate at full output power into a maximum VSWR of around 2:1; above this a transmitter may automatically reduce its output power to limit the voltages and currents in internal circuits. The power rating of an antenna is determined by both average and peak envelope power (PEP). The average power determines the required current-carrying capacity of the antenna conductors, while the peak power determines the peak electric fields on the antenna, which set requirements on insulator and conductor configurations to avoid arcing or corona discharge, especially in wet weather or at locations where insulators may become contaminated with salt or pollution. Mechanical requirements are usually determined by wind speed, ice loading, operating temperatures, and corrosive environments. Some applications demand limited antenna size or height, or the whole antenna system may need to be portable for military or emergency relief operations. The tensions in the components of an HF antenna are usually computed using a standard nonlinear finite-element software package. Care must be taken in design to ensure that all joints are fully articulated to avoid failure by the workhardening of conductors. Insulating synthetic ropes are widely used to support radiating elements and transmission lines, but primary ceramic insulators (typically made from alumina, which has a low dielectric loss factor) are necessary at higher powers to avoid damage by corona discharge. Many HF antennas are of electrically balanced design and are fed from a coaxial cable through a balun in the form of a transmission line device or a ferrite-cored transformer. It has become common practice to load HF antennas with resistive or reactive loads; this technique increases the impedance bandwidth but reduces the available gain. The user should always check the power gain of loaded antennas when they are to be used for transmission.

28.3 ANTENNAS MOUNTED ABOVE GROUND Effect of the Ground HF antennas are usually operated less than two wavelengths above the ground, whose proximity modifies both the input impedance and radiation patterns. Although accurate evaluation of ground effects is difficult, many properties of HF antennas can be understood by regarding them as being situated over an infinite, flat ground plane. If the ground plane is perfectly conducting, its effect is to create an image of the antenna (see Figure 28-2). For ground of finite conductivity, simple image theory does not give an exact solution, but will often provide useful approximate results.


HF Antennas 28-5

HF ANTENNAS

Direct wave E

θ

Element

θ

θ

For perfect ground, the elevation pattern may be computed by assuming an image radiating in antiphase with the source element.

Image FIGURE 28-2

Reflected wave rE

Direct and ground-reflected rays from an element mounted above ground

Radiation Patterns of Antennas Above Ground To compute the signal from an antenna at an elevation angle q ο, we must add the ground-reflected wave to the direct wave, allowing for the amplitude and phase of the reflection coefficient of the ground. The proximity of the ground modifies the currents in the elements of an array placed close to it; so although an approximation can be obtained by this method, a full solution for arrays lower than about one wavelength requires the manipulation of the matrix of all the currents in the array including the effect of the ground. The reflection coefficient for waves with horizontal polarization rh is given by

ρh =

sin θ − (ε r − jx ) − cos 2 θ sin θ + (ε r − jx ) − cos 2 θ

(28-1)

and the complex reflection coefficient for waves with vertical polarization rv is given by

ρv =

(ε r − jx)sin θ − (ε r − jx) − cos 2 θ (ε r − jx)sin θ + (ε r − jx) − cos 2 θ

(28-2)

where x = 18 × 10 3 σ / fMHz

(28-3)

q = grazing angle e = relative permittivity of the ground s = conductivity of the ground Siemens/meter (S/m) fMHz = operating frequency in MHz The ground reflection coefficient for horizontally polarized waves is a well-behaved and slowly varying function. For vertically polarized waves both the magnitude and phase of the reflection coefficient vary with the grazing angle (q ) and the frequency. The angle


HF Antennas CHAPTER TWENTY-EIGHT

1

1

0.8

0.8

0.6

0.6 Mod ρv

Modulus of ρh

28-6

0.4

0.4 Mod R, 2MHz Mod R, 4MHz Mod R, 8MHz Mod R, 16MHz Mod R, 32MHz

0.2

0.2

0 0

20

40

60

Mod R, 2MHz Mod R, 4MHz Mod R, 8MHz Mod R, 16MHz Mod R, 32MHz

0

80

0

20

40

60

80

Grazing Angle (Deg)

Grazing Angle (Deg)

0 Arg R, 2MHz Arg R, 4MHz Arg R, 8MHz Arg R, 16MHz Arg R, 32MHz

−80

Arg R, 2MHz Arg R, 4MHz Arg R, 8MHz Arg R, 16MHz Arg R, 32MHz

150

Arg ρv

Arg ρh

−40

100

−120

50

−160 0

20

40 60 Grazing Angle (Deg)

80

Magnitude and phase of reflection, H-polarization

0

0

20

40 60 Grazing Angle (Deg)

80

Magnitude and phase of reflection, V-polarization

Reducing conductivity has a similar effect to increasing frequency—see Eq. 28-3.

FIGURE 28-3 Reflection coefficients for horizontal and vertical polarized waves on reflection by ground with er = 5 and s = 10 mS/m (typical of fertile farmland)

at which the reflection coefficient is minimum is called the Pseudo-Brewster Angle, and at this angle the phase of the reflection coefficient is 90°. Figure 28-3 is calculated for er = 5 and s = 10 mS/m. It will be seen from the form of x in Eq. 28-3 that the effects of falling conductivity and increasing frequency are similar, so over poor ground the values tend to fall even further from complete reflection as the ground absorbs more energy from the incoming wave. By selecting the polarization and height of the antenna above the ground, we can effectively control the elevation angle of the main beam maximum (the TOA). The ground— even poorly conducting ground—acts as a good reflector at low angles of elevation for horizontally polarized signals, and the gain of a horizontally polarized antenna is effectively increased fourfold when the antenna is located above good ground. (The reader may consider how this comes to be true, as radiation is apparently only confined to one halfspace, so why is the gain not merely doubled?) As shown in Figure 28-3, the magnitude of rh is not strongly dependent on ground conductivity at low angles, but the contribution added by reflection from the ground falls off as the TOA increases. For many types of horizontally polarized antennas the effective height of the radiating region above ground is fixed, with the result that the TOA falls as the frequency increases. This suits the general behavior of HF paths, because long paths with low TOAs generally operate at higher frequencies than high-TOA short paths.



28-7

FIGURE 28-4 Elevation patterns of a monopole over ground of different conductivities

The effect on a vertical radiator close to the ground is more complex because the current distribution on the element is modified by coupling to the ground, but the general principle still operates. The finite conductivity of the ground reduces the gain of the antenna at low elevation angles, raising the elevation main beam maximum. This can be seen in Figure 28-4, which also illustrates that the effect of finite ground conductivity can be reduced by the use of a ground mat, in this case comprising wires laid in the direction of propagation and typically spaced by around l /10 at the highest operating frequency. When thinking about the radiation patterns of HF antennas, it is important to remember that the fields in the azimuth plane (q = 0°) may be very small, and it is usual to specify the “azimuth” pattern for some specified takeoff angle(s)—effectively the cuts of a cone on the solid 3D pattern. Computer-Aided Design HF antennas are often designed using computer programs that solve Maxwell’s equations numerically for arbitrary antenna geometries, including the effects of the ground and any ground mat.7 They calculate the current distribution on the antenna, and from this they derive the radiation patterns, making allowance for ground effects. In general the radiation patterns are computed to an acceptable degree of accuracy, but the computation of feedpoint impedances should be regarded as a good estimate, rather than reliable fact. It is often worth changing the segmentation of the model and details close to the feedpoint to check that the result is stable and is not sensitive to small changes in the way the real antenna is represented.

28.4 MAJOR FORMS OF HF COMMUNICATIONS ANTENNAS Many HF antennas comprise derivatives of simple dipoles and vertical monopoles. To meet the requirements of extended bandwidth, these simple prototypes are modified into the forms shown in Figure 28-5.


HF Antennas 28-8


FIGURE 28-5 (a) Dipole; (b) Multiple dipole; (c) Loaded dipole; (d) Fan dipole; (e) 3D biconical dipole

When additional gain is needed, or when it is desirable to form a directional beam, the most common practice is to use some form of log-periodic or curtain array. When an antenna is only required to receive signals, it is possible to use compact low-gain elements and to array these to provide the required directivity to enhance the received signal-to-noise ratio. Dipoles and Dipole Derivatives The simplest HF antenna is a wire dipole suspended above the ground (see Figure 28-5a). For NVIS use it should be no more than 0.35l above the ground to avoid an overhead pattern null. The multiple dipole (Figure 28-5b) provides operation at several separate frequency bands, while the loaded dipole (Figure 28-5c), loaded fan dipole (Figure 28-5d), and biconical dipole (Figure 28-5e) will provide low VSWR over extended bandwidths. Monopole Derivatives These designs require a ground mat that forms the second terminal of the antenna and increases effective ground conductivity, enhancing radiation at low takeoff angles. A full ground mat for this purpose comprises a set of between 60 and 180 radial copper (or copper-plated steel) wires, l /4 long at the lowest operating frequency, laid on the ground or buried within a few centimeters of the surface (deeper burial leads to losses in the ground above the wires). The inner ends of these wires are joined together at the base of the monopole; they form the second terminal of the antenna and are connected to the outer conductor of the feeding coaxial line. The biconical monopole (see Figure 28-6) can be constructed using a single metallic mast placed on an insulated base. Internal resonances between the tower and the cone can be removed by connecting the mast to the cone at one or more points. In an alternative form of this antenna the mast is grounded and the lower cone wires are insulated and fed against ground. Midway up the mast the cone wires are connected to the mast to form an inductive loop. This acts as a matching device, allowing the antenna to provide a low VSWR when it is only 0.17l high. Conical monopoles comprise only the lower half of the biconical antenna, and their mechanical construction can pose some interesting challenges.


HF Antennas 28-9

HF ANTENNAS

FIGURE 28-6 A biconical monopole antenna

The azimuth pattern of conical and biconical monopoles is circular at all frequencies, but the elevation pattern changes with frequency in the manner shown in Figure 28-4. Horizontally Polarized Log-Periodic Dipole Array (LPDA) The basic LPDA (see Figure 28-7) comprises a series of half-wave dipoles spaced along a balanced transmission line and connected to alternate sides. Successive dipole lengths and spacings have a constant ratio t (typically between 0.8 and 0.95). The bandwidth is limited by the lengths of the longest and shortest dipoles. The number of radiators depends on the frequency range and the value of t. The parameters t and a , as discussed in detail in Chapter 13, determine the gain, input impedance, and maximum VSWR of the antenna.8,9

Direction of fire

Rn Rn-1

Ln-1

Ln

Ln-1/Ln = Rn-1/Rn = τ

FIGURE 28-7 A transposed-dipole LPDA


HF Antennas 28-10


At any given frequency only those dipoles approximately l /2 long are excited, and these form a single active region. The active region moves along the curtain from the longest to the shortest dipole as the frequency is increased. Because of this movement, it is possible to mount an LPDA so that the electrical height of the active region is constant (the apex of the antenna is placed at ground level) or variable by raising the apex so the active region at the HF end of the array is higher (in terms of the wavelength) than at the lower frequencies. This allows the takeoff angle to be optimized to some extent to match the path characteristics. It is important to choose t and a carefully because of the possibility of unwanted resonant effects—energy that normally radiates from the active region excites a second active region where the radiators are 3l /2 long, causing impedance and pattern perturbations. The antenna is fed close to the shortest element. The balanced feedline is loaded by the elements that modify its effective characteristic impedance. Careful attention to detail is needed if the antenna is to provide a low VSWR across the whole HF band; parameters requiring careful optimization include the characteristic impedance of the feedline and the distances by which it extends beyond the array at both ends. The input impedance of a horizontally polarized LPDA is strongly dependent on the height of the array above ground, and problems can be experienced over sloping ground. (Raising the HF end of the array is usually less problematic than a configuration in which the HF end of the array is electrically closer to the ground.) A vertically polarized array can be formed by rotating an LPDA into the vertical plane and deforming its shape so the lower end of each element is at the same distance above ground— typically about 1 m. The dipoles are suspended from an insulating catenary attached to a mast positioned behind the longest element; the input end of the feedline is supported on an insulating pole close to the front of the antenna, while the far end of the transmission line is attached by insulators to the supporting mast. A shorter form of vertically polarized LPDA can be constructed by placing the main feedline parallel with the ground and using only quarter-wavelength radiating elements— effectively a log-periodic monopole array. A practical difficulty with this array is the need to provide the effective 180° phase shift along the feedline between the radiating elements; this has been overcome by various methods including loading or meandering the transmission line between the elements. The complexities of feeding, together with the requirement for a ground mat, have limited the use of this form of array to circumstances in which the height of a standard LPDA is unacceptable—for example, close to airports. Adequate VSWR performance can be obtained when each radiating element is constructed from a single wire. A long thin element has a high characteristic impedance, and the voltage at the open end is very high even for modest input powers; for peak powers above about 20 kW it may cause corona discharge problems at the open ends. The power rating of the antenna can be increased by forming each element from a number of parallel wires forming a cage perhaps 5–20 cm in diameter. Cage elements have wider impedance bandwidths, so the maximum VSWR of an array can be lower than that possible with single wire elements. Other forms of radiating element and feed systems such as trapezoidal or triangular teeth can be used for the same purposes. A typical horizontal LPDA using halfwave radiators has a gain of 10 to 12 dBi and an azimuth beamwidth of 60–80°; its TOA depends on the height of the active region and can be adjusted by changing the inclination of the antenna relative to the ground. A vertical half-wave LPDA has a typical gain of 11 dBi and an azimuth beamwidth of 110°. Transformations of the LPDA The basic topology of the LPDA has proved very adaptable, and the antenna preserves its wideband characteristics even when subject to transformations that provide a variety of


HF Antennas 28-11

HF ANTENNAS

(a)

(b)

(d)

(c)

(e)

FIGURE 28-8 Transformations of the standard LPDA (a). (b) and (e) NVIS derivatives, using the ground to reflect the downward-directed primary radiation. (c) A common single-mast horizontally polarized version. (d) Sometimes used for high-power rotatable antennas. The arrows indicate the general directions of fire.

new properties. Some of these are shown in Figure 28-8, which shows various ways in which the two half-curtains of the standard LPDA can be mutually rotated and the radiating elements realigned. Horizontally polarized LPDAs can also be stacked above one another to obtain increased gain or arrayed in the azimuth plane to provide omnidirectional or sector coverage. Rotatable LPDAs Rotatable LPDAs are often used when the direction of transmission must be changed from time to time and a compact antenna is required—they have been a common sight on the roofs of embassy buildings in capital cities worldwide. For input powers below about 10 kW these antennas often make use of tapered tubular radiating elements but may also be made from wires tensioned by insulated supporting catenary ropes. The antenna has constant physical height above ground, so its electrical height increases with frequency. As a result, elevation grating lobes form at higher frequencies, reducing gain in the principal lobe and introducing nulls at elevation angles that may be required to serve active propagation modes. These antennas are often mechanically complex, with 3D systems of supporting ropes to support the weight of the radiating elements. They are not easy to erect because they are large, heavy, and fragile; they require regular maintenance. The rotating mechanism may be situated at ground level or at the masthead; in both cases careful design is necessary to manage the large torque generated during rotation and periods of high wind. The large dimensions of a rotatable LPDA can be reduced by loading or folding parts of the longer elements. To minimize antenna dimensions, the highest possible lowerfrequency bound should be adopted for any application. Spiral and Conical Log-Periodic Antennas A major problem at many HF communications stations is a lack of land available for the required number of transmitting antennas. To help solve this problem, multimode antennas


HF Antennas 28-12


FIGURE 28-9

Multimode spiral antenna (Courtesy of TCI)

have been developed that permit several HF transmitters to feed one antenna simultaneously and that provide good isolation between the transmitters. A commonly used structure for such an antenna is a four-arm spiral, as illustrated in Figure 28-9. An antenna of this type operating from 2 to 30 MHz with a gain of approximately 7 dBi can be designed with a height of 37 m and a diameter of about 115 m. Smaller antennas may operate to higher low-frequency limits or may use resistive loading to achieve acceptable VSWR values, reducing their radiation efficiency and power gain. The four-arm spiral can be excited in four independent modes, three of which are balanced and one that is unbalanced. The balanced modes are horizontally polarized and omnidirectional in azimuth at all elevation angles. The unbalanced mode is vertically polarized and also omnidirectional10 (see also Chapter 13). In practice the orthogonality of the normal modes can provide interport isolation between 20 dB and 30 dB. The most common feed system excites the three balanced normal modes. Two of these modes radiate energy directly overhead, and the third mode has a null at the zenith. The configuration shown in Figure 28-9 provides a low TOA at higher frequencies and high TOA at lower frequencies. Excitation of these modes requires that the antenna be equipped with a combination of hybrid and balun transformers. Some multimode spiral antennas are inverted with their apex located near the ground. This arrangement produces TOAs that are nearly constant with frequency. Other configurations of multimode spirals also excite the fourth (unbalanced) mode, in which the tower and antenna are excited as a vertical radiator.

28.5 OTHER TYPES OF HF COMMUNICATIONS ANTENNAS Vertical Whip The vertically polarized whip antenna, usually 2–5 m long, is widely used because it is simple, inexpensive, and easy to mount on a vehicle or the roof of a building. An electrically short whip has a narrow impedance bandwidth and low efficiency; an associated tuning unit is needed to match it at the required operating frequency. The tuning unit may



28-13

be automatic in operation, but is not usually fast enough to support frequency hopping, and its input power is usually limited to a few hundred watts. Whips are commonly used on vehicles, but they suffer the disadvantage that their radiation pattern has a deep overhead null, limiting their effectiveness for short NVIS paths. It is common to see whips bent over the top of the carrying vehicle in an effort to reduce the overhead null, but this unfortunately further reduces their bandwidth and efficiency. Traveling Wave Antennas A traveling wave antenna comprises some form of leaky transmission line fed at one end and terminated by a resistive load at the remote end. It behaves ideally like a matched transmission line and has a relatively constant input impedance over a very wide bandwidth. Its gain, efficiency, and radiation patterns vary significantly with frequency, reducing its useful operational bandwidth to about an octave. The rhombic antenna was widely used for point-to-point communication links. It has generally been superseded for transmitting use by the LPDA, which is smaller and requires less expensive supporting structures, and by LPDAs or active loop arrays for receiving. The sloping-V comprises two wires connected to a feedline at the top of a supporting mast, the wires being arranged in a V-shape in plan with their ends grounded through terminating resistors (each typically 300 Ω). As long as the two legs have an electrical length longer than about two wavelengths, it provides a very simple and quickly deployed antenna with some directional characteristics and modest gain. The inverted-V comprises a single wire element fed at ground level, suspended above ground at its midpoint, and terminating at ground level in a resistive load. A single or multiple wire ground mat is needed to provide the second terminal for both the input and the terminating resistor. The radiation from a long wire has a main lobe firing obliquely off the axis of the wire; an optimum geometry aligns the wires so the directions of radiation of the two wires coincide. The sloping- and inverted-V, which can be viewed as half-rhombic antennas, find application as quick-erect broadband antennas for transportable operations. The Beverage antenna comprises a single wire (or more than one connected in parallel) mounted a few meters above ground level. It is terminated with a resistive load at one end and fed at the other end. It provides a simple low-cost directional receiving antenna for use at the lower end of the HF band (or in the MF band). Loop Antennas When space is very restricted or an unobtrusive antenna is essential, a small loop can be used as a transmitting antenna. A small vertical loop in free space has a frequencyindependent radiation pattern. Its performance can be approximated by a half loop mounted on a conducting plane (see Figure 28-10b). The radiation resistance of a loop is proportional to A2f 4, where A is its area and f is the frequency, so the radiation resistance of small loops is very low at low frequencies. The minimum acceptable size is determined by the bandwidth needed at the lowest operating frequency—a bandwidth of 2 kHz can be provided by a loop 1 m high and 2 m wide. The loop has a narrow impedance bandwidth, and it must be tuned at each operating frequency (see Figure 28-10c). A variable capacitor tunes the main radiating loop to resonance while a small coupled drive loop transforms its resistance. It is possible to tune the antenna automatically using a microprocessor-based system that monitors the current in the loops and sets the capacitor to an appropriate value—as with the short whip, a stepper motor is too slow for frequency-hopping applications.


HF Antennas 28-14


FIGURE 28-10 Vertically polarized transmitting loop: (a) azimuth radiation patterns at different elevation angles; (b) half loop mounted on ground plane; (c) equivalent circuit of half loop on ground plane

The small radiation resistance of the loop at low frequencies makes it inevitable that resistive losses reduce efficiency. Losses are minimized by using a high-Q vacuum capacitor and by ensuring that the current paths in the loop and ground screen have high conductivity. At 3 MHz the efficiency of a well-designed loop will be about 5 percent. Because the radiation resistance of the loop increases with the f 4 but the resistive losses increase only with √f, efficiency rises rapidly with increasing frequency, reaching about 50 percent at 10 MHz and 90 percent above 18 MHz. There is continuing interest in using loops as NVIS antennas for vehicular use.11

28.6 HF RECEIVING ANTENNAS The HF environment is generally very noisy, the principal sources of noise being solar and galactic activity, tropical thunderstorms, and human-made noise. If we receive a signal on a very short dipole and gradually increase its length, the resulting signal-to-noise ratio rapidly increases, but once the incoming signal is large enough to dominate internal receiver noise no further increase occurs. Noise Limitation Fe is the level of external noise in a 1-Hz bandwidth relative to kT0, where 10 log10 kT0 = −204 dBW. Fe can be determined from published noise maps12 or by direct measurement by using a calibrated receiver and antenna. Figure 28-11 shows a typical example of the frequency variation of atmospheric noise; rising noise at low frequencies conveniently compensates for the diminishing effective height of a small loop. Atmospheric noise levels vary with geographical location, season, and time of day. The local human-made noise level depends on whether the receiving site is in an urban, suburban, or rural location; in urban areas it can exceed the atmospheric noise level.



28-15

FIGURE 28-11 Example of an external atmospheric and galactic noise figure: (a) atmospheric daytime noise; (b) atmospheric nighttime noise; (c) galactic noise; (d) worst noise condition. For data related to specific locations see “Radio Noise.”12

The receiver noise figure Fr is obtained from the receiver specification. The transmissionline noise figure Ft is the reciprocal of its attenuation expressed in dB. For systems with filters, preamplifiers, etc., the noise figure Frt can be determined by computing the noise figure for the equivalent cascaded network. The antenna noise figure Fa is given by Fa = D − G + LM

(28-4)

where D = directivity, dBi G = gain, dBi LM = mismatch loss = 10 log10 (1 − | r |2), dB where r is the voltage reflection coefficient The quantity D − G can be expressed in terms of the efficiency h by using the relation D − G = −10 log10 h. The efficiency includes resistive losses in both the antenna and the surrounding ground, which may be significant when earth conductivity is low and there is no ground screen. When fe < fa ft fr , the receiving system is externally noise-limited because the external noise exceeds the internal noise. When fe = fa ft fr , the system is internally noise-limited, and system performance can be improved by reducing the antenna or equipment noise figures. In many geographical locations, atmospheric and noise levels from human activity are extremely low, so receiving systems must have low noise figures to avoid internal noise limitation. Conventional receiving antennas, designed for transmission, can also be used for reception. These are very efficient, with noise figures often less than 3 dB, making them suitable for use at very quiet receiving sites. Because a high HF receiving-antenna noise figure is often acceptable, a receive-only antenna can be inefficient. Consequently, it can be small and if necessary can be resistively loaded to improve its performance.


HF Antennas 28-16


Receiving Loops A practical antenna having a noise figure near the upper limit allowable for most situations is the small vertical loop. A typical loop about 1.5 m in diameter, mounted about 2 m above the ground, has a radiation pattern covering both low and high elevation angles. Successful use has been made of balanced loops (which have classical bi-directional azimuth patterns) and unbalanced loops (which have cardioidal azimuth patterns). The noise figure of a loop depends strongly on its area and the operating frequency. A loop 1.5 m in diameter has a noise figure of about 50 dB at 2 MHz and 20 dB at 30 MHz. It is common to connect the loop to the coaxial feeder cable using a broadband low-noise amplifier that should have good linearity to ensure that intermodulation products are not produced in the intended signal environment. The amplifier typically has a gain of 10–15 dB and serves to match the loop to the feedline and to establish the receiver system noise figure. A number of loops can be arrayed to provide a variety of directional radiation patterns. In most arrangements mutual coupling between the loops is negligible, and simple array theory will predict the radiation patterns. Beams are formed by bringing the coaxial cable from each loop into a beamforming network typically constructed using small ferrite-cored hybrid combiners. The most common arrangement comprises eight loops in an end-fire array with spacing between loops of about l /2 at the maximum operating frequency. If balanced loops are used, the array can be made bi-directional by splitting the signal from each loop and feeding two beamforming networks. Four bi-directional arrays can be arranged in a circular rosette that provides eight independent beams covering 360° in azimuth. Unbalanced loops have been successfully used to form end-fire, broadside, planar, and circular arrays. Adaptive Antennas The availability of cheap microprocessors has made it possible to use active antenna elements as the basis of adaptive antenna arrays. In an adaptive system the amplitudes and phases of the signals from a multi-element array are adjusted dynamically to give a radiation pattern maximizing the wanted signal and minimizing interfering co-channel and adjacent-channel signals. The low mutual impedance between small loops makes them ideal as components of an adaptive array. Circular Arrays Arrays of circularly disposed monopoles or LPDAs are used in monitoring or emitter-locating systems (see Chapter 47). A typical array comprises 18 to 36 elements. The output of each element is fed into a beamformer containing delay lines and power splitters forming a narrow azimuth beam. It is possible to generate N equispaced beams simultaneously in an N-element array by splitting the power from each element N ways. In arrays of LPDAs it is usual to point the elements radially inward. This configuration keeps the effective electrical diameter of the array nearly frequency invariant, making the azimuth beamwidth nearly independent of frequency and eliminating azimuth grating lobes. The disadvantage of the inward-looking array is that each element fires through those opposite, and this complicates the calculation of the radiation pattern.

28.7 BROADBAND DIPOLE CURTAIN ARRAYS Broadband dipole curtain arrays are used for high-power HF ionospheric broadcasting. A rectangular array of dipoles is mounted about 0.25l 0 in front of an untuned reflecting screen consisting of closely spaced horizontal wires (see Figure 28-12). This provides an Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.


FIGURE 28-12 HRS 2/4/h)

28-17

Basic HF dipole curtain array (antenna shown is designated

independent choice of azimuth and elevation patterns, the possibility of steering (slewing) the beam direction in azimuth or elevation, high power gain, wide impedance bandwidth, and high power-handling capacity. Standard Nomenclature Dipole curtain arrays are described by the internationally agreed nomenclature HRRS m/n/h, where H denotes horizontal polarization R denotes an array having a reflector curtain R (if not omitted) denotes that the direction of radiation is reversible S (if not omitted) denotes that the beam is slewable m is the width of the horizontal aperture in half wavelengths at the design frequency n is the number of the dipoles in a vertical stack h is the height of the lowest dipole above ground, in wavelengths at the design frequency The design frequency f0 is approximately equal to √(f1 f2), where f1 and f2 are the lower and upper frequency limits; l 0 is the corresponding design wavelength. Radiation Pattern and Gain The TOA (qmax) and first null elevation angle depend on the average electrical height of the dipoles in the vertical stack and are given approximately by TOA = sin −1 ( f0 λ0 / 4 fH avg )

(28-5)

Null = sin −1 ( f0 λ0 / 2 fH avg )

(28-6)


HF Antennas 28-18


TABLE 28-1

Takeoff Angle of Dipole Array with Reflecting Screen

Number of elements in vertical stack n, half-wavelength spacing

Height above ground of lowest element in wavelengths, h 0.25

0.5

0.75

1.0

1

45°

29°

19°

15 and 48°†

2

22°

17°

14°

11°

3

15°

12°

10°

9°

4

11°

10°

8°

7°

5

9°

8°

7°

6°

6

7°

7°

6°

5°

†

Two lobes present

The TOAs at f0 are given for several configurations in Table 28-1. The lower and upper −3-dB points are located at approximately qmax/2 and 3qmax/2, respectively. The level of the minor elevation lobes is determined by the number of dipoles in the stack and their spacing. The azimuth half-power beamwidth depends primarily on the width of the array, but also depends weakly on the TOA. The beamwidth at f0 is approximately 76° for arrays onedipole wide, 50° for two-elements wide, and 24° for four-elements wide, with beamwidths at other frequencies ( fMHz ) being obtained approximately by multiplying these values by fo /f. Dipole arrays with adjustable horizontal beamwidth have been constructed that use an RF switching system to excite one or more pairs of vertical stacks. The gain at f0 of an array of half-wave dipoles is shown in Table 28-2. The approximate gain at frequencies different from f0 can be obtained by adding 20 log (f / f0) to the values in the table. Slewing Early forms of dipole curtains used full-wave dipoles. They had narrow-bandwidth feed systems and thin dipoles, so they were capable of operating only in one or two broadcast bands. The radiation patterns were slewed in azimuth by up to 15° using a switched feed arrangement TABLE 28-2

Gain of Dipole Array with Perfect Reflecting Screen (dBi)

Number of half-wave elements wide, m Number of 1 2 4 elements in vertical stack n, Height above ground of lowest element in wavelengths, h half-wavelength spacing 0.25 0.50 0.75 1.0 0.25 0.50 0.75 1.0 0.25 0.50 0.75

1.0

1

12.5 13.1 13.9 13.4

13.5 14.3 15.1 14.6

16.1 17.1 17.8 17.4

2

13.9 14.9 15.5 15.7

15.3 16.4 16.9 17.1

18.1 19.2 19.7 19.9

3

15.5 16.3 16.8 17.1

17.0 17.8 18.3 18.6

19.8 20.6 21.1 21.4

4

16.6 17.3 17.8 18.1

18.1 18.8 19.3 19.6

20.9 21.6 22.1 22.4

6

18.2 18.7 19.2 19.5

19.8 20.3 20.7 21.0

22.6 23.1 23.6 23.9



28-19

FIGURE 28-13 Slewing system and corporate feed of a broadband HRS 4/4/h dipole array. Dipole length = 0.46 l 0, dipole spacing = 0.50 l 0 center to center, and screen-to-dipole spacing = 0.25 l 0 .

to provide a phase difference between the two lateral halves of the array. If greater slew angles are attempted, the horizontal pattern develops large secondary lobes that reduce gain by up to 3 dB and may interfere with co-channel transmissions from other stations. To achieve larger slew angles it is necessary to reduce the horizontal spacing between the dipoles. In modern arrays the dipole length is usually slightly less than l 0 /2, and the spacing between the stacks is l 0 /2. The slewing system (see Figure 28-13) enables slew angles to be as large as ±30°. Systems with up to five slew angles are in common use. Dipole arrays can be slewed vertically to change their takeoff angle and consequently the distance to the optimum reception area. Vertical slewing is accomplished by altering the excitation phases of the dipoles in each vertical stack using a switching system operating on the elevation feedlines. Bandwidth The bandwidth of a dipole array depends on the bandwidth of the dipoles (the fatter the better) and of the feed system. The feed system is usually in the form of a branched tree system (see Figure 28-13) using wideband transmission-line transformers (tapers or multisection designs).13 A well-designed broadband dipole array has a VSWR of 1.5:1 or less in its operating bands. For some arrays and slews, mutual impedances cause some dipoles to exhibit a negative driving-point resistance that can cause large currents to circulate in the feed system. Large circulating currents produce high voltages in parts of the array and rapid changes in antenna input impedance, often accompanied by high input VSWR. Arrays must be designed to eliminate these resonances or to move them to frequencies outside the operating bands. Practical Considerations To achieve wide bandwidths, broadband arrays are constructed as fat multiple-wire cage dipoles in the form of rectangular boxes or cylinders. The dipoles are usually folded to raise their input impedances and to provide additional impedance compensation. A folded dipole operates simultaneously in two modes. The radiating mode depends only on the length and


HF Antennas 28-20


equivalent diameter of the dipole cage. The nonradiating transmission-line mode comprises currents that flow in a loop around the dipole and through the short circuits at the end. The impedance variations of the two modes tend to cancel, so the transmission-line mode can be used to compensate and to tune the radiating mode. This compensation is optimized by moving the folded-dipole short circuit away from the end of the dipole toward the feedpoint. HRS 4/4 arrays can be designed to have a power rating of 750 kW average (4MW PEP), which is large enough to accommodate two fully modulated 250-kW AM transmitters fed into the antenna simultaneously by using a diplexer. HRS 4/6/h arrays can be designed to handle the power of two diplexed 500-kW transmitters. The octave bandwidth of a dipole array allows three antennas, 6/7/9/11-MHz, 9/11/13/15/17-MHz, and 13/15/17/19/21/26-MHz, to cover the whole international broadcast spectrum, with multiple-antenna coverage of many of the bands.

28.8 SITING HF ANTENNAS Sloping and Hilly Terrain The ground over which an antenna is mounted is seldom flat and uniform. This is of special significance when designing an antenna with a very low takeoff angle. At low elevation angles the zone over which ground reflections add to the direct signal extends far in front of the antenna, and any non-uniformities will affect the shape of the resulting elevation patterns.14,15 Ground Conductivity For vertically polarized antennas, high ground conductivity is important for the effective operation of transmitting antennas. Vertical transmitting antennas mounted very near seawater, which has a high conductivity (4 S/m), do not require ground screens for radiationpattern enhancement. For most types of soil, conductivity is significantly smaller than this value, so metallic ground screens are sometimes used when a low takeoff angle is needed. For horizontally polarized antennas, good ground conductivity is not important, although the ground losses below the antenna can be reduced by the use of a screen of wires parallel with the direction of polarization. Coupling Between Co-Sited Antennas When energy is coupled between one transmitting antenna and another, it can cause the generation of intermodulation products that can interfere with other spectrum users; it can also cause the incorrect operation of reflected power protection circuits. A typical minimum acceptable coupling is between −20 dB and −30 dB, depending on the technology of the transmitters, but greater isolation will be needed if the transmitter powers involved are substantially different. Receivers are not tolerant of high signal levels presented at their inputs. When high power transmitters are in use, very high levels of isolation are needed to avoid blocking and cross-modulation being caused in the receiver. The small effective height of active loop receiving antennas results in lower internal intermodulation products than may be provided by an LPDA or other more conventional receiving antenna.



28-21

For antennas in each other’s far-field region, an estimate can be made using the standard Friis transmission equation: Pr /Pt = GtGr (l /4p r)2

(28-7)

where Pr /Pt is the numerical ratio of received to transmitted power Gr and Gt are the respective power gains of the two antennas r is their separation (m) l is the wavelength When antennas are cross-polarized in the direction joining them, an additional isolation of 10–20 dB can be assumed. The magnitude of the coupling depends on the polarization of the antennas. For vertically polarized antennas the coupling factor is Cv = 20 log10 (l /8p d ) + Gt + Gr dB

(28-8)

where d = spacing between antennas Gt = gain of transmitting antenna, dBi Gr = gain of receiving antenna, dBi When antennas are only a few wavelengths apart, the mutually induced currents may cause modification of the expected radiation patterns. If this is likely to be important, a computer analysis may be necessary to assess the possible extent of the interaction. Side-by-Side Antennas Horizontally polarized antennas with directive radiation patterns can be placed side by side, often using common towers, with an isolation exceeding 20 dB. Equations 28-7 and 28-8 do not apply in this situation because the antennas are too close. Coupling should be calculated by using computer analysis.

REFERENCES 1. K. Davies, Ionospheric Radio, IEEE Electromagnetic Wave Series 31 (London: Peter Peregrinus, Ltd., 1990). 2. High-Frequency Radio Automatic Link Establishment (ALE) Application Handbook (Boulder, CO: Institute for Telecommunication Sciences, 1999). 3. “HF Propagation Prediction Method,” Rec 533-8, ITU-R, Geneva, 2005. 4. “Computation of Reliability and Compatibility of HF Radio Systems,” Rec 842–3, ITU-R, Geneva, 2005. 5. R. Silberstein and F. Dickson, “Great-Circle and Deviated-Path Observations on CW Signals Using a Simple Technique,” IEEE Trans AP, vol. 13, no. 1 (January 1965): 52–63. 6. A. B. Rogers, C. M. Earrington, and T. B. Jones, “Large HF Bearing Errors for Propagation Paths Tangential to the Auroral Oval,” IEE Proc. on Microwaves, Antennas and Propagat., vol. 144, no. 2 (April 1997): 91–96.


HF Antennas 28-22


7. Numerical Electromagnetic Code (NEC) is a moment-method program developed for the U.S. Navy by Lawrence Livermore Laboratory in 1977. Its derivatives are widely used and can be found on many web sites. 8. E. C. Jordan and K. G. Balmain, Electromagnetic Waves and Radiating Systems, 2nd Ed. (Englewood Cliffs, NJ: Prentice-Hall, 1968). 9. W. L. Stutzman and G. A. Thiele, Antenna Theory and Design (New York: John A. Wiley & Sons, 1981). 10. R. G. Corzine and J. A. Mosko, Four-Arm Spiral Antennas (Norwood, MA: Artech House, 1990). 11. B. A. Austin and W. C. Liu, “An Optimised Vehicular Loop Antenna for NVIS Applications,” 8th Int. Conf. on HF Radio Systems and Techniques, IEE Conf. Publ. No. 474 (July 2000): 43–47. 12. “Radio Noise,” Rec 372–8, ITU-R, Geneva, 2003. 13. G. L. Matthei, L. Young, and E. M. Jones, Microwave Filters, Impedance-Matching Networks, and Coupling Structures, Chap. 6 (New York: McGraw-Hill, 1964). 14. W. F. Utlaut, “Siting Criteria for HF Communications Centers,” Tech. Note 139, National Bureau of Standards, April 1962. 15. Siting of Radiocommunications Facilities—LF, MF, and HF Transmitting and HF Receiving Facilities, Australian Standard AS 3516.1-1988, Australian Standards Association, Homebush NSW, 1988.

BIBLIOGRAPHY Further useful general information will be found in: Handbook on High-Frequency Directional Antennae, CCIR (Geneva: International Telecommunications Union, 1966). “Transmitting Antennas in HF Broadcasting,” BS.80-3 ITU-R, Geneva, 1990. “Transmitting Antennas for Sound Broadcasting in the Tropical Zone,” BS.139-3 ITU-R, Geneva, 1990.



Chapter 29

VHF and UHF Antennas for Communications and Broadcasting Brian S. Collins BSC Associates Ltd. and Queen Mary, University of London CONTENTS 29.1 INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

29-2

29.2 DESIGN FOR RELIABILITY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

29-2

29.3 SOME UNIVERSAL PRINCIPLES . . . . . . . . . . . . . . . . . . . . . . . . . .

29-9

29.4 PERFORMANCE, RELIABILITY, COST, AND THE CHOICE OF MATERIALS . . . . . . . . . . . . . . . . . . . . . . . . . 29-10 29.5 SAFETY: RADHAZ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29-15 29.6 PROPAGATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29-16 29.7 POINT-TO-POINT LINKS AND BROADCAST RECEIVING ANTENNAS. . . . . . . . . . . . . . . . . . . . . . 29-17 29.8 OMNIDIRECTIONAL ANTENNAS . . . . . . . . . . . . . . . . . . . . . . . . . . 29-23 29.9 ANTENNAS FOR BROADCAST SERVICES AND BASE STATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29-28 29.10 BASE STATION ANTENNAS FOR MOBILE PHONE SERVICES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29-36 29.11 SYSTEM CONSIDERATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29-41 29-1 Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.

VHF and UHF Antennas for Communications and Broadcasting 29-2

CHAPTER TWENTY-NINE

29.1 INTRODUCTION The VHF and UHF bands are used for a wide variety of point-to-point and broadcast services with huge economic and social value, generating requirements for many different types of antennas. This chapter includes subjects dealt with in separate chapters in earlier editions of this Handbook and has been extended to cover recent developments in base station antennas, emphasizing methods and techniques that can be applied to these and to many other practical antenna designs. We first review matters of reliability, hardware design, construction, and cost, and then examine the design of antennas for typical applications. Antenna arrays for broadcasting and for mobile communications share many design methods, and engineers working with one of these will find many useful ideas by looking at the methods that have been applied in the other application.

29.2 DESIGN FOR RELIABILITY Any broadcast or communication system must provide an adequate level of reliability and availability. A link may become unusable if the signal-to-noise ratio falls below the design level; it is important that the design objectives for a system specify the fraction of time for which this may occur. A downtime of 0.01% or even less may be necessary for a link to a lifesaving emergency service, but 1% downtime may be as little as can be economically justified for a radiotelephone in a boat used for leisure-time fishing. Fading due to statistical fluctuations in the propagation path is usually guarded against by a fade margin in the power budget. In mobile systems it is standard practice to use a diversity system to reduce the impact of fading on system reliability. This takes advantage of the low correlation between fading events over two physically separate paths, at two frequencies, or for two polarizations. Antenna failure is an important cause of system unavailability. In critical systems reliability is often increased by using two separate antennas fed through a hybrid combiner so that in the event of the failure of one antenna the system continues to operate with a reduced level of signal. When designing an antenna it is very important to understand the user’s requirement for reliability and to ensure consistent performance in every delivered unit. These objectives are achieved by understanding and guarding against the mechanisms that lead to failure and inconsistency. It is very difficult to predict the reliability of an antenna by any formal method, and users will review a proposed design in the light of their experience, looking particularly for weaknesses that may have caused them problems in the past. No telecommunications or broadcasting operator wants antennas to fail in service, and well-designed antennas can provide reliable operation over many years even in difficult climatic conditions. Over-design is expensive and is not necessary; under-design costs more money and endangers vital services. The challenge for engineers new to the field is to develop an intuitive feeling for best practice in this area. The close relationship between mechanical design and electrical function means that antennas must be designed from the outset with their mechanical construction in mind. A good antenna design brings together the best practice in electrical, mechanical, and production engineering. We now examine important causes of antenna system failure. Wind-Induced Mechanical Failures The oscillating loads imposed by wind on antennas and their supporting structures cause countless failures. Aluminum and its alloys are very prone to fatigue failure, and the antenna engineer must be aware of this problem. To achieve real reliability: Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.

VHF and UHF Antennas for Communications and Broadcasting VHF AND UHF ANTENNAS FOR COMMUNICATIONS AND BROADCASTING

29-3

1. Examine available wind-speed data for the location where the antenna is to be used. 2. Consider possible local effects such as turbulence around tall buildings or accelerated airflow over steeply sloping ground. 3. Choose suitable materials for critical high-stress components, and use de-rated permissible stress levels to allow for fatigue. 4. Check antenna designs for mechanical resonance. 5. Damp out, stiffen up, or guy parts of the antenna system that are prone to vibration or oscillation. Base station antennas for mobile radio usually appear to be immune from wind-induced failure, but a large, heavy antenna mounted only on brackets at each end forms a mechanically resonant structure that imposes large fluctuating loads at the support points. The occasional occurrence of bracket failure has led some networks to adopt the practice that every antenna is secured by a stainless steel safety lanyard to prevent a failure in which a roof-mounted antenna could fall into a busy street. More extensive information on design against wind failure will be found in Simiu and Scanlan.1 Corrosion The effects of corrosion and wind-induced stresses are synergistic, each making the other worse. They are almost always responsible for the eventual failure of any antenna system. Every antenna engineer should also be a corrosion engineer; it is always rewarding to examine old antennas to see which causes of corrosion could have been avoided by better design. The essence of good corrosion engineering is ●

●

Selection of suitable alloys for outdoor exposure and choice of compatible materials when different metals or alloys are in contact. A contact potential of 0.25 V is the maximum permissible for long life in exposed conditions. Specification of suitable protective processes—electroplating, painting, galvanizing, etc.

There is an enormous variety in the severity of the corrosion environment at different locations, ranging from dry, unpolluted rural areas to hot, humid coastal industrial complexes. Detailed information on corrosion mechanisms and control is provided in Uhlig2 and Fontana.3 Ultraviolet Degradation Plastics do not corrode, but they degrade by oxidation and the action of ultraviolet light. These effects are reduced by additives to the bulk materials. The extensive use of plastics in buildings and other outdoor applications has led to the availability of well-stabilized grades of polystyrene (PS), unplasticized polyvinyl chloride (UPVC), acrylonitrile butadiene styrene (ABS), and aminoisoprene-styrene copolymer (ASA). At higher frequencies the lower dielectric loss of PS makes it an obvious choice for radomes, but all these plastics are useful for radomes and other external components. Fiberglass (GRP—glass reinforced plastic) is widely used in antenna structures and radomes. For radomes it is usual to specify polyester resin and type-E glass to ensure the lowest losses. GRP must always be protected against UV degradation by specifying a suitable stabilizer in the gel-coat. A radome often has a significant effect on the impedance of the antenna elements inside it, so care must be taken to make sure that the potentially high water absorption of polyester resin does not impair the function of the antenna. Vinyl ester resins have much lower water absorption than polyester resins; although slightly more lossy they can be used where water absorption may be a problem. Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.


CHAPTER TWENTY-NINE

Water Ingress Antennas are exposed to the action of rain and of wide daily and seasonal temperature cycling. Water will penetrate incorrectly designed external joints and fill internal voids in antenna or feed components. The designer must ensure that any voids are either self-draining or are hermetically sealed—there is no reliable alternative. Ice and Snow The accumulation of ice and snow on an antenna causes an increase in the input voltage standing wave ratio (VSWR) and a reduction in gain. The severity of these effects, caused by the capacitive loading of antenna elements and absorption of radio-frequency (RF) energy, increases as the frequency rises. It is important to check that the antenna and its mounting are strong enough to support the weight of snow and ice that will accumulate on them. This is vital because even when the risk of a short loss of service due to the electrical effects of ice can be accepted, the collapse of the antenna is certainly unacceptable! Ice falling from the upper parts of a structure onto antennas below is a major cause of failure; safeguard against it by fitting lightweight antennas above more solidly constructed ones, or provide vulnerable antennas with shields to deflect falling ice. On tall structures, falling ice creates a danger at ground level; feeder gantries must be robust enough to prevent damage by falling ice, and personnel must be kept clear of areas where falling ice presents a hazard. In moderate conditions, antennas may be provided with radomes to cover either the terminals of driven elements or whole antennas. In more severe conditions heaters may be fitted inside antenna elements or to prevent the buildup of ice and snow on radomes. A wide range of surface treatments has been used to prevent the adhesion of ice; some of these show initial promise but become degraded and ineffective after a period of exposure to sunlight and surface pollution. Flexible radome membranes and non-rigid antenna elements have been used with some success. Sand and Dust Antennas in desert areas are subject to sand-laden wind, which rapidly abrades many surfaces. Fine dust accumulates in any unsealed cavity. Breakdown under Power An inadequately designed antenna will fail by the overheating of conductors, dielectric heating, or tracking across insulators. The power rating of coaxial components may be determined from published data, but any newly designed antenna should be tested by a physical power test. An antenna under test should be expected to survive continuous operation at 1.5 times rated mean power and at 2 times rated peak voltage; for critical applications even larger factors of safety should be specified. For multichannel systems with n channels: paverage = p1 + p2 . . . if all channels have equal mean power = npchannel v peak = v1 + v2 + . . . ppeak = ( p1 + p2 + . . .)2 = n 2 pchannel

if all channels have equal power



29-5

In-service failures of high-power antennas often occur at the joints of internal conductors. The passage of high currents through spring connections generates heat, which in turn reduces the temper of the spring, resulting in increased joint resistance and further heating; this leads to a runaway failure. Systems subject to large thermal cycling generate relative physical movement between conductors caused by differential heating and thermal expansion; this movement can cause wear and loss of contact integrity. External insulating components are subject to the action of water and ice; as well as reducing the surface resistance of the material, water absorption can greatly increase the dielectric loss factor of insulating materials such as fiberglass. Exposed components carrying high surface fields must be designed very conservatively, and materials that are likely to suffer surface tracking should be avoided wherever possible. Lightning Damage Antennas mounted on the highest point of a structure are prone to lightning damage. The provision of a solid, low-inductance path for lightning currents in an antenna system reduces the probability of severe damage to the antenna. Electronic equipment is best protected by good antenna design and system grounding, supplemented (for low-power systems) by gas tubes connected across the feeder cables. Figure 29-1 shows a typical system with good grounding to prevent side-flash damage and danger to personnel.4–6 Precipitation and discharge noise are caused when charged raindrops fall onto an antenna or when an antenna is exposed to an intense electric field in thunderstorm conditions. Precipitation noise can be troublesome at the lower end of the VHF band and may be

FIGURE 29-1 Typical example of good grounding practice



CHAPTER TWENTY-NINE

experienced frequently in some locations. When problems arise, antenna elements may be fitted with insulating covers. These prevent the transfer of charge from individual raindrops into the antenna circuit. Passive intermodulation products (PIMs) are generated when radio-frequency currents flow across any joint between conductors. They will generally encounter some nonlinearity in the resistance of the joint, caused by the presence of oxide layers on the conductors and made worse by the increase in current density because of the micro-scale roughness of the contacting surfaces. It has been a longstanding practice to use single antennas for the transmission of multiple radio and TV signals, but the only limit on PIM was the radiation of unwanted spurious signals, perhaps 100 dB below the wanted signals. The advent of mobile radio base stations led to the use of a single antenna for transmission of multiple carriers in duplex operation and use of the same antenna for multichannel reception. This has led to the usual adoption of the requirement for PIM to be lower than –153 dB with respect to 2 × 20W carriers. This ratio (1015.3) is approximately the ratio of the distance from the earth to the sun compared with the thickness of a piece of thin paper, so achieving this limit on a consistent production basis presents a major challenge. The achievement of low PIM depends on following clear rules in design and at all stages of production: ●

●

●

●

●

●

●

Avoid any interconductor joint that is not essential to the electrical operation of the antenna. Ensure that every essential joint is tightly compressed and that the force on the joint does not depend on any compressible or creep-prone material. Avoid dry solder joints—well-soldered joints seem to cause few problems. Ensure PCBs are cleanly etched and well washed after processing. Protect all essential joints from corrosion (and wherever possible from contact with water). Avoid the use of metals with nonlinear conductivity or contact potentials—nickel is a known PIM source, and aluminum joints must be provided with corrosion protection and high contact forces. Maintain a high level of cleanliness in storage and assembly areas.

Coaxial Connectors A wide range of coaxial connector interfaces is available, and the choice between them will usually be made on the basis of the mean power ratings that the connectors must support. For permanent installation, most tower-mounted installations use Type N, 7/16-DIN, and the familiar range of bolted EIA flange connectors. If a connector is underrated, the spring inner conductor components will overheat; this leads to softening of the metal from which they are made, reduction of the contact pressure, and consequent increase in contact resistance leading to even more heating and eventual failure. The quality of the fitting of connectors is important in achieving low PIM, low VSWR, and long-term reliability. For long-term reliability it is standard practice to provide additional weatherproofing to connectors, either by fitting a purpose-designed shrink-on boot or by wrapping with tape. Self-amalgamating tape provides physical protection, but some types do not form a waterproof bond at the connector surface; they should be overwrapped with a layer of petroleum jelly–impregnated tape, which remains waterproof for decades. Where a cable is terminated directly into connection points on the antenna—for example, in domestic radio and television antennas, it is worth applying silicone sealant to the end of the cable and the terminals of the antenna to prevent water penetrating the cable and increasing its loss.



29-7

Coaxial Cables For applications at VHF and above, the ratings of cables and connectors are generally determined by their attenuation or the mean power they are required to carry. The maximum power rating of a cable is determined by many factors: ●

●

●

●

●

●

●

●

●

The cable diameter Large cables have lower surface current densities and more surface from which heat can be radiated or conducted. The loss factor of the dielectric Foam dielectrics have lower losses than the same material in solid form. The thermal conductivity of the dielectric Heat generated in the inner conductor must escape through the dielectric and the outer part of the cable. The operating frequency As the frequency rises, the skin depth on both inner and outer conductors diminishes, so resistive losses increase, reducing the ability of the cable to carry power. The ambient temperature The cable has a fixed maximum internal operating temperature, so in high ambient temperatures it can carry less power. The thermal environment of the cable Cables installed underground, in floor ducts, or in full sunshine, may require additional derating factors to allow for these conditions. It may be economic to fit a sunshade over a cable rather than to use a larger cable. The maximum operating VSWR The cable will become hottest at points of maximum current. In general the effect of an operating VSWR s will be that the maximum power of the cable is reduced by a factor 1/s. The possibility of operation with a high VSWR caused by fault conditions or ice accumulation must be considered. The VSWR for this calculation should be based on the maximum VSWR into which the transmitter will operate before protective trips reduce its output power. At UHF the axial thermal conductivity of a large cable allows heat to flow away from the points of maximum current, reducing the derating factor. Fault conditions In fault conditions the power division ratio of a junction may effectively change, increasing the power into parts of the antenna. Any fault that increases the power in part of the antenna must be considered when choosing cable power ratings. Peak voltages The peak voltage rating of a cable may be exceeded before its mean power rating is reached, so its peak voltage rating or equivalent peak power rating must be checked against the operational requirement. Peak power rating limitations are common at MF and HF and may determine the choice of cables for multichannel VHF and UHF antennas.

Derating factors for these effects are applied cumulatively, so the effective power rating of a cable may be far lower than that given by the power/frequency curve in the cable specification. In a complex antenna system the cables at every level of the feed system must be assessed separately. Information on derating factors will be found in manufacturer’s catalogs—often in inconspicuous tables at the back. Large coaxial cables are heavy and difficult to bend. It is essential to allow adequate bending radii and enough clearance along the cable route to allow it to be installed and formed into position. It is always worth consulting an experienced installer to check the feasibility of complex cable runs. Arrangements for lifting and supporting cables should always be planned by reference to manufacturers’ specifications. Cables for receiving antennas are selected on the basis of their attenuation per unit length (dB/m). The least lossy cables for a given outside diameter are those with helical or foam polyethylene dielectrics. Cables with continuous copper outer conductors have the best screening against pickup when passing through regions of strong



CHAPTER TWENTY-NINE

signals—perhaps past transmitting antennas. More flexible cables have braided or tape screens, with multiple screens used for improved shielding when necessary. Some lowcost cables have sparse braids that offer relatively poor immunity to pickup of noise and spurious signals. Large antenna systems often comprise multiple radiating elements connected by coaxial cables that may be prone to damage by wind vibration, falling ice, and personnel climbing within the array during installation or maintenance operations. To ensure high reliability, secure the cables adequately and route them where they are unlikely to be trodden on or used as “handles.” Staff working on these antennas must be trained to appreciate these issues. Antenna Design for High Operating Power Many of the factors just listed are relevant when designing an antenna to handle any power that is high enough to cause local heating and damage to its components. A small antenna operating at 2 GHz and carrying 100 W may be vulnerable to power failure if the designer failed to recognize the need to limit power dissipation and to avoid thermally insulating the components that will generate heat under power. A multichannel GSM omnidirectional antenna designed to carry 500 W provides a critical test of the designer’s understanding of thermal design. For components that will operate at elevated temperatures, make sure the chosen materials will not degrade at their expected operating temperature and that when power is turned on or off, the resulting thermal expansion/contraction can be accommodated without causing stresses. Where thermal expansion may cause relative movement between parts, the design must allow for this by providing a reliable low-resistance current path and avoiding wear by repeated small movements. Power Testing There is no substitute for conducting a power test, either on a complete antenna or on critical components like radiating elements and power splitters. The mean and peak power levels for the test should be chosen to provide the factors of safety suggested in this section. To provide in-service reliability it is important to understand the operating temperatures that may be experienced by internal components, relating this to the softening point of plastics and to the maximum service temperature for spring connections and solder joints. The tools for temperature measurement include infrared thermal cameras for external components and thermochromic paints and stick-on labels for internal components. Thermochromic labels may heat up by dielectric heating in high RF fields; take care that as far as possible they are protected from them. When conducting a power test, monitor the temperature of the largest masses to make sure the test is long enough for everything to reach thermal equilibrium. Record the ambient temperature at the time of the test, as the critical parameter is the temperature rise above the ambient temperature recorded at various points of the system; if the antenna will be used in a higher ambient temperature than during the test, allow for this when deciding on service ratings. Installation Design and Practice An excellent guide to good practice is provided by the ETSI Guide.7 Large antennas will be transported in sections and reassembled on site, so all components must withstand transport and rough handling. They must be designed such that reassembly can be carried out on site by an antenna rigger who has no knowledge of the subtle points of antenna design. Components should fit together only one way (the right way) and any (unavoidably) interchangeable parts must be very clearly marked. Any components that were earlier tested as subassemblies should be marked so the same components are reassembled in the same way.



VHF AND UHF ANTENNAS FOR COMMUNICATIONS AND BROADCASTING

29.3 SOME UNIVERSAL PRINCIPLES The mechanical design of any antenna system needs to provide a number of basic features: Repeatability of electrical characteristics between manufactured samples Robustness to withstand transport and installation Stability of electrical characteristics in service Resistance to lightning damage Low RF losses Low levels of passive intermodulation products

●

●

●

●

●

●

TABLE 29-1 Typical Mechanical Details Illustrating Good Design Principles for Antenna Design Detail

Practical Imperfections

Preferred Detail Example

1 (a) Strip conductors joined to one another or to ground.

Contact pressure at circled point is not defined. Loss, power rating, and PIM are uncertain. Impedance changes if contact point moves.

Right angle connection between flat conductors.

Any lack of square or deflection of the joint displaces the contact point: loss, PIM, unstable Z.

(b) (b) Can also be achieved by bending or pressing strip end(s).

2

Move the fixing closer to the bend, or press the flange to define the contact point.

3

Connection through a dielectric layer.

Most thermoplastic materials creep under pressure, so the joint integrity is lost and heating occurs. Brittle material may be fractured if joint is overtightened.

Add a sleeve over the fastener to remove stress from the dielectric and to form a stable current path.

4

Clamping through a dielectric layer.

The problems are the same as those of the preceding example, but this is worse because the pressure under the fastener is higher for a given joint closure force.

Add a sleeve and a washer to reduce pressure: also observe precautions as in (1).

5

Current path relies on contact through thread. 6

This is particularly to be avoided. It shares the problems of (3), and the point of current flow between the screw and the upper conductor is undefined. Don’t rely on current flowing from a screw into the threaded component.

Defined current path.

These examples are exaggerated to show clearly what is happening. Spring plug contact is smaller than hole.

Current path is incorrect: impedance uncertain and current density at contact point is high. Plug will overheat.

Contact is at entry to hole. it is stable and less dependent on spring temper.



CHAPTER TWENTY-NINE

Their relative importance varies with the application, but fortunately, the same design methods improve all these desirable features and, with experience, many of them can be incorporated into a design with minimal extra cost, adding real value to the product. Many important design techniques relate to the configuration of joints between conductors, whether these are part of a radiating structure or an internal feed system. RF currents flow only in the surface of the conductors that carry them, so the achievement of our objectives depends on the way in which the conductor surfaces relate to one another at every interconnection. There must be a well-defined surface current path, maintained independently of production tolerances and bearing sufficient mechanical pressure to ensure a stable current path whatever forces the joint may experience on account of temperature change, vibration, wind load, or other causes. It is essential to remember that in the real world no surfaces are flat—they are always rough and slope in one direction or another. No surfaces are exactly square, and no conductors are completely rigid. Table 29-1 shows examples that illustrate simple principles that must be followed in every antenna design. These examples are not exhaustive. If the reasons for avoiding the details in the left-hand column are appreciated, a whole variety of good solutions may often be found. The criticality of these details increases as the power and/or design frequency increase, or as the requirement for low PIM or resistance to lightning damage becomes more severe.

29.4 PERFORMANCE, RELIABILITY, COST, AND THE CHOICE OF MATERIALS On many occasions the antenna designer will receive a complete electrical and mechanical specification for an antenna and will create a design that provides the best ratio of performance to cost. On other occasions the designer will have knowledge of the system requirement for the antenna and can optimize the technical and economic aspects of the antenna design together with its supporting structure, cables, and associated hardware; this process often provides more value to the user because the designer can balance a variety of parameters that contribute to the performance and costs of the system. The reliability of antenna systems continues to increase, reflecting the availability of improved manufacturing techniques, better materials, and the experience of designers and users. There is a huge repertoire of solutions to obtain any needed electrical performance, so much of the antenna designer’s time is now spent matching available design concepts to low-cost materials and production methods. Customers’ expectations of “more for less” apply to antennas just as they do to any other product. Products that demand hours of patient adjustment to finesse them to meet their specification are not commercially viable; someone will have designed a simple product that works every time—the target we must achieve! In many respects broadcast services operate under the most difficult constraints, as they must provide a very high standard of coverage in a target region while observing stringent restrictions on signals radiated outside the target. This is made difficult because their target regions are often defined by commercial and political constraints that don’t fit well with antenna design equations. The designer must provide the required overall performance for the lowest cost. Reliability of 100% is often very difficult and costly to achieve and is only necessary for a small number of services. By comparison, 99.9% availability will entirely satisfy many users and can be provided much more readily; the user may be unable to justify the high cost of that extra 0.1%.



29-11

Cost-effective design is only obtained by: ●

●

●

●

Identifying the availability and coverage needed Determining the environment at both ends of the path Estimating the propagation characteristics of the path and judging the reliability of the estimates Selecting equipment and antennas to meet the communications and reliability objectives

This is a general process, applicable to point-to-point and broadcast systems. For a pointto-point system we may have control of the equipment at both ends of the link, and resources can be balanced to provide the best solution. In a broadcast or mobile radio system, coverage plans are based on assumptions about the effective isotropic radiated power (EIRP) of the transmitter, the height and gain of the user’s antenna, and the sensitivity of the receiver. Trade-Offs For any major scheme it is always advisable to work through the following checklist: 1. Examine the interactions of structure height, transmitter power, feeder attenuation, and antenna gain. 2. Consider using split antennas and duplicate feeders to increase reliability. 3. Consider the use of diversity techniques to achieve target availability instead of a single system with higher power and gains. 4. Review the propagation data, especially the probability of multipath or co-channel interference. Don’t engineer a system with 99.9% hardware availability and find 3% outage due to co-channel problems. Check the cost of antennas designed to reduce co-channel problems by nulling out the troublesome signals and the use of diversity techniques to combat multipath fading. 5. Visit the chosen site—if that’s not possible, request maps and photographs of the site and its surroundings. General wind data are useless if the tower is near a cliff edge—a careful estimate of actual conditions must be made. A nearby industrial area may mean a corrosive environment, and nearness to main roads indicates a probable high electrical noise level. Look for local physical obstructions in the propagation path. Check access arrangements to confirm assumptions that it will be possible to transport the antenna to site without taking it to pieces. 6. Don’t over-design to cover ignorance. Find out! Cost Issues A glance through the early chapters of this Handbook demonstrates that many different types of antenna can be used as the basis of design for most applications. The choice between approaches sometimes relates to physical attributes, for example, to minimize dimensions or to ensure robustness. In the majority of cases the designer must compare a number of possible approaches and select candidates for further development on the basis of their suitability for production—the potential for using low-cost techniques and materials and avoiding critical tolerances to ensure reproducibility in production. However hard the path, it is never good enough to produce just “a design that works.”



CHAPTER TWENTY-NINE

There are no unique solutions to minimizing costs. The optimum design for any application will depend on labor and material costs in the location where production will take place, although a design that offers the user a higher performance/price ratio may succeed over a simple minimum-cost approach. A design that minimizes the number of components— using each physical component to deliver multiple functions—often provides a result with low cost, high reliability, and clear customer appeal. For a design that is going to be produced in large numbers, figure out the cost of materials, and count the number of parts, holes, fasteners, and critical dimensions. Then rework the design several times to reduce them all! A major contribution to excess cost is overspecification of mechanical tolerances in an effort to control production variability. There is no substitute for carrying out a sensitivity analysis to check the dependence of a design on component tolerances (both mechanical and electrical)—this allows noncritical parameters to be assigned relaxed tolerances, and the design of critical areas can be reviewed to find less sensitive solutions. It is easy to assume that components manufactured using CNC (computer numerical control) machines have vanishingly small mechanical tolerances, but the designer who relies on this will soon be disappointed. Always ensure that the mechanical engineer who will translate your favorite electrical design into hardware understands which tolerances are critical, which must avoid cumulative errors, and which can be relaxed. Materials that appear interchangeable to a mechanical designer may be very different in electrical properties. Design for Production Many antenna designs are created to be manufactured in large volumes. Not only must the designer devise a structure with all the right electrical properties— patterns, gain, VSWR, and the rest—but it must be possible to produce the design using standard industrial processes. The performance of a prototype must exceed that required in production by a large enough margin so that almost none of the units produced will be outside the required performance specification. The cost of producing the design in the target volumes, including the costs of scrap and reworking any failures, must meet the requirements of the market and the shareholders. It is not efficient to begin a design using one technique and collection of materials, and then to decide late in development that to meet production demands the techniques and materials must be changed. An understanding of production techniques that would be economic given the forecast production volumes and rate of production must precede electrical design. At every step of electrical optimization the designer must bear in mind the practicability of converting the ideas into producible hardware and must understand the consequences of dimensional tolerances and variable material properties. The design task includes identifying areas of greatest risk and modifying the design to reduce their potential effects; changing materials or processes should only be considered if no cost increase would result. Materials When selecting materials for antenna construction it is increasingly necessary to consider the need to design for recycling and to ensure that products do not contain toxic materials. The European RoHS directive and initiatives in China, Japan, and some states of the United States have led to the elimination of toxic materials from electrical and electronic products; this trend will continue, embracing a wider variety of products and potential toxins. The approach of many major international companies is to insist on compliance with at least the RoHS limits8 for any product they purchase. More extensive checklists follow the Japan Green Procurement9 or EIA recommendations.10 This process is seen as responsible citizenship rather than simply compliance with regulations. The currently restricted materials include lead, mercury, hexavalent chromium, cadmium, PBBs, and PBDEs (compounds previously used as flame-retardants in plastics). As well as prompting the move to lead-free solders, the regulations have changed practices in metal finishing, and



29-13

in the stabilization and fireproofing of resins, adhesives, and plastics. Design for recycling is a developing art that will continue to challenge old assumptions about the way materials are selected, components are assembled, and products are packed. Designers everywhere need to be aware of the rules that their companies must conform to when trading in international markets. The Internet is an excellent source of information about current legislation and best practice, both in your country and in international trade. Conductors The conducting materials used for the construction of most VHF/UHF antennas include galvanized steel, aluminum alloys, and copper (usually in the form of transmission lines and printed circuit laminate). Other materials—including zinc, brass, and stainless steel—are sometimes used in special applications, and also to provide protective coatings. The results of using materials with lower conductivities must be understood in the context of the intended operating frequency and power. At UHF most of the current may flow in the plating rather than the base material, so plating materials must be chosen with their conductivity in mind. Efforts to enhance the conductivity of a less conductive base material are not very reliable; the conductivity of a plated metal (for example, silver) may be lower than that of the material in bulk because of the nature of the crystalline structure of the plated coating. Insulators Most insulators are made from a surprisingly small range of materials. Antenna system power ratings are often determined by dielectric heating, so for many applications the choice of insulating materials is limited to a small group of low-loss materials: polyethylene (PE), polypropylene (PP), polystyrene (PS, or better, high-impact polystyrene—HIPS), and polytetrafluoroethylene (PTFE). Of these the first three are readily available at low cost and can be molded into complex shapes. PE, PP, and PTFE change shape significantly under prolonged mechanical loading (a process known as creep), especially at elevated temperatures. PTFE has a thermal coefficient of expansion that is nonlinear over the usual range of ambient temperatures. PE, PP, and PS suffer surface tracking when exposed to high electric fields—especially if the surface is exposed to rain—and tracking immediately carbonizes the surface, leading to rapid failure. PTFE can be used successfully with significantly higher tangential electric fields along the surface than the other materials. When designing with plastics it is advisable to specify the exact material (including manufacturer and grade). Permittivity and dielectric loss may vary between different grades of the same generic material, especially if recycled material has been included. Ceramic insulators are expensive and are not commonly used in VHF/UHF antenna systems. When using a material for the first time always check its relative permittivity, loss factor, and water absorption. Printed Circuit Boards These are increasingly used for the construction of antennas at frequencies from the UHF TV band upward. They provide a low-cost method of reproducing large volumes of devices with tight tolerances for use at low and medium powers. The main problems associated with their use are as follows: Power limitation is caused by lossy dielectric, limitation of current density in narrow conductors, and high field strengths associated with the inevitable sharp edges of etched conducting tracks. When working with laminates based on PTFE or other high-temperature materials, remember that the failure point of the laminate may occur at a very high temperature, so the power rating of an antenna may be limited by the power capable of unsoldering the input connector. Unless provision is made to transfer heat from the critical components, failure may occur at a much lower power than expected. Performance variability is caused by tolerances in the dielectric constant of the laminate, etching tolerances (over-/under-etch), and problems on long antennas caused by stretching of the photographic film carrying the negative image of the tracks. Errors in



CHAPTER TWENTY-NINE

registration between the faces of double-layer structures and tolerances in the positions of holes and other features must not be overlooked. When manufacturing long boards with close tolerances it is important to keep the film negatives for different layers together, so that any stretch does not cause progressive registration errors, and they should preferably be stored flat or laid out flat for some time before they are used. Under- and over-etching cause the width of microstrip lines and other features to vary and can be detected by positioning telltale features at strategic locations. Gap-and-track or wedge features can make very small under-/over-etch clearly visible to the naked eye. Mechanical stability When designing with PTFE-based laminates it is very important to remember the very poor creep strength of this material, even when reinforced with glass. No electrical connection should rely on compression of the laminate—it will simply flow away, and the integrity of the electrical contact will be lost. None of these limitations is insuperable, but every designer must understand them, control their extent, and manage their effects. This becomes more critical as the operating frequency is increased. There are few fixed rules, as many of the effects depend on the processes used by the board fabricator and the etching shop. The mobile radio base station market has driven many changes in the cost and variety of available materials and the availability of processors who can cope with boards up to 2.5 m long. Computer Design Tools The art and science of antenna design has been hugely advanced by the availability of powerful low-cost computers and specialist electromagnetic simulation packages. For some purposes a computer model will provide a highly accurate estimate of the performance to be expected from a real antenna and allows optimization to meet a practical requirement. Some antennas are still too complex to model as a whole (for example, a large multi-tier TV broadcast array), and the whole design problem must be divided into separate areas for optimization—in this case the design of the panel and the power dividers, the optimization of the geometry and excitation of a tier of the array to create the required azimuth radiation pattern, and the design of the excitation of each tier to provide the required elevation radiation pattern. Simulation is very useful for investigating the effects of mechanical and material tolerances. As the operating frequency increases, the precision with which an antenna is modeled becomes increasingly important, especially in the input region. Software tools can analyze the performance of an antenna structure but in general are unable to synthesize antennas or arrays to meet specific requirements. A computer does not understand how an antenna works, or which aspects are important to the user; it simply predicts the performance of the input structure. In the frequency bands addressed by this chapter, moment-method and time domain simulation tools are useful. Full 3D tools are needed for many applications, but 2D or 2½-D tools are useful for some antenna structures and microstrip feed systems. Computer analysis of aerodynamic and structural design is usually carried out using finite-element methods and allows designers to optimize the shape of critical antenna components. A wide variety of commercial software is available, and every organization engaged in antenna design needs to have access to programs appropriate to the complexity of the work in which it is engaged. The designer should always view simulation outputs showing the fields, currents, or mechanical stresses present in an antenna design. An intuitive feel for their behavior is one of the hallmarks of the successful designer.



29-15

29.5 SAFETY: RADHAZ The maximum permitted RF field and power flux levels to which personnel may be exposed are specified in various national and international regulations11 and corresponding national standards. These ensure that the effects of the electromagnetic fields on the human body are controlled at a level at which no health effects are known to occur. Those working around LF, MF, and HF antennas should also recognize that the limiting parameter may be the intensity of the H-field rather than E-field or power flux density. Antenna engineers and riggers are probably the most exposed group in the population, and they should be particularly aware of the exposure standards and the need to avoid overexposure of themselves and others. Antenna engineers become accustomed to working with antennas carrying powers that may be sufficient to cause death or serious injury. Whether climbing close to broadcasting antennas or mobile radio base stations, or conducting power tests in a laboratory, all those working with live antennas need to be made aware of the potential hazards created by non-ionizing radiation. When possible all power should be removed from antennas before climbing near them; transmitters should be locked off, and the person working on the antenna should take the keys or the fuses so that no one can inadvertently apply power while work is in progress. When it is necessary to pass close to operating antennas a competent person should calculate the probable exposure to high fields, and the personnel concerned should carry personal alarms to warn them if they move into areas of unacceptably high field strength. When designing new antennas that are intended to allow internal access (like VHF/UHF panel arrays), care must be taken to make sure that internal access ways are well enough screened to allow safe access, if necessary after reducing normal operating power to permit access. Explicit advisory notices must be placed at the base of the antenna supporting structure to advise workers of the necessary precautions to ensure their safety. Routine production testing of the passive intermodulation performance of base station antennas creates a significant hazard, and facilities must provide foolproof protection of personnel using the facility and reassurance for those in the vicinity; this can be provided using infrared beams or pressure-sensitive mats to switch off the power in the event that someone is too close to the antenna under test. Public concerns regarding the possible hazards posed by radio installations appear increasingly likely to be unfounded.12 Typical power flux densities to which the public is exposed are usually three orders of magnitude below the international limits. Antenna engineers have a responsibility to provide others with accurate estimates of exposure levels and to monitor the developing research literature on potential health effects.13–15 Field strengths many wavelengths distant from an antenna—for example, at points on the ground or on nearby rooftops—can be calculated on the basis of the power flux per unit area. The power flux through the surface of a sphere of radius r surrounding an isotropic radiator radiating Wt watts is Wt/4p r2 watts per square meter. For an antenna with gain G, the power flux is GWt /4p r2.W/m2. By knowing the mean input power to the antenna, its gain in the direction of interest, and the distance to the point of potential exposure, the likely power flux can be calculated and compared with that permitted. At points close to the antenna a simple and reasonably accurate power density estimate is obtained by assuming that Wt watts flow out through a cylindrical surface with the same length as the antenna and a width equal to the distance between the 3-dB points at the relevant distance from the antenna. Assume a back-to-front ratio of only 10 dB for persons close to the antenna, because they cannot be expected to stay in the rear minimum.



CHAPTER TWENTY-NINE

29.6 PROPAGATION The study of propagation effects in the VHF/UHF bands is beyond the scope of this chapter, but all designers of systems required to produce reliable end-to-end connection need to familiarize themselves with tropospheric effects, terrain blocking, and the effects of the built environment. While the prediction of loss over a point-to-point path is relatively straightforward, the prediction of coverage for broadcast and cellular systems has to be approached on a statistical basis to allow for the variations in the position and path conditions to large numbers of users. The possibility of signal impairment by multipath propagation affects many systems. When designing antennas for a long link, consider the possible effects of fading and ducting, in which typical signal levels are either reduced, causing the potential loss of established links, or increased, causing potential co-channel interference from distant users of the same frequency. Propagation effects influence antenna system design in a variety of ways including the provision of adequate gain and fade margin, the height at which antennas must be mounted, the choice of polarization, and the provision of space, polarization, or frequency diversity. Detailed prediction of coverage areas of broadcast and cellular systems is usually carried out using computer simulations, and a wide variety of planning tools is available, embodying terrain models of various degrees of detail and sophistication. A comprehensive set of graphical data is provided in Shibuya.16 Use of the finite resource of the electromagnetic spectrum is optimized not only by the adoption of optimum modulation systems and signal formats, but also by the choice of antenna characteristics. Frequency reuse in cellular systems is managed by the positioning of base stations, antenna pattern specifications, and the use of elevation downtilt. The avoidance of co- and adjacent-channel interference in domestic TV systems is managed by the radiation pattern specifications applied to broadcast stations and domestic receiving antennas.

Choice of Polarization Base stations for mobile services formerly used vertical polarization because it was simple to provide omnidirectional antennas at the mobile terminals and the base station. Cellular systems now often use 45°-slant linear polarization because the signals received from a typical mobile on a pair of antennas polarized at +45º and –45º have sufficiently low correlation to provide useful diversity gain. Circular polarization is often used for radio services to reduce the influence of the polarization of the receiving antenna and to make it easier for the untrained user to find a consistent high-quality signal. There is sometimes an advantage in using horizontal polarization for obstructed pointto-point links in hilly terrain, but the choice of polarization is often determined by the need to control co-channel interference. Orthogonal polarizations are often chosen for antennas mounted close together in order to increase the isolation between them. The use of circular polarization (CP) can reduce the effects of destructive interference by reflected multipath signals, so CP should be considered for any path where this problem is expected. CP has been used with success on a number of long grazing-incidence oversea paths where problems with variable sea-surface reflections had been expected to be troublesome. Both ends of a CP link must use antennas with the same sense of polarization. In some countries circular polarization is used for TV services, but for this application the receiving antenna is usually horizontally polarized.



29-17

29.7 POINT-TO-POINT LINKS AND BROADCAST RECEIVING ANTENNAS Most low/medium gain antennas are of end-fire design. When higher gain is needed, it is usually obtained by forming arrays of basic end-fire antennas. The antenna styles described here are only a small subset of all the possible designs, but they represent a large proportion of antennas used worldwide for practical applications. In some countries, bowtie dipole and corner reflector designs are popular for domestic TV reception, and they are described later in this section. From a cost/performance point of view they have little significant advantage. Simple log-periodic antennas will cover the UHF band. A variety of hybrid Yagi-Uda, log-periodic, and other solutions is used to provide multiband operation over the VHF and UHF bands. Many of these designs have only modest gain, but the high signal levels available outdoors in many cities allow them to provide adequate reception. In situations with co-channel interference or multipath (ghosting), a more conventional antenna with a narrower azimuth beamwidth and lower sidelobes is often preferred. Yagi-Uda Antennas The Yagi-Uda antenna is the most successful general-purpose directional antenna design at frequencies up to at least 2.5 GHz. It is inexpensive and simple to construct, and will provide gains of up to about 17 dBi (or more if a multiple array is used). Its bandwidth can be chosen to suit the application; almost an octave bandwidth can be achieved, but at the expense of a lower gain than can be provided over a narrower band. At low frequencies the realizable gain is limited by the physical size of the antenna; in the upper UHF band a reflector antenna may be simpler, less costly, and more reliable if a large gain is required. Yagi-Uda antennas provide unidirectional beams with moderately low side and rear lobes. The characteristics of the basic antenna can be modified in a variety of ways, some of which are shown in Figure 29-3. The basic antenna (a) can be arrayed in linear or planar arrays (b). When the individual antennas are correctly spaced, an array of N antennas will have a power gain N times as large as that of a single antenna, less an allowance for feeder losses. Table 29-2 indicates typical gains and arraying distances for Yagi-Uda antennas of various sizes. Different array spacings may be used when a deep null at a specified bearing is required, but the forward gain may be slightly reduced. To achieve closely definable radiation patterns it is important to feed the radiating element using a suitable balun to avoid unwanted radiation from the feed cable. Typical YagiUda antennas will provide cross-polar discrimination (XPD) of at least 20 dB in the main beam direction; higher XPD can be obtained by observing good symmetry and a wellbalanced feed/balun system. Yagi-Uda antennas can be built to support high input powers, and they are commonly used for directional broadcast transmission. Only one element and its drive circuit must be specially designed to support high input power. The bandwidth over which the front-to-back ratio is maintained may be increased by replacing a single reflector rod by two or three parallel rods (c)—as often used for domestic TV receiving antennas. The back-to-front ratio of a simple Yagi-Uda antenna may be increased either by the addition of a screen (d) or by arraying two antennas with a quarterwavelength axial displacement, providing a corresponding additional quarter wavelength of feeder cable to the forward antenna (e). A well-designed screen will provide a backto-front ratio of as much as 40 dB, while that available from the quadrature-fed system is typically 26 dB.



CHAPTER TWENTY-NINE

FIGURE 29-2 Configurations of Yagi-Uda antennas: (a) Standard six-element antenna; (b) Stacked and bayed arrays; (c) Double reflector rods; (d) Reflector screen; (e) Increased F/b ratio by l /4 offset; (f ), (g), (h) Arrangements to produce azimuth radiation patterns for special applications



29-19

TABLE 29-2 Typical Data for Yagi-Uda Antennas Number of Elements 3

Typical Gain, dBi

Spacing for Arraying, l

7

0.7

4

9

1.0

6

10.5

1.25

8

12.5

1.63

12

14.5

1.8

15

15.5

1.9

18

16.5

2.0

Circular polarization can be obtained by using crossed Yagi-Uda antennas: a pair of antennas mounted on a common boom with their elements set at right angles. The two antennas must be fed in phase quadrature or be fed in phase and mutually displaced by a quarter wavelength along the boom. (Using the in-phase method the equality of the phases and current amplitudes on the two components is less disturbed by the VSWR of the individual antennas as long as they are identical.) There has been some interest in slow-wave end-fire arrays that use long, closed forms for their elements, such as rings or squares. They have not proved popular in practice, probably because they have narrower bandwidths than an optimal Yagi-Uda antenna and are more complex to manufacture. The domestic TV receiving antenna market is sometimes subject to fashion, and some interesting local variants are seen in cities round the world. Log-periodic antennas are widely used for applications in which a large frequency bandwidth is needed. The gain of a typical VHF or UHF log-periodic antenna is about 10 dBi, but larger gains can be obtained by arraying two or more antennas. The disadvantage of all log-periodic designs is the large physical size of an antenna with only modest gain. This is because only a small part of the whole structure is active at any given frequency. The most common design used on the VHF-UHF bands is the log-periodic dipole array (LPDA). (See Chapter 13 and Kraus and Marhefka.17) After selecting suitable values for the design ratio t and apex angle a, the designer must decide on the compromises necessary to produce a practical antenna at reasonable cost. The theoretical ideal is for the crosssectional dimensions of the elements and support booms to be scaled continuously along the array; in practice, the elements are often made in groups by using standard tube sizes, and the support boom is often of uniform cross section. The stray capacitances and inductances associated within the feed region are sometimes troublesome, especially in the UHF band, and can be compensated only by experiment. The coaxial feed cable is usually passed through one of the two support booms to avoid the need for a wideband balun. Printed-circuit techniques can readily be applied to LPDA design in the UHF band and above, as the antenna is easy to divide into two separate structures that can be etched onto two substrate surfaces. At the lower end of the VHF band the dipole elements may be constructed from flexible wires supported from an insulating catenary cord. A typical well-designed octave-bandwidth LPDA has a VSWR less than 1.3:1 and a gain of 10 dBi. Log-periodic antennas have been widely used both as TV receiving antennas and also as transmitting antennas at low-power relay stations in the VHF and UHF broadcast bands. LPDAs are produced commercially for surveillance and measurement applications with bandwidths exceeding a decade. Other log-periodic designs that can be found in Chapter 13 are less common in commercial application in the VHF/UHF bands, probably for reasons of cost and windload.



CHAPTER TWENTY-NINE

Helices A long helical antenna has an easily predicted performance and is simple to construct and match. A VSWR as low as 1.2:1 can be obtained fairly easily over a frequency bandwidth of 20%, and wider bandwidths are possible if the helix is tapered or stepped in diameter. Conductive spacers may be used to support the helical element from the central support boom, so the antenna can be made very robust. At higher frequencies it may be more convenient to support the helical element by winding a conducting tape onto a dielectric rod or tube. The maximum gain that can be obtained from a single helix is limited by the physical length that can conveniently be supported, typically ranging from 12 dBi at 150 MHz to 20 dBi at 2 GHz. Helices can be arrayed for increased gain; to obtain correct phasing the inputs must be driven in phase, and the start position of each helix in the array must be the same. Low gain helices—in particular, short quadrifilar helices—find application as reflector feed units and as antennas for small satellite terminals.18 Ceramic-loaded quadrifilar helices have been used as receiving antennas for GPS systems.19 For further information on the design and performance of helices see Chapter 12. Bowtie Dipoles and Reflectors Arrays comprising a column of bowtie dipoles mounted in front of a reflecting screen are common as horizontally polarized domestic TV receiving antennas in some countries. The triangles can be approximated with wire mesh, provided the mesh spacing is less than onetenth wavelength, or by a fan of rods connected at the feedpoint. Flare angles between 60º and 80° combined with a half-length up to 0.58l provide satisfactory input impedance and radiation patterns. To provide more gain and an increased front-to-back ratio, dipoles are usually stacked vertically and mounted approximately 0.25l in front of a reflecting screen made from parallel rods spaced 0.1l apart (at fmax). In this configuration the dipoles are fed by an open balanced transmission line. The measured gain for one and two dipoles over a flat screen is also shown in Figure 29-4. Commercially available antennas typically have a VSWR less than 2.0, a front-to-back ratio greater than 15 dB, and sidelobe levels less than 13 dB below the peak gain over 90% of the UHF band. Corner Reflectors Well-designed corner-reflector antennas are capable of providing high gain and low sidelobe levels, but below 100 MHz they are mechanically cumbersome. Before using a corner reflector, make sure that the same amount of material could not be more effectively used to build a Yagi-Uda antenna, or perhaps a pair of them, to do the job better. In the UHF band, corner reflectors can be very simply constructed from solid or perforated sheet and a variety of beamwidths and back-to-front ratios obtained by the choice of the apex angle, spacing of the dipole from the vertex, and the width of the reflector. The apex of the corner is sometimes modified to form a trough (see Figure 29-5). The provision of multiple dipoles extends the antenna aperture and increases the available gain. Panel Antennas An antenna that comprises a reflecting screen with simple radiating elements mounted over it, in a

FIGURE 29-3 Bowtie dipole



29-21

FIGURE 29-4 Gain characteristics: (a) Bowtie dipole (a = 70°, A = 190 mm); (b) Bowtie dipole in front of screen (550 × 480 mm); (c) Two tiers stacked 600 mm apart in front of a screen

broadside configuration, is generally termed a panel antenna. An array may comprise one or more panels connected together. Typical panels use full-wavelength dipoles, half-wave dipoles, or slots as radiating elements (see Figure 29-6). For some applications they have advantages over Yagi-Uda antennas: ●

●

●

●

More constant gain, radiation patterns, and VSWR over a wide bandwidth—up to an octave More compact physical construction—the phase center is maintained closer to the axis of the supporting structure, providing better control of the azimuth radiation pattern Very low coupling to the mounting structure Low side lobes and rear lobes

FIGURE 29-5 Corner and trough reflectors



CHAPTER TWENTY-NINE

FIGURE 29-6 Panel antennas: (a) Two full-wave dipole elements; (b) Two batwing slot elements; (c) Skeleton-slot elements

Panel antennas for frequencies in the UHF band lend themselves to printed-circuit design methods, as the radiating structures, feedlines, and matching system may all be produced by stripline techniques. At lower frequencies the radiating elements are often mounted at voltage minimum points using conducting supports, so a strong, rigid construction can be produced. A solidly built but lightweight panel for a military application is shown in Figure 29-7a. Here an all-welded aluminum frame and a skeleton-slot radiator are used so that the antenna will resist rough use in the field. Panels are often used as building blocks for complex arrays for radio and TV broadcasting. Paraboloids The design of a high-gain antenna may be reduced to a problem of illuminating the aperture necessary to develop the specified radiation patterns and gain. The size of the aperture is determined only by the gain required, whatever type of element is used to fill it. As the cost of the feed system and the radiating elements doubles for each extra 3-dB gain, a stage is reached at which it becomes attractive to use a single radiating element illuminating a reflector that occupies the whole of the necessary antenna aperture. The design task is reduced to choosing the size and shape of the reflector and specifying the radiation pattern of the illuminating antenna. If the antenna aperture is incompletely filled or its illumination is non-uniform, the gain that is realized decreases. The ratio of the achieved gain to the gain obtainable from the same aperture when it is uniformly illuminated by

FIGURE 29-7 Robustly constructed antennas for military use: (a) Skeletonslot-fed panel (225–400 MHz); (b) Grid paraboloid (610–1850 MHz) (Courtesy of Jaybeam Ltd.)



29-23

TABLE 29-3 Typical Paraboloid-Antenna Configurations Frequency (MHz)

Diameter (m)

f/d ratio

Construction

Feed Type

200

10.0

0.5

Mesh paraboloid

4-element Yagi-Uda

700

3.0

0.25

Solid skin

Dipole and reflector

900

7.0

0.4

Perforated steel sheet

Horn

610–960

1.2

0.25

Grid of rods

Slot and reflector

1500

2.4

0.25

Solid skin

Dipole and disk

1350–2500

1.2

0.25

Grid of rods

LPDA

lossless elements is termed the aperture efficiency of the antenna. In a receiving context, this quantity represents the proportion of the power incident on the aperture that is delivered to a matched load at the terminals of the antenna. In the VHF and UHF bands, a reflector may be made of solid sheet, perforated sheet, wire mesh, or a series of parallel curved rods. As the wavelength is large, the mechanical tolerance of the reflector surface is not very demanding, and various methods of approximating the true surface required are possible. Table 29-3 indicates some of the combinations of techniques currently in use and illustrates the diversity of the methods that are successful for various purposes. Grid paraboloids are attractive to produce because the curvature of all the rods is exactly the same; only their length varies across the antenna. A typical example is shown in Figure 29-7b. The main deficiency of grid paraboloids is the leakage of energy through the surface, restricting the front-to-back ratio that can be achieved. For example, at 1.5 GHz a front-toback ratio of –30 dB is a typical limit. If a greater front-to-back ratio is needed, it may be possible to adopt an offset geometry. Alternatively, the reflector bars can be extended in depth, or an orthodox continuous skin of solid or perforated sheet can be used in place of the grid; the consequent increase in weight and wind-loaded area must be accepted as a necessary penalty for improved electrical performance. Electrically small reflectors have poor radiation pattern performance because the feed has physical dimensions independent of the reflector diameter, so feed blockage is more troublesome and leads to high sidelobes. Radomes are frequently fitted to feeds or complete antennas in order to reduce the effects of wind and snow. They may be made from fiberglass or in the form of a tensioned membrane across the front of the antenna. In severe climates it is possible to heat a radome with a set of embedded wires, but this method can be applied only to a plane-polarized antenna. Point-to-point links using tropospheric-scatter propagation require extremely high antenna gains and typically use a reflector that is an offset part of a full paraboloidal surface constructed from mesh or perforated sheet. Illumination is provided by a horn supported at the focal point by a separate tower. For a full discussion of the design of reflector antennas refer to Chapter 15.

29.8 OMNIDIRECTIONAL ANTENNAS Simple Low-Gain Antennas The simplest types of antennas will provide truly omnidirectional azimuth coverage only when mounted in a clear position on top of a tower. Figure 29-8 shows standard configurations for ground-plane and coaxial dipole antennas and demonstrates that these forms are closely related. They are cheap and simple to



CHAPTER TWENTY-NINE

FIGURE 29-8 Low-gain base-station antennas: (a) Standard ground plane with radials; (b) Ground plane with sloping radials; (c) Ground plane with closed ring; (d) Coaxial dipole

construct, and may be made to handle high power. Exact dimensions must be determined by experiment, as the stray inductance and capacitance associated with the feedpoint insulator cannot be neglected. The use of a folded feed system can provide useful mechanical support and gives better control over the antenna impedance. (Both the relative diameter of the feed and grounded conductors and the point of their interconnection can be varied.) The operational bandwidth of the coaxial dipole d depends critically on the characteristic impedance Z0, of the coaxial section formed by the feedline (radius r) inside the skirt (radius R). If this section has too small a Z0, radiating currents will flow on the outside of the feeder line unless the skirt length is exactly l/4. The impedance, gain, and radiation pattern of the antenna then become critically dependent on the positioning of the feedline on the tower, severely limiting the useful bandwidth of the antenna. Discone Antennas The discone and its variants are the most commonly used low-gain wideband antennas. The useful lower frequency limit occurs when the cone is a little less than l/4 high, but the upper frequency limit is determined almost entirely by the accuracy with which the conical geometry is maintained near the feedpoint at the apex of the cone. Discones may be made with either the disk or the cone uppermost. The support for the upper part of the antenna usually takes the form of low-loss dielectric pillars or a thinwalled dielectric cylinder, fitted well outside the critical feed region. Variants of the basic discone use a bicone in place of the conventional cone and replace the disk by a short cone with a large apex angle. At the lower end of the VHF band discone antennas can be mounted at ground level, so a minimal skeleton disk that couples to the ground may be used if some loss of efficiency and the propagation effects associated with a low antenna elevation can be accepted. Collinear Arrays The ground plane and coaxial dipole have about the same gain as a halfwavelength dipole. When more gain is needed, the most popular omnidirectional antennas are simple collinear arrays of half-wave dipoles. The original array of this type is the Franklin array shown in Figure 29-9a. This design is not very convenient owing to the phase-reversing stubs that project from the ends of each half-wave radiating section, but various derivatives are widely used. The arrangement at Figure 29-9b uses non-inductive meander lines to provide phase reversal; that at c is a rearrangement of the original, while those at d and e use coaxial line sections. Arrangements such as these may be mounted in fiberglass tubes to provide mechanical support, and the designs at b, c, and f are suitable for production by printed-circuit techniques. In each of these arrangements, the elements are connected in series; an input-matching section transforms the input impedance of the lower section, which may be l/2 or l/4 long, to 50 Ω. A set of quarter-wavelength radial



(a)

(b)

(c)

(d)

(e)

29-25

(f)

FIGURE 29-9 Collinear dipole arrays: (a) Franklin array; (b) Array with meander-line phase reversal; (c) Array with transposed coaxial sections; (d) and (e) Alternative coaxial forms; ( f ) Planar version

elements or a quarter-wavelength choke is used to suppress currents on the outside of the feeder cable. The maximum useful gain available from these arrays, typically about 10 dBi, is limited by two factors: ●

●

There is mechanical instability in a very long antenna with a small vertical beamwidth. The available excitation current diminishes away from the feed as a result of the power lost by radiation from the array.

In the case of the coaxial-line designs, each section is shorter than a free-space halfwavelength so that the correct phase shift is obtained inside the section. The examples shown would typically provide a gain of 9 dBi at the design frequency. The useful bandwidth of series-fed collinear antennas is inherently narrow because of the phase error between successive radiating sections that occurs when the frequency is changed from the design frequency. The typical behavior of the major lobes of the vertical radiation pattern of these arrays is shown in Figure 29-10. A further problem with long arrays is that as the array length is increased, the series connection of the elements results in an increased input impedance; as the transformation ratio of the input feed network increases, so the input impedance bandwidth is reduced. Parallel-Fed Arrays Much greater control is obtained by using an array of fat dipoles with an internal, branched, parallel-connected feed system. Arrays of this type provide stable gain, radiation patterns, and input VSWR over wide bandwidths. A well-optimized eight-element array is able to provide acceptable gain (~10 dBi), radiation patterns, and input VSWR (< 1.7) over the band 225–400 MHz and is suitable for ground-to-air communications; at

FIGURE 29-10 Vertical radiation pattern of a typical end-fed array



CHAPTER TWENTY-NINE

FIGURE 29-11 Pole-mounted dipoles: (a) Inline; (b) Four dipoles spaced around a pole; (c) Eight dipoles spaced around a pole

this frequency the diameter of a structural GRP radome is around 250 mm. Antennas may be designed using a combination of parallel and series feeding; for example, an array can be designed as two end-fed arrays end-to-end, and fed in the center where they join. As requirements for elevation pattern stability increase, the array must be further subdivided into parallel-fed units. Dipoles on a Pole A simple offset pole-mounted array is shown in Figure 29-11a. This will provide a gain of about 10 dBi in the forward direction but typically only 4 dBi rearward, depending on the pole diameter and the spacing between the dipole and the pole axis. An attempt to avoid this problem is shown in Figure 29-11b, but this type of antenna has distorted vertical radiation patterns caused by the phase shifts that result from the displacement of the dipoles; gain is also reduced to about 6 dBi for the four-element array shown. The solution in Figure 29-11c, in which dipoles are placed in pairs and are cophased, is more satisfactory, as the phase center of each tier is concentric with the supporting pole. However, the antenna is relatively expensive, as in this arrangement eight dipoles provide only 6-dB gain over a single dipole. One possibility is to use the inline stacked array in Figure 29-11a and to place the base station toward the edge of the service area. The rearward illumination may be improved if the spacing between the dipoles and the pole is optimized for the pole size and operating frequency. Analytical solutions to the azimuth pattern are available, and simpler computer programs provide results in good agreement with measurements. When designing the feed networks for multi-element arrays of this type, take care to allow for the effects of mutual impedances, especially when unequal currents or asymmetric geometries are used. Antennas on the Body of a Tower Figure 29-12a shows a measured horizontal radiation pattern for a simple dipole mounted from one leg of a lattice tower 2-m face width. The distortion of the circular azimuthal pattern of the dipole is very typical and is caused by blocking and reflection from the structure. By contrast, Figure 29-12b shows what can be achieved by an antenna comprising three dipole panels mounted on the same structure. The penalty of adopting this improved solution lies in the cost of the more complex antenna, so before an optimum design can be arrived at, the value of the improved service must be assessed.



29-27

FIGURE 29-12 Typical azimuth patterns of (a) VHF dipole mounted off one leg of a triangular mast and (b) three dipole panels mounted on the same structure (linear field plot)

The horizontal radiation pattern of a complete panel array is usually predicted from measured complex radiation-pattern data for a single panel, using a suitable computer program. For each azimuth bearing, the angle from each panel axis is found, and the relative field in that direction is obtained. The radiated phase is computed from the excitation phases and physical offsets of the phase centers of the individual panels. Depending on the cross-sectional size of the structure, the most omnidirectional coverage may be produced with all panels driven with the same phase or by a phase rotation around the structure; for example, on a square tower the element current phases would then be 0, 90, 180, and 270°. When phase rotation is used, the individual elements may be offset from the centerlines of the faces of the structure to give a more omnidirectional azimuth pattern, as in Figure 29-13. A panel array comprising four tiers, each of four panels, is an expensive installation, but if properly designed, it can have a useful bandwidth of as much as 25%. This allows several user services to be combined into the same antenna, each user having access to a very omnidirectional high-gain antenna. Groups of panels may be arranged and fed to produce an azimuth pattern tailored to the arbitrary requirements of the service area or to provide nulls necessary to meet co-channel protection objectives. Panel arrays are discussed in further detail with reference to their use by broadcasting services in Section 29.9. VHF/UHF base-station antennas are sometimes situated on the bodies of large towers, perhaps up to 10 m in diameter. It is not economically possible to provide smooth omnidirectional coverage from such a large structure. However, by use of some lateral FIGURE 29-13 Plan arrangement of an omnidirectional panel thinking, it may be possible to array on a large tower



CHAPTER TWENTY-NINE

make a virtue out of a necessity. It is often possible to provide solid coverage of an arc of 270°, or to use two frequency channels, providing overlapping coverage in the region of greatest traffic. If the station operator can guarantee the rearward null in the azimuth patterns at both frequencies, the frequency planner may benefit, and the operator may get two channels where they matter most. Special-Purpose Arrays For applications in which the largest possible coverage must be obtained, the azimuth radiation pattern of the antenna must be shaped to concentrate the transmitted power in the area to be served, for example, an airway, harbor, or railroad track. Energy radiated in other directions is wasted and is a potential cause of interference to others. Antennas with cardioidal azimuth radiation patterns are useful for a wide range of applications. Simple two-element arrays (dipole plus passive reflector) or dipoles mounted off the face of a tower may be adequate, but a wider range of patterns is available if two driven dipoles are mounted on a single supporting boom and excited with suitably chosen currents and phases. When a signal must be laid down over an arbitrarily shaped area of terrain, Yagi-Uda arrays or LPDAs may be arranged as at Figure 29-2f and g. Due allowance must be made for the separation of the phase centers of the antennas when computing the radiation patterns. As an approximate guide, the phase center of a Yagi-Uda antenna lies one-third of the way along the director array, measured from the driven element. Further tiers of antennas may be used to increase the gain of the system without modifying the azimuth radiation patterns.

29.9 ANTENNAS FOR BROADCAST SERVICES AND BASE STATIONS One of the characteristics shared by broadcasting antennas and mobile base station antennas is that they are usually linear vertical arrays of some form of standard radiating element. This format derives from the requirement to lay down a signal over some defined area of the surrounding terrain while concentrating the radiated energy into a well-focused beam in the elevation plane; radiation skyward is wasted, so the objective is to concentrate as much energy as possible toward the outer edges of the served area (sometimes, but not necessarily, the horizon). The characteristic of linear arrays is that the radiation patterns obtained in the vertical (elevation) plane and the horizontal (azimuth) plane are mathematically separable. That is, the elevation pattern is determined by the way in which the radiating elements are physically arranged and electrically fed in the vertical plane—the nature of the individual elements only matters to the extent that the pattern is the product of the elevation pattern of the individual elements and the array factor generated by the chosen spacing and complex feed currents. The azimuth pattern is independent of the way in which successive vertical tiers are fed and only depends on the arrangement of a single tier (bay) of elements. It is important to appreciate that this separability collapses if the main beam of the antenna is deflected downward away from the horizontal direction, or if the successive tiers of the antenna do not have identical azimuth patterns. The main difference between broadcasting and base station antennas is that in broadcast practice it is usual for antennas to have wider and more complex azimuth patterns than base station antennas. Before looking at the methods for the design of these antennas it is worth reviewing some basic array theory. First consider an array of elements spaced in the vertical plane, providing omnidirectional azimuth coverage. If the aperture is large, say more than eight wavelengths, the total directivity is only a function of the length of the array. We can fill it



29-29

with elements with low directivity (isotropic radiators) or higher directivity (for example, vertical dipoles one wavelength long), but as long as we fill the aperture with elements, the directivity will be the same. (This doesn’t apply in the same way to short arrays because the effective aperture can “overhang” the physical aperture by a more significant proportion.) As a rule of thumb, the maximum directivity of an omnidirectional array will be about 1.15 times per wavelength of vertical aperture. If the aperture is underfilled—the elements are too far apart—the result will be lower directivity because energy is lost from the main beam into characteristic high elevation sidelobes (grating lobes). Overfilling the aperture with elements produces no benefit in directivity and increases the coupling (mutual impedance) between adjacent elements, making it more difficult to control the excitation of the array. In the azimuth plane the situation is different. If we select a radiating element whose property is to uniformly illuminate a 180º sector with no signal at all in the other half-space, then it is clear that such an element has a directivity of two (3 dB). Because directivity is properly defined as a 3D property, this ratio is commonly known as the azimuth maxto-mean ratio. It is numerically equal to the area of the circle circumscribing the polar radiation pattern when plotted in linear field coordinates, E(q ), to the area of the plotted radiation pattern itself (because both power and area are proportional to E2). This makes it very easy to visualize what is happening. The maximum total directivity of a vertical array is the product of the directivity in the elevation plane times the azimuth max/mean ratio. Azimuth Pattern Shaping Each tier of an array may comprise a single radiating element; this is common in base station practice and for simple broadcast antennas. The desired pattern is often obtained by arraying elements against a mounting pole, or against a reflecting plane when more suppression of radiation is needed in the rearward direction. Where the required azimuth pattern cannot be achieved using a single radiating element (with or without a reflector) each tier is constructed from several elements, each often constructed in the form of a panel comprising an integrated radiating structure and reflector. Panels are usually designed to cover wide frequency bandwidths and to maintain very consistent radiation pattern, polarization, and impedance over the operating band. Typical panels cover entire broadcast bands. Designing arrays in this manner has the advantage that by using standard, highly optimized, mass-produced panels the antenna designer can construct a wide variety of azimuth pattern shapes. The adoption of sector coverage for mobile radio base stations has removed the requirement for separate panels, but the design of the radiating element has the same constraint of requiring constant performance over wide bandwidths. To reduce the mutual impedance between adjacent elements individual elements are often surrounded by fences or cavities. Elevation Pattern Shaping There is little point in directing the elevation maximum of our vertical array in exactly the horizontal plane: in flat terrain half the radiated power will be lost over the head of any ground-based receiver. The further we get from the antenna the more the curve of the earth will result in the main beam passing overhead. The main beam is therefore usually tilted downward, typically so the upper –3-dB point lies in the horizontal plane; the exact tilt may depend on the relative elevation of the antenna and the served users. A further result of tilting the beam downward is to reduce the level of unwanted illumination of users located beyond the effective served area. The served area is always surrounded by a region covered by signals too weak to use but too strong to ignore; the size of this zone is reduced when beamtilt is used. For ground-to-air services a small upward beamtilt is sometimes applied.



CHAPTER TWENTY-NINE

A column of elements with identical currents in each radiating element creates an elevation radiation pattern with a simple (sin x)/x form. This pattern is generally modified to create some additional shaping by feeding the tiers of the array with currents of different phases and/or amplitudes. There is no unique solution to the creation of a radiation pattern with some arbitrary shaping (as we generally need to match only a requirement for amplitude and not for both amplitude and phase), so a variety of procedures can be adopted to derive a suitable set of element currents. Practical considerations include ensuring that the chosen function does not require very different currents in adjacent elements, which will exaggerate the effects of mutual impedances. The radiation pattern must not be sensitive to small changes in some of the array currents, as this may make it difficult to achieve the required pattern over a wide frequency band. A simple method of filling nulls and suppressing sidelobes is based on the principle of superposition. Given that a uniform array has an elevation pattern in the form of a sinc function, supplementary excitations can be added to the principal excitation with chosen amplitudes and relative phases. To fill a null we apply a sinc function with its maximum directed at the angle of the null and with its phase in quadrature with the principal excitation; to suppress a sidelobe an additional sync function is added with its maximum in the direction of the sidelobe and in antiphase with the main excitation. Additional sync functions can be added to fill successive nulls or to suppress as many sidelobes as required. There is some interaction between the added sinc functions because each has its own sidelobes at –13 dB relative to its maximum, but by using a computer program the method can be implemented in an interactive manner. At the end of the optimization process the element currents associated with each sync function are added vectorially to give the complex current that will be needed to generate the specified pattern. When assessing the performance of an array it is useful to obtain a sample of the radiating currents from each element by using inline couplers or external sampling loops. The currents can be displayed in polar coordinates using a vector network analyzer to allow the designer to verify that the actual currents are close to the design currents and to maintain the correct relationship within some acceptable limit over the operating frequency band. The sensitivity of the radiation pattern to small variations in the element currents can be investigated using a Monte Carlo analysis in which random variations in amplitude and phase, with a known standard deviation, are applied to the array currents, and the resulting patterns are evaluated for compliance with the target specification. The most common reason for element currents deviating from their design values is the imperfect impedance match of the elements; this causes the power division at each junction to deviate from the expected value in a manner that is often strongly frequency dependent. The design of many radiating elements in broadcast and base station arrays is often governed by the requirement that individual elements (or element groups) have a sufficiently low VSWR to allow radiation pattern constraints to be met over extended bandwidths. The use of Wilkinson power dividers can limit the extent of current variation, but unless good matching is achieved this may be at a price of losing RF power into balancing loads. Impedance Characteristics The maximum permissible input VSWR of an antenna is determined by the effect that a significant reflection would have on the transmitted signal. The applicable criterion is commonly more stringent than considerations of transmitter currents and voltages or of maximum power transfer to the antenna. With respect to the input impedance of the whole array, there are several important considerations: The radiation of signals with relative time delays If we examine the typical arrangement shown in Figure 29-14, we see that a transmitted pulse, which may be part of any analog or digital transmission, travels from the transmitter to the antenna, experiencing some loss and transit time in the interconnecting cable. If the antenna has a return loss L db,



∼

Cable with attenuation AdB and transit time t sec

29-31

Antenna with return loss L dB

Transmitter with reverse reflection coefficient r dB FIGURE 29-14 Transmitter feeding a mismatched load

a signal L dB below the forward signal at the antenna now moves along the cable in the reverse direction. After suffering a delay t and attenuation A, the signal encounters the output port of the transmitter. In general the impedance looking back into a transmitter is very far from 50 Ω, so much of the signal is now reflected with reflection coefficient r dB. After again passing along the cable, the signal originally reflected at the antenna is reradiated with a relative level of –(L + 2A + r) dB and a delay of 2t relative to the original signal. The significance of both the amplitude and the relative delay of this secondary signal depends on the signal format being transmitted. In analog TV and digital systems using high-order digital modulation formats it is the essential determinant of the required antenna reflection performance, especially for antennas fed by long, lowloss cables. Frequency-dependent load impedance at the transmitter If the antenna input impedance differs from the characteristic impedance of the main feeder, the impedance seen by the transmitter varies with frequency. (The impedance of the antenna forms circles round the center of the Smith Chart.) The result of a transmitter working into an impedance that varies with frequency is often to distort the output waveforms in both phase and amplitude. Again the effect and its acceptable limits depend on the signal format. With some systems it is usual to adjust the transmitter characteristics to compensate for some degree of distortion caused by the varying load impedance, but the extent to which this can be done is limited, especially when the antenna is connected by a long, low-loss cable. Cable power ratings The effective power rating of a cable is reduced when it is operating at a high VSWR, as the high-current points in the standing wave cause hotspots on the cable; for this reason the cable rating must be reduced by a factor of 1/VSWR. These effects usually define the required impedance characteristic of the whole antenna. The effect of reflection loss only becomes significant at high values of VSWR, the ruleof-thumb values being 0.5 dB at 2:1 and 1 dB at 2.6:1 (return losses of 9 dB and 7 dB, respectively). There is a further reason why the VSWR of individual elements needs to be controlled. If two loads are connected in parallel at a transmission line junction, the current division at the junction is in inverse ratio to the (complex) impedances seen at the inputs to the separate branches. If we join two mismatched antennas to a junction through cables of equal length, this presents no particular problem. Although mismatched, the impedances are equal so the current divides equally in amplitude and phase. If the lines connecting the loads to the junction are of unequal length—perhaps because we require a phase shift between the currents in the connected elements—the impedances presented at the junction are no longer equal because they lie at different points on a circle round the center of the Smith Chart. The result is that the load currents will be neither equal nor of the expected relative phase, modifying the radiation pattern of the arrayed elements in a manner we may not have expected. This imposes a new criterion for the input match of the radiating



CHAPTER TWENTY-NINE

elements, depending on the accuracy with which design currents must be achieved and maintained over extended bandwidths in order to create the required radiation patterns. If the equality of the currents must be preserved despite mismatched impedances, the use of a hybrid junction or Wilkinson splitter will reduce the unwanted effect of the mismatch on the power division. In situations in which antenna matching is very critical it is common to use a variable transmission line device to center the complex impedance plot of the antenna relative to the exact Zo of the coaxial feedline. The variable components may take the form of short sleeves (“slugs”) fitted to the inner conductor of a section of transmission line—the equivalent of small patches on a microstrip line. Where close impedance matching is needed at a number of different frequencies, a long section of transmission line is used and is fitted with several slugs at different computed positions. Designing Radiating Panels The commonest designs for radiating panels provide an azimuth beamwidth of 90º (total) at –6 dB relative to boresight. When two panels driven with equal co-phased currents are mounted close to one another on the sides of a square tower, the field in the diagonal direction is equal to the field in the main beam direction of each panel. A successful panel must have an azimuth beamwidth that is very stable with frequency, and the level of any subsidiary lobes must be very low. Typical panels comprise two full-wave dipoles mounted against a reflecting screen, usually made from solid sheet above about 400 MHz and formed from conducting rods at lower frequencies. Because antenna installers use horizontal bars as handholds and climbing steps, it is common to make these bars from hot-galvanized mild steel. In the UHF band the design and production of wideband panels with well-controlled azimuth beamwidth and low VSWR (typically < 1.1:1 over 470– 860 MHz) presents an interesting challenge. Combining Panels into Groups The most usual feed method is to connect all the panels that form one tier of an array together into one power divider, constructed from coaxial line sections for use at high powers or various forms of microstrip for low powers. The input ports of each tier are then combined, usually in a branched tree pattern. For base station antennas a single feed is usually provided to the whole array, while in broadcast practice it is usual to provide separate inputs to each (vertical) half of the whole array, and to connect the antenna to the final stage of combining at ground level. This configuration gives access to the two separate half-antennas, allowing some redundancy in the case of failure of some part of the array, and access for maintenance when one half array is disconnected and grounded—in these cases the gain of the antenna is obviously reduced when only one half is driven, but loss of some coverage is better then no transmission. It is obviously also possible to combine a broadcast array by interconnecting all the panels on each face with a power division network and then to interconnect the faces in a final stage of combining. This method is not favored because it uses more cable, and a single fault may grossly modify the azimuth radiation pattern. When multiple radiating elements on each panel are used to create circular polarization it is usual to combine the elements of each panel together before feeding them from the main feed network. For analog TV transmissions it is necessary to achieve a very low VSWR across each transmission channel, and arrays often have extremely wide bandwidths. For this reason the phase rotation method shown in Figure 29-15 has been extensively used. It can be extended by feeding adjacent tiers in phase quadrature (and rotating them mechanically by 90º) and repeating this with successive groups of two and four tiers. Complete half antennas are often fed in phase quadrature without a corresponding mechanical rotation to



29-33

provide elevation beamtilt equal to half the total elevation beamwidth. These methods are not so easily applied when the azimuth radiation pattern is directional. Directional radiation patterns are achieved by a combination of several methods: ●

●

●

Adjustment of the amplitude and phase of panels forming each tier of the array Adjustment of the physical positions of the panels in each tier Use of unequal number of panels on different faces of the array (in this case the physical configurations of different tiers are not identical)

Azimuth pattern requirements are driven by a mixture of the minimization of interference with stations operating on the same or adjacent channels (which may be contained in international regulatory requirements), topography, and the distribution of the target population. Azimuth patterns can be predicted by computer programs (which often make use of measured data from individual panels) but usually need validation by practical measurement—at least of an appropriate subsection of the complete array. The most common sources of difference between predictions and measurements occur when: ●

●

●

The array has substantial physical displacement of panels—potentially allowing blockage of one panel by another. Current paths in the screening surfaces are physically interrupted; such a problem occurs in a horizontally polarized array if the screening panels are not bonded to one another at each corner, when the gaps at each corner operate as slot radiators that were not included in the computer prediction. Radiating elements with imperfect matching are connected together by cables of unequal length (see above).

High-power broadcast arrays require very careful attention to mechanical design details, and the advice in Table 29-1 needs to be thoroughly applied. The accumulation of ice and snow will change element matching with the result that radiation patterns, gain, and input VSWR will all be degraded. High voltages at element feedpoints are likely to create dielectric heating, and any insulators are likely to suffer surface tracking both from exposure to the weather and accumulation of contaminants from urban rain, birds, and salt spray at coastal locations. This problem can be addressed by the use of a radome, fitted either over the whole array or over the elements or their feedpoints—in this last case the radome must be sufficiently large that it does not experience strong surface fields, or the result will be failure by surface tracking. Omnidirectional Broadcasting Antennas Broadband high-power omnidirectional antennas are usually built from tiers of four (occasionally three) panels that may have linear or circular polarization. Depending on the cross-sectional size of the structure, the most uniform omnidirectional coverage may be produced with all panels driven with the same phase or by a phase rotation around the structure; for example, on a square tower the element current phases would then be 0º, 90º, 180º, and 270°. When phase rotation is used, the individual elements may be offset from the centerlines of the faces of the structure to give a more omnidirectional azimuth pattern, as in Figure 29-13. Omnidirectional vertically polarized antennas can be produced by mounting two vertical dipoles either side of a supporting pole, carefully choosing the spacing to provide a uniform azimuth pattern. If the support is too large (in wavelengths), then three or more dipoles may be needed to provide uniformity. Horizontally polarized omni antennas can be constructed using a turnstile format in which two separate elements are arranged at right angles and fed in phase quadrature.



CHAPTER TWENTY-NINE

An alternative method is to create a square loop from four horizontal half-wave dipoles fed in phase. This produces a good omnidirectional azimuth pattern but the large capacitive coupling across the corners of the array limits the achievable impedance bandwidth (see Figure 29-15a). Circularly polarized antennas are sometimes based on loop-dipole derivatives and short radial-firing multiple-start helices. These are often mounted off the side of the supporting structure, providing quasi-omni patterns with degraded polarization circularity. An omnidirectional array can be designed using broadband panels with crossed-dipole elements, but this arrangement is costly and has a very substantial weight and windload. The pole-mounted Lindenblad array20 (see Figure 29-15b) is formed from four half-wave dipoles mounted on a square that is a half-wavelength across, with the elements lying at 45° to the horizontal and driven with equal cophased currents. When viewed from the side of the square, it is easily seen that the geometry ensures that the vertical and horizontal field components are equal in magnitude and have a quarter wavelength relative to the physical displacement of their phase centers. The arrangement provides an omnidirectional azimuth pattern with low ripple and good polarization purity. An array comprising multiple tiers, centrally mounted on a supporting pole, can provide high gain and low VSWR across the whole FM radio band 97.5–108 MHz and is suitable for high-power multi-channel operation. A short four-start helix can be imagined as a derivative of a Lindenblad antenna using curves radiating elements. A modified form of Lindenblad array is sometimes used in which the four dipoles are fed with phase rotation, the spacing between dipoles reduced, and the tilt angle changed to around 30°. This provides hemispherical coverage but is not suitable for terrestrial broadcast use. Feeding Multiple Arrays Simple low-power arrays operating at frequencies below 500 MHz are usually fed using a coaxial-cable branching network (see Figure 29-16). Stripline power dividers are attractive for applications above about 500 MHz and can be designed to provide arbitrary power division ratios and any required number of ports; they provide a high level of reproducibility in volume production and can be designed using readily available computer software. If an array is to operate at high mean input power, it may be necessary to use large-diameter fabricated coaxial transformers.

(a)

(b)

FIGURE 29-15 (a) A horizontally polarized square loop comprising four half-wave dipoles; (b) A circularly polarized Lindenblad array in which the dipoles are rotated by 45º



29-35

FIGURE 29-16 Simple branching feeder systems: (a) Two-way; (b) Four-way; (c) Two-way, compensated; (d) Two-way, high-power

Multiway broadband power dividers are usually designed using multiple quarter-wave sections to achieve a very low input reflection coefficient over a very wide bandwidth. The general principle used is illustrated in Figure 29-17. At the common point of the branching output lines the impedance is low and requires to be transformed up to 50 Ω. If this is done in a single step over a wide bandwidth, the transformed impedance will take the form of an arc on the Smith Chart as in Figure 29-17a. A first transformation is made to an intermediate point on the Smith Chart, and this is followed by a second transformation with a similar impedance ratio. The action of the second quarter-wavelength is again to wind the intermediate impedance into an arc in which the angle traversed is proportional to frequency; this compensates the behavior of the first transformation, and the result is a very small impedance plot lying close to 50 Ω (see Figure 29-17b). When large numbers of branch lines are to be fed it is good practice to transform the outgoing line impedances up before connecting them in parallel, to avoid the need to make lines with extremely low Zo. Where large power division ratios are needed the outgoing impedances can be transformed up so

FIGURE 29-17 (a) The dispersive effect of a simple quarter-wave section over a wide bandwidth. (b) A two-stage transformation has been used in which the effect of the second section is to reduce the dispersion caused by the first section.



CHAPTER TWENTY-NINE

(a)

(b)

FIGURE 29-18 (a) Panel arrays of full-wave dipoles for 174–230 MHz and (b) bent half-wave dipoles for 87.5–108 MHz. The rigger is a reminder of the need to plan for safe access and to ensure that equipment cannot be damaged by being used as footsteps or handles. (Courtesy of Kathrein-Werke KG)

the impedances at the common point have the required ratio and are stable with frequency; this ensures that the power division ratio is stable. Designs of this type can be realized in coaxial or microstrip lines.21–24 Alternative designs can be created by thinking in lumped element terms and using short lines of high Zo (inductors) and of very low Zo (capacitors). This technique can produce some very compact splitters with wide frequency bandwidths. Wilkinson power dividers have characteristics similar to those of a hybrid junction. Figures 29-18 and 29-19 show typical examples of high power radio and TV broadcast antennas embodying many of the techniques that have been discussed.

29.10 BASE STATION ANTENNAS FOR MOBILE PHONE SERVICES In many respects the design of base station antennas parallels that of a directional broadcast array. In both systems the shaping of the azimuth radiation patterns is used to control coverage and co-channel interference in the coverage area of other stations. Both systems employ columnar arrays of radiating elements in which the currents in different tiers are closely controlled in order to shape the elevation radiation pattern—elevation null-fill is used to avoid holes in local coverage. The main differences are the azimuth pattern (base stations have radiating elements firing in only one direction), transmitted power (a few hundred watts rather than tens of kilowatts), and the scale of production—these differences have led to the widespread use of microstrip techniques in both radiating elements and feed systems.



29-37

The use of circular polarization in many broadcast systems is paralleled by the use of polarization diversity in base station antennas (transmitting independent linear components rather than combining them in phase quadrature). Broadcast antenna engineers will find many interesting techniques in base station antenna design, where the changing demands of customers and intense competition have fostered rapid development of simple but effective broadband methods adapted to large-scale production. The parameters of a base station antenna are an essential means by which the coverage and capacity of the network are optimized. Space-diversity receive systems employing two vertically polarized antennas are common in rural areas; in urban areas polarization diversity is often used and is provided by antennas supporting two orthogonal linear polarizations (±45º). Because of the closely spaced channels used by a base station, FIGURE 29-19 Typical broadcast antenna installathe transmitters are often combined using tions on a concrete supporting structure. A UHF TV 3-dB hybrids (incurring a 3-dB loss). To antenna is mounted on top of the tower. It is protected minimize combining losses, all available by a GRP radome that is provided with helical strakes antenna ports are used for multichannel to reduce its vulnerability to mechanical oscillation at certain wind speeds. Below this is a VHF FM radio transmission as well as for reception, so antenna built from dipole panels, and below this is a there are very stringent requirements on variety of LPDAs. (Courtesy of Kathrein-Werke KG) passive intermodulation products. The development of an increasing number of frequency bands for mobile radio has led to the development of dual-band antennas (usually 850/1900 MHz or 900/1800 MHz); the advent of 3G services in the 2100-MHz band has encouraged the development of wideband antennas operating over the ranges 826–960 MHz and 1710–2170 MHz. Further developments in base station antennas include the use of remotely adjustable beamtilt, azimuth beamwidth, and azimuth pointing direction, and the integration of lownoise amplifiers into antenna systems. The continuing pressure to meet rising capacity and coverage needs has led to the development of inconspicuous “street-works” base stations with three sector antennas, each with remote tilt facilities, housed in a single radome and mounted on a street-lighting pole. A vertically polarized antenna comprises an array of vertically polarized dipoles mounted in front of a reflecting screen and protected by a radome, usually made from UVstabilized ABS or polystyrene. Successful designs have used ●

●

●

●

Punched or pressed dipoles fed by miniature coaxial cables, using cable or microstrip power dividers Microstrip dipoles with a distribution network integrated in a single PCB or fabricated assembly Dipoles fed by suspended microstrip feeds, all cut from aluminum sheet Stacked patches with microstrip feed networks



CHAPTER TWENTY-NINE

Dual-polar antennas have been produced using ●

●

●

Crossed dipoles constructed by printed-circuit or sheet metal techniques, integrated with microstrip or cable feed systems. Squares comprising four half-wave dipoles, usually of stamped or cast metal construction with cable feed. Patch elements, usually air-spaced and standing 10–20 mm above the reflecting plane, fed as stacked patches or by some other indirect feed mechanism below them. (Directly edge-feeding a wideband patch gives rise to pattern squint at frequencies far from the design frequency.)

Dual-polar antennas require the specification of several additional parameters to ensure optimum operation in a mobile radio system: Azimuth radiation patterns and gain are defined assuming the antenna under test is illuminated by a plane-polar signal with a polarization of either +45º or –45º. Because the sense of slant polarization in the rearward direction is opposite to that at 0º, the cross-polar rear lobe often limits network performance and should always be measured. Some networks define a sum-power limit, but this is pessimistic as no real antenna could receive the sum power in both polarizations. The azimuth pattern is defined and measured as a conical cut of the 3D solid pattern, with a half-angle equal to the complement of the nominal elevation beamtilt. Errors in the azimuth pattern relate to the design of the radiating elements and their immediate environment in the array. Elevation beamtilt is conventionally defined in the sense that positive beamtilt is downward. The specified tilt must be achieved within some defined limit that must be maintained over the whole operating bandwidth. Changes in tilt result in uncertainty in the coverage footprint of the base station between transmit and receive frequencies and variable overlap with adjacent stations. The impedance matching of the upper and lower halves of the array is the first matter to investigate if an array has an unstable beamtilt; any significant mismatch will cause a frequency-dependent phase error between them. Squint is the angle between the azimuth directions of the beams formed in the two polarizations, usually measured as the mean direction between the –3-dB points of the azimuth patterns in each polarization. Squint is controlled by the electrical symmetry of the radiating elements and their means of excitation. It is usually greatest for large tilt angles because the use of ±45º polarization puts any element squint into the diagonal planes: the larger the elevation tilt, the more any element squint expresses itself in the azimuth pattern of the array. Tracking is the gain difference between the two polarizations measured over some specified range of azimuth angles surrounding the main beam direction, so this parameter includes the effects of both squint and beam-center gain difference. If there are significant frequency-dependent tracking errors, a mobile may be handed-off into the cell on the basis of a strong signal in the broadcast control channel (BCCH), but if the polarization tracking is poor and the assigned channel is transmitted in the other polarization from the BCCH, the signal may be several dB below what would have been expected. Precise tracking is achieved by minimizing squint and making the +45º and –45º subsystems as mechanically and electrically equal as possible. Cross-polar or interport isolation (XPI or IPI) is the attenuation measured between the +45º port and the –45º port. This is significant because of the practice of using both ports for transmission—the coupled energy between transmitters must lie below –30 dB because this is the level at which the transmitters meet their spurious emission specifications. This is a stringent requirement and various coupling compensation mechanisms are used, including tabs on the elements, rods, bars, and “mushrooms” placed between elements and along the edges of the antenna chassis, metal tapes inside radomes, and feedline coupling. The XPI usually increases with increasing elevation tilt because the phase shift across the array results in the coupled signals being increasingly out of phase. Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.


29-39

Cross-polar discrimination (XPD) is the relative response of one section of the antenna to co- and cross-polar signals. A more logical parameter to define is polarization orthogonality, which is the parameter that permits the antenna to provide polarization diversity, but this is mathematically complex and is not easy to measure. It is a common feature of directional slant-polar antennas that their polarization on boresight is close to 45º, but as the observer moves away from boresight the polarization tends to become dominated by the vertical component (at 90º from boresight it is easy to understand that the polarization becomes completely vertical). In practice, the diversity gain obtained from a dual polar antenna does not fall significantly for values of orthogonality or XPD larger than about 10 dB,25 so this parameter is not often defined or measured. The achievement of a large value of XPI does not ensure that the XPD is high, and the XPD should always be checked, especially if large amounts of compensation are necessary to achieve the required XPI. A 45º-slant polarization has the unusual property that if the transmitting antenna has +45º polarization, the polarization appears as –45º to the antenna that receives it. A base station antenna is used for both transmission and reception, and there is no clear convention as to how ports should be labeled. If they are used or measured in any manner in which the distinction is relevant, the port naming should be regarded as arbitrary and the actual polarization checked. Cross-band isolation is defined for multiband antennas to ensure that the signals presented to the base station filter system lie within the expected limits. In antenna terms, insufficient cross-band isolation is likely to indicate that currents in the high-band part of the antenna will excite radiating currents in the low-band elements, corrupting the highband radiation patterns. If an antenna has adequate radiation patterns but some deficit in cross-band isolation, it is acceptable to add filtering in the feed system. Some antennas are required to provide a single input port for both frequency band groups, in which case a diplexer will be needed to combine the feed systems of the highand low-band arrays. This can be formed from high-pass and low-pass elements engineered in microstrip and needs sufficient isolation to prevent the radiation pattern distortion explained earlier. Passive intermodulation products (PIMs) of all orders are generally required to lie below –153 dBc for two carriers at a level of +43 dBm (20W). For dual band antennas the first-order transmit-band products may lie in the receiver band and can be particularly troublesome (900 + 900 = 1800; 1800 – 900 = 900). Good PIM performance is only achieved by the exercise of good design practices and of care and cleanliness in manufacture (see Section 29.2). Input VSWR is usually required to be less than 1.4:1 across the relevant operating bands. It is usually most difficult to achieve this if the array has a 0º electrical beamtilt because all the elements are driven in phase so the reflections also arrive back in phase at the input. As the beamtilt is increased the reflections are no longer cophased, so the required VSWR is easier to achieve. Some design margin is needed on these parameters between the specification and what is measured on prototypes and pilot production, in order to accommodate production variation without products failing to meet specification. In this respect PIM, XPI, and VSWR are the parameters most likely to vary significantly between samples; variation is controlled by adopting good design practices, ensuring components are accurately made, and educating production personnel about consistency and cleanliness. Adjustable Elevation Tilt As noted earlier, the use of beamtilt has long been a feature of both base station and broadcasting antennas. The use of remotely controlled adjustable tilt for base station antennas was introduced as a method by which the network operator can control the area covered by a base station antenna. Reducing the transmitted power also reduces coverage in buildings and shadow areas close to the base station, while downtilting the main beam allows close control of the position and depth of the overlap area



CHAPTER TWENTY-NINE

Control

Drive

Power divider

FIGURE 29-20 A typical array with adjustable electrical tilt, fed as five subarrays, each of two elements, using tapped-line phase shifters powered from a single drive motor. Feedlines are usually dimensioned so that when the phase shifters are at one end of their travel, the tilt is 0º.

between base stations. Electrical downtilt (produced by introducing a progressive phase shift over the aperture of the antenna) is used in preference to mechanical tilting because mechanical tilt changes the shape of the footprint as well as its size, and physically downtilted antennas are visually more obtrusive than untilted antennas. The requirement for very low PIM levels and substantial transmitted power has led to the development of electromechanical phase shifters driven by stepper motors. Three different forms of design are in general use: transmission lines whose electrical length is changed with moving dielectric loading, lines with capacitively coupled moving taps, and lines with trombone sections. In applying these to a linear array it is necessary to arrange that the phase shift for each element is proportional to its distance along the array. To reduce the total phase shift needed it is common to use a differential scheme in which, to tilt the beam down, elements at the bottom of the array are delayed in phase and simultaneously those at the top are advanced. To reduce the complexity of the phase shifters, the elements of the array are often grouped together into fixed subarrays of two or three elements, and variable phase shift is applied between them (see Figure 29-20). Grouping elements into subarrays to some extent compromises the performance of the array; a truly linear phase shift produces lower elevation sidelobe levels even at extreme tilts (see Figure 29-21). An industry-standard digital control interface has been defined and implemented by all major antenna manufacturers—this interface26 provides control for antennas, tower-mounted amplifiers, and other tower-top hardware. 00

0

-5 −10

-15

Amplitude (dB)

Level (dB)

Amplitude (dB)

-10 −10

-20 −20 -25 -30 −30

−20

−40

-35 -40 −40

−60

-150 −150

-100 −100

-50 0 −50 0 Angle (deg) Angle (deg)

50 50

100 100

150 150

08 −80

06 −60

04 −40

2 0 02−020 )ge0 d( elgnA20

0440

0660

0880

Angle (deg)

FIGURE 29-21 Examples of the azimuth (left) and elevation patterns (right) of a base station antenna showing excellent consistency of beam shape and sidelobe levels over a range of electrical elevation beam tilt angles of 0º–10º (Courtesy of Argus Technologies)



29-41

Extension of remote control techniques to the adjustment of azimuth pattern characteristics—beamwidth and pointing direction—is relatively easy and provides increased adaptive control of network optimization, allowing real-time response to changing demands for capacity and coverage in the network. Such techniques are based on cross-layer communication originating from radio resource management; they are symmetrical in their enhancement of uplink and downlink and are protocol independent. Antenna systems embodying them have been described as semi-smart antennas.27

29.11 SYSTEM CONSIDERATIONS Mounting Arrangements When mounting any antenna it is important not to impair its performance by the influence of the supporting structure. The inevitable effect of the supporting structure on the radiation pattern of a dipole has been referred to in Section 29.8. This effect is accompanied by a modification of the input impedance, which may be unwelcome if a low VSWR is needed. In any critical application the change of the radiation patterns and gain must be taken into account when estimating system performance. Impedance matching of the antenna must be undertaken in the final mounting position or a close simulation of it. If Yagi-Uda arrays are mounted with their elements close to a conducting structure, they too will suffer changes of radiation patterns and impedance. The effects will be greatest if tower members pass through the antenna, as they do when an array is mounted on clamps fitted at the center of the cross boom. If at all possible, when an array is center-mounted, the member to which it is clamped should lie at right angles to the elements of the array. Currents induced in diagonal members of the supporting structure will cause reradiation in polarization planes other than that intended. This will result in the cross-polar discrimination of the antenna system being reduced from that which would be measured on an isolated antenna at a test range. When polarization protection is important, the tower should be screened from the field radiated by the antenna with a cage of bars spaced not more than l /10 apart, lying in the plane of polarization. (A square mesh is used for circular polarization.) Panel antennas are designed with an integral screen to reduce coupling to the mounting structure. Long end-mounted antennas are subjected to large bending forces and turning moments at their support points. These forces can be reduced by staying the antenna, using nylon or polyester ropes for the purpose to avoid degrading its electrical characteristics. In severe environments antennas may be provided with radomes or protective paints. It is very important that the antennas are tested and set up with these measures already applied, especially if the operating frequency is in the UHF band. Coupling A further consideration when planning a new antenna installation on an existing structure is the coupling that will exist between different antennas. When a transmitting antenna is mounted close to a receiving antenna, problems that can arise include ●

●

●

Radiation of spurious signals (including broadband noise) from the transmitter Blocking or desensitization of the receiver Generation of cross-modulation effects by the receiver

The last two effects depend critically on the isolation between the antennas and on parameters of the transmitters and receivers; these parameters are generally specified by their manufacturers.



CHAPTER TWENTY-NINE

FIGURE 29-22 Typical isolations between Yagi-Uda antennas

The isolation between two antennas may be predicted from Figure 29-22, measurements, or from standard propagation formulas. Antenna isolations may be increased by using larger spacings between them, arranging that they are cross-polarized, or by using arrays of two or more antennas spaced to provide each with a radiation-pattern null in the direction of the other. An alternative method of increasing the isolation between the antenna-system inputs is to insert filters. If a suitable filter can be constructed, the antenna isolation may be reduced until, in the limit, a single antenna is used with all equipment, transmitters, and receivers coupled to it through filters. When receivers are connected to a common antenna, the signal from the antenna is usually amplified before being divided by a hybrid network. The number of services that use a single antenna can be extended to six or more, provided adequate spacings are maintained between the frequencies allocated to different users. The whole system is expensive, but the cost may be justified if the antenna itself is large or if tower space is limited. Multiple User Systems To reduce the environmental impact and cost of antenna systems, users share antenna resources by installing a single broadband antenna system and using it to support multiple radio transmitters or receivers (or both). Not only does this practice reduce the profile of an antenna installation, but it also allows several services to make best use of locations where antennas can be most favorably sited.

REFERENCES 1. E. Simiu and R. H. Scanlan, Wind Effects on Structures: An Introduction to Wind Engineering (New York: John Wiley, 1986). 2. H. H. Uhlig, Corrosion and Corrosion Control (New York: John Wiley, 1985).



29-43

3. M. G. Fontana, Corrosion Engineering, 3rd Ed. (New York: McGraw-Hill, 1986). 4. Lightning Protection Code, ANSI C5.1-1969 (New York: American National Standards Institute, 1969). 5. Code of Practice for the Protection of Structures Against Lightning, BS6651:1990 (London: British Standards Institution, 1990). 6. R. H. Golde (ed.), Lightning, Vol. 2: Lightning Protection (New York: Academic Press, 1977). 7. “Electromagnetic Compatibility and Radio Spectrum Matters (ERM); Radio Site Engineering for Radio Equipment and Systems,” ETSI Guide, ETSI EG 200 053, (Sophia Antipolis, France: European Telecommunications Standards Institute, 2004). This is an excellent general guide to good site practice, available from www.etsi.org. 8. On the restriction of the use of certain hazardous substances in electrical and electronic equipment, Directive 2002/95/EC of the European Parliament (and documents cited therein). 9. Guidelines for Standardization of Material Declaration, Japan Green Procurement Survey Standardization Initiative, 2nd Ed., June 2004. 10. Material Composition Declaration of Electronic Products: Joint Industry Guide (JIG) (Arlington, VA: Electronic Industries Association (EIA), April 2005). 11. “Guidelines for Limiting Exposure in Time-Varying Electric, Magnetic, and Electromagnetic Fields (up to 300 GHz) ”, International Commission on Non-Ionising Radiation Protection, (ICNIRP), Health Physics, 74 (1998): 494–522. 12. B. S. Collins, “RADHAZ: The Unmentionable Hazard?” Electronics World (October 2004). 13. Establishing a Dialogue on Risks from Electromagnetic Fields (Geneva: World Health Organization, 2002). 14. J. E. Moulder, K. R. Foster, L. S. Erdreich, and J. P. McNamee, “Mobile Phones, Mobile Phone Base Stations, and Cancer: A Review,” Int J Rad Biol, vol. 81 (2005): 189–203. 15. The most current information will always be found on the web sites of the major national public health institutions, for example, http://www.fcc.gov/oet /rfsafety/. 16. S. Shibuya, Basic Atlas of Radio Wave Propagation (New York: John Wiley, 1987). 17. J. D. Kraus and R. J. Marhefka, Antennas for All Applications (New York: McGraw-Hill, 2001). 18. C. Kilgus, “Shaped-Conical Radiation Pattern Performance of the Backfire Quadrifilar Helix,” IEEE Trans. AP, vol. 23, no. 3 (May 1975): 392–397. 19. O. Leisten, J. C. Vardaxoglou, P. McEvoy, R. Seager, and A. Wingfield, “Miniaturised Dielectrically-Loaded Quadrifilar Antenna for Global Positioning System (GPS),” Electronics Letters, vol. 37, no. 22 (October 25, 2001): 1321–1322. 20. N. E. Lindenblad, (October 1940): U.S. Patent 2 217 911. 21. S. B. Cohn, “Optimum Design of Stepped Transmission-Line Transformers,” IEEE Trans MTT, vol. 3, issue 3 (April 1955):16–20. 22. C. W. Davidson, Transmission Lines for Communications (New York: Palgrave Macmillan, 1989). 23. K. Chang, Handbook of Microwave and Optical Components: Vol. 1, Microwave Passive and Antenna Components (New York: John Wiley, 1997). 24. G. L. Matthei, L. Young, and E. Jones, Microwave Filters, Impedance Matching Networks and Coupling Structures (Norwood, MA: Artech House, 1980). 25. B. S. Collins, “The Effect of Imperfect Antenna Cross-Polar Performance on the Diversity Gain of a Dual-Polarized Receiving System,” Microwave Journal, vol. 43, no. 4 (April 2000): 84–94. 26. Control Interface for Antenna Line Devices, Standard AISG v2.0, Antenna Interface Standards Group (AISG Ltd.), 2006. Available from www.aisg.org.uk. 27. C. G. Parini et al, Semi-Smart Antenna Technology, Ofcom Document No. 830000081/03 (London: Ofcom, 2006).


VHF and UHF Antennas for Communications and Broadcasting



Chapter 30

Portable TV Antennas Mitsuo Taguchi Nagasaki University

CONTENTS 30.1 INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-2 30.2 DIGITAL TERRESTRIAL TELEVISION IN THE WORLD . . . . . . . . . . 30-2 30.3 MINIATURIZATION TECHNIQUES FOR PORTABLE TV ANTENNAS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-3 30.4 RECEPTION ANTENNA MOUNTED ON VEHICLE . . . . . . . . . . . . . . 30-6


Portable TV Antennas 30-2

CHAPTER THIRTY

30.1 INTRODUCTION Digital terrestrial television broadcasting is being introduced in many countries. This is providing a variety of services on a single channel, such as high-definition television (HDTV) and multimedia data. Obtaining portable TV reception generally is difficult compared to obtaining fixed reception at home. For fixed reception, a directional antenna with high gain and a narrow beam width, such as the Yagi-Uda dipole array antenna, is used.1 For portable reception, an omnidirectional antenna or low-gain antenna is used. The objective of this chapter is to describe portable reception antennas suitable for digital terrestrial TV, such as small antennas with amplifier circuits, mobile phone antennas, and reception antennas mounted on cars.

30.2 DIGITAL TERRESTRIAL TELEVISION IN THE WORLD Table 30-1 shows the frequency allocation plan in the world.2 In many countries (but not in the United States or Canada), the high-band VHF from 174 MHz to 240 MHz and the UHF band from 470 MHz to 862 MHz are used for the digital terrestrial television. TABLE 30-1

Frequency Allocation Plan and Channel Plan for Digital Terrestrial Television in the World

54 72 Ch. 2-4 54 72 Canada (March 2003) Ch. 2-4

USA (Nov. 1988)

76 88 5-6 76 88 5-6

216

174

470

Ch. 7-13 216

174 Ch. 7-13

England (Sept. 1998) Ireland (2006)

174

France (March 2005)

174

209 Ch. 5-9 230 Ch. 5-12

Spain (Nov. 2005) Portugal (unknown) Germany (Nov. 2002)

202 174 Ch. 5-6,8 174

Finland (Aug. 2001)

174

Denmark (March 2006)

174

Netherland (April 2003) Sweden (April 1999) Norway (2006-2007) Italy (April 2004)

239 230

470

230

470

862 MHz 830 MHz 862 MHz 862 MHz 862 MHz 790 MHz 862 MHz

Ch. 5-12 216 174 470 Ch. 5-10 230 174 470 Ch. 5-12 240 470 174

Ch. 21-69 862 MHz Ch. 21-69 862 MHz Ch. 21-69 862 MHz Ch. 21-69

230

174

854 MHz

470 710 MHz

470 Ch. 13-52 174

216

752 MHz

470

Ch. 7-13

Ch. 14-60 602 MHz 530 Ch. 24-35

Taiwan (July 2004)

806 MHz Ch. 21-62 790 MHz 470 614 Ch. 24-38 Ch. 39-60 520 610 750 806 MHz -Ch. 67 Ch. 28-

470

HongKong (-2007) Singapore (Feb. 2001) Australia (Jan. 2001)

806 MHz

Ch. 21-60

Ch. 5-12

Japan (Dec. 2003) Korea (Oct. 2001)

698 MHz 608 614 Ch. 38-51 470 698 MHz 608 614 Ch. 14-36 Ch. 38-51 470 550 630 Ch. 41-62 Ch. 21-30 470 Ch. 21-69 470 Ch. 21-65 470 Ch. 21-69 582 470 Ch. 21-69 470 Ch. 14-36

174

230 Ch. 6-12


Portable TV Antennas PORTABLE TV ANTENNAS

30-3

30.3 MINIATURIZATION TECHNIQUES FOR PORTABLE TV ANTENNAS This section shows examples of small portable TV reception antennas such as the small antenna with amplifier circuit and the mobile phone antenna. Small Antenna with Amplifier Circuit for UHF-TV Figure 30-1 shows the structure of a coplanar waveguide (CPW)-fed dipole antenna and a loop antenna for a television receiver and DC-equivalent biasing circuits.3,4 The antenna element and CPW are printed on a polyimide film of thickness 45 mm. The relative permittivity of film is 3.5. The characteristic impedances of CPW in two antennas are chosen as 75 Ω for impedance matching to the coaxial feeder of the television receiver. The antenna element and CPW are covered by a film of thickness 50 mm for protection. The silicon transistor 2SC2585 or 2SC3604 is integrated at the feedpoint of the dipole or loop antenna, respectively. One of the primary problems with the active antenna is that the unwanted radiation from the amplifier circuit and the bias line will perturb the antenna characteristics. To overcome this problem, the amplifier circuit is mounted on CPW. CPW also serves as a bias supply line to the amplifier circuit. The available power gain of 2SC2585 is 8 to 14 dB at frequencies from 90 to 770 MHz. The nominal value of the noise figure of this transistor is about 1.6 dB at 2 GHz. To suppress the undesired radiation from the amplifier circuit, the fixed-bias circuit with collector feedback, the simple biasing circuit, is adopted for the active dipole antenna. On the other hand, the available power gain of 2SC3604 is 24 to 18 dB at frequencies from 90 to 770 MHz. As the available power gain of 2SC3604 is larger than one of 2SC2585, the self-bias circuit is adopted to suppress oscillation and to increase stability. The nominal value of the noise figure of this transistor is about 1.6 dB at 4 GHz. The input impedance and radiation field of dipole and loop antennas are calculated by using the computer program WIPL-D based on the Method of Moment.5 In the calculation, antenna elements are approximated by the wire antenna whose radii are Wi / 4 (i = 1,2). As the dielectric film is very thin, the existence of film is not considered.

FIGURE 30-1 Dipole and loop antennas loaded with amplifier circuits



CHAPTER THIRTY

(a) Dipole antenna, L =12 cm

(b) Loop antenna, W2 = 12 mm

FIGURE 30-2 Input impedance of dipole and loop antennas

Figure 30-2 shows the input impedance characteristics of dipole and loop antennas. The calculated input impedances fairly agree with the measured results. Figure 30-3 shows the scattering parameters of transistors for active antennas. In the television 1–12 channel frequencies 90 to 222 MHz, the dipole antenna is not matched to the transistor circuit. Although the input reactances of dipole antennas are capacitive in the frequencies 90 to 222 MHz, those of loop antennas are inductive. This means that the conjugate impedance matching between antenna and amplifier is improved in the loop antenna. Figure 30-4 shows the calculated and measured actual gains expressed by relative values to the half-wave dipole antenna. As the impedance matching between a dipole of 24cm in length and transistor 2SC2585 is reasonable at the television 13–62 channel frequencies from 470 to 770 MHz in Japan, actual gains of more than 8 dBd are obtained. In the active

(a) 2SC2585

(b) 2SC3604

FIGURE 30-3 S parameters of silicon transistors



(a) Dipole antenna, L = 12 cm

30-5

(b) Loop antenna, W2 = 12 mm

FIGURE 30-4 Actual gains of dipole and loop antennas

loop antenna, the impedance matching between loop and transistor 2SC3604 is improved in the frequencies 90 to 222 MHz. Actual gains of more than 6.5 dBd are obtained for 13–62 channels, from 470 to 770 MHz. Portable Antenna on Mobile Phone Digital terrestrial television broadcasting in Japan is designed so that each channel is divided into 13 segments. Among these 13 segments, 12 segments are used for HDTV transmission and 1 segment is used for mobile receivers.6 Maximum video resolution is 320 by 240 pixels, and maximum video bit rate is 128 kbit/s. Maximum audio bit rate is 64 kbit/s and the additional data broadcasting occupies the remaining 60 kbit/s. Digital Multimedia Broadcasting (DMB) is a digital radio transmission system for sending multimedia (radio, TV, and datacasting) to mobile devices such as mobile phones. Currently, DMB is being put into use in a number of countries such as South Korea, Germany, France, Switzerland, China, the UK, and Indonesia.7 Figure 30-5 shows an example of the embedded antenna for digital terrestrial broadcasting in Japan (one segment).8 Its size is 38 mm by 10 mm by 2 mm. Figure 30-6 shows the antenna module, consisting of the antenna, the tuning circuit, and the low noise amplifier (LNA). Figure 30-7 shows the antenna tuning. The antenna is directly connected with the LNA, and is tuned to be synchronized with the channel selection. The antenna gain is self-controlled with the sensitivity of the tuner module.

FIGURE 30-5 Antenna module for Japanese digital mobile TV



CHAPTER THIRTY

FIGURE 30-6 Block diagram of antenna module FIGURE 30-7 Antenna tuning

30.4 RECEPTION ANTENNA MOUNTED ON VEHICLE This section shows examples of reception antennas mounted on vehicles, such as a window glass antenna, a diversity reception antenna system, and an adaptive beam-steering reception system. Reception Antenna on Rear Window of Sedan-Type Car Figure 30-8 shows an example of window-glass antenna printed on the rear window of a car, composed of four diversity reception antenna elements.9 TV antennas No. 1 and No. 2 are located close to the defogger (heater) element. TV antenna No. 4 is directly connected to a defogger element. Figure 30-9 shows the measured electric field radiation patterns in the horizontal plane. Figure 30-10 shows the average gain around all directions in the horizontal plane. The directivity is measured with a 10 dB amplifier circuit. In these figures, the angle is measured from the front direction. Diversity Reception Antenna System Figure 30-11 illustrates the diversity reception antenna system for TV.10–12 The antenna elements are printed symmetrically on the left and right quarter windows of the car. The antenna consists of two parts: the slanted element consisting of three thin lines located

FIGURE 30-8 Window glass antenna printed on the rear window of sedantype car



30-7

FIGURE 30-9 Measured electric field radiation patterns in the horizontal plane

close to the rear pillars, and a horizontal element printed at the center of the quarter glass. A switching diversity system was adopted because the synthetic diversity system was not sufficiently effective at suppressing multipath distortion. The switching diversity system can effectively follow changes in field strength by selecting appropriate antennas for better reception. The antenna switching is performed during every vertical blanking period in the TV video signals. The diversity reception system for the terrestrial TV has been on the market since the middle 1980s.11,12 Two V-shaped pole elements were first used for TV reception. A new antenna system consisting of four antenna elements printed on the rear quarter window glass was developed in 1986. Subsequently, an adaptive array antenna for terrestrial digital TV broadcasting was developed, and you will soon be able to enjoy TV programs in highquality digital while riding in an automobile.13

FIGURE 30-10 Average gain around all directions in the horizontal plane



CHAPTER THIRTY

FIGURE 30-11 Diversity antenna system for TV

On-glass Mobile Antennas for Digital Terrestrial Television Figure 30-12 shows the geometry of the horizontally polarized on-glass mobile antenna.14 The antenna has been developed by modifying a batwing antenna.15 Batwing antennas have a broad bandwidth but are too large to be installed in vehicles, so development has focused on size reduction. The proposed antenna is normal to ground and fed by a coaxial cable. The antenna height, H, was reduced by bending the wires in that direction without narrowing the bandwidth. The height of the new antenna is approximately half of the original antenna height. The prototype antenna for field experimentation using a passenger van is shown in Figure 30-12b. Two antenna elements and a mesh ground plane were printed on the inside of the rear glass window of the passenger van. The dielectric constant of the glass was 6.7, and the loss tangent was 0.008. The thickness was about 5 mm, which is very thin compared with the wavelength of the UHF band. Experiments confirmed that the glass had a very small effect on antenna characteristics. The patterns were printed using silver paste, and the mesh ground plane was connected electromagnetically to the vehicle body upon installation of the window. The conductivity of the silver paste was about 2 × 106 S/m, which is lower than that of copper (5.813 × 107 S/m). Antenna gain loss was confirmed to be less than 0.3 dB within the frequency band by comparison with the measured results of an antenna composed of copper.

(a) Antenna element

(b) Prototype on-glass antenna

FIGURE 30-12 Geometry of on-glass antenna



30-9

FIGURE 30-13 Measured radiation pattern (vertical plane, 485 MHz)

Figure 30-13 shows the radiation patterns of the antenna installed on the left side of the passenger van at 485 MHz. As expected, the elevation pattern peak of Ef , the co-polarization of DTV, was in the horizontal plane. Figure 30-14 shows the radiation patterns in the horizontal plane at 485 MHz. The antenna element printed inside the left side glass radiates strongly to the left side of the vehicle, and the element inside the right side glass radiates strongly to the right side. A peak gain of 4.5 dBi was achieved by each antenna element, further showing the suitability of this type of antenna for use in a diversity system. A diversity system can achieve a high C/N ratio in built-up urban areas where multipath fading is strong. Maximum Ratio Combining (MRC) was selected as the signal processing method.16 The radiation pattern of MRC was estimated by combing the measured radiation patterns of four antennas installed on the passenger van, as shown in Figure 30-15. An average gain of 4.6 dBi was achieved on the horizontal polarization. Figure 30-16 shows the average gain of the pattern. The solid line represents the gain of the proposed antenna and the dotted line represents a comparison using the average gain of four commercial pole antennas installed on the roof. These pole antennas were leaned at an angle of 45 degrees to the roof. The average gain of four crossed-dipole antennas installed on the corner of the roof was also measured at 485 MHz.17 From 470 MHz to 710 MHz, the average gain of the proposed antenna was greater than 3.1 dBi. This was 4.7 dB higher than that of the commercial pole antennas, which is the minimum value within the bandwidth. It was also 1.6 dB higher than the average gain of crossed-dipole antennas at 485 MHz.

FIGURE 30-14 Measured radiation pattern (horizontal plane, 485 MHz, element)



CHAPTER THIRTY

Modified H-shaped Antenna Figure 30-17 shows a modified H-shaped antenna for automotive digital terrestrial reception.18 Two wires a-b-c-d-e and f-g-h-i-j are symmetrically placed around the center m of the antenna and connected by the wire c-m-h. The parts a-b-c and f-g-h of the two wires are longer than the parts c-d-e and h-i-j. The antenna is placed in the xy plane. A control system for combining received signals is shown in Figure 30-17b. The received signals of the four antennas are downconverted, weighted, and combined. The weight vector for each signal is controlled, based on the MRC method. The four antennas mounted at the top of the front and rear windows are also depicted. The antenna FIGURE 30-15 Radiation pattern plane, which includes the antenna element, is inclined (four-element array) from the horizontal plane in an actual car. A design concept of the coverage in the x y plane is summarized in Figure 30-18. Mechanism of the antenna is described with Figure 30-17a and Figure 30-18. The antenna has three resonant modes. Each resonant mode is excited when a part of the antenna becomes a half-wavelength. The frequency band for digital terrestrial services is divided into three bands. The wire a-b-c-m-h-g-f is resonated at the low frequency, which is series resonance. Parallel resonance occurs at the middle frequency. The two wires a-b-c-d-e and f-g-h-i-j are resonated. Series resonance occurs again at the high frequency, in which the wire e-d-c-m-h-i-j is resonated. The radiation pattern of a single element is rotated clockwise with increasing frequency. The minimum level of relative amplitude increases when the maximum radiation is directed toward the x-axis direction. The four antennas are symmetrically placed around the x-axis and y-axis directions in Table 30-2, which is assumed to be equivalent to installation at the top of the front and rear windows. The symmetrical arrangement allows coverage of 360 degrees. The wires shown in Figure 30-17a were etched on a typical FR4 substrate having dielectric constant of 4.6 and thickness of 0.8 mm. The width of the lines was 1 mm. The lengths LL, LS , and LC were experimentally adjusted to 102, 67, and 30 mm, respectively, with constant value of H = 60 mm, since the value of H was specified for car installation. The prototype antenna has LC balun and a parallel feedline between the LC balun and the feedpoint m.

FIGURE 30-16 Average gain of the adaptive array pattern (Ef)



PORTABLE TV ANTENNAS

(a) Antenna element

(b) Control system

FIGURE 30-17 Modified H-shaped antenna and control system for combining received signals for automotive digital terrestrial reception

The LC balun transforms impedance from 110 Ω to 50 Ω as well as balance-unbalance mode. The LC balun consists of a high-pass filter (HPF), a low-pass filter (LPF), and a T-junction. Both filters have the fifth-order Butterworth function. A 3 dB cutoff frequency is set at a low frequency of 272 MHz for the HPF and a higher frequency of 1228 MHz for the LPF of the frequency band from 470 to 710 MHz so that both filters may give flat amplitude and out of phase. Insertion loss of the LC balun was measured to be 0.3 dB across the frequency band. The parallel feedline has a length of 40 mm. The width of each line and the distance between the centers of the two lines were chosen to be 0.8 and 1.1 mm, so that characteristic impedance would be 110 Ω. Figure 30-19 shows the measured radiation patterns at the three frequencies. The figure-8 radiation patterns were rotated clockwise with increasing frequency as the calculation results predicted. VSWR was measured to be less than 3 from 470 to 710 MHz. Four prototype antennas were mounted at the top of the front and rear windows of the car, as shown in Figure 30-17b. The distance between the edges of the metal roof and the nearest side of the prototype antennas was set at 20 mm. The spacings of antenna elements were, respectively, 600 mm in the x’-axis direction and 1600 mm in the y’-axis direction. Figure 30-20a–d show typical measured radiation patterns of the four prototype antennas in the f’ plane at 530 MHz. Each prototype antenna had gain to the x’-axis direction as well as to the y’-axis direction because the four antennas in free space had inclined figure-8 radiation patterns. The peak plot combined with the radiation patterns in Figure 30-20a–d is shown in Figure 30-20e. A near omnidirectional pattern was achieved in the f’ plane. In terms of vertical plane, gain of the peak plot decreased with increasing the angle q ’, 484 MHz 576 MHz

638 MHz

FIGURE 30-18 Concept covering 360 degrees in the xy plane using the four antennas across the frequency band for DTV

FIGURE 30-19 Measured radiation patterns of a prototype antenna in the xy plane



CHAPTER THIRTY

FIGURE 30-20 Measured radiation patterns in f’ plane at 530 MHz. Four prototype antennas were mounted at the top of the front and rear windows of a car. Peak plot of the combined pattern based on the MRC method is also presented.

because radiation patterns of the mounted prototype antennas were affected by the effect of the car body. The values of the minimum gain of the peak plot were –2.3 dBi for q’ = 70 degrees, –3.7 dBi for q ’ = 80 degrees, and –6 dBi for q’ = 90 degrees. Adaptive Beam Steering Reception System To improve the reception quality of television on mobile terminals, the adaptive beam steering system has been studied in Japan.19 Figure 30-21 shows the adaptive beam steering reception system. This system adopts the diversity technique, combining received signals prior to FFT processing on the OFDM demodulator. The experiment was conducted in the Nagoya suburbs using a van with four antennas on the windows, and was focused on traveling on highways. From the experiment, it was shown that the mobile reception area using the adaptive beam steering system is almost the same as the fixed reception area.19

FIGURE 30-21 Adaptive beam steering reception system



30-13

REFERENCES 1. E. B. Joy, Antenna Engineering Handbook, Chap. 29, R. C. Johnson (ed.) (New York: McGraw-Hill, 1993). 2. Federal Communications Commission, “FCC Online Table of Frequency Allocations,” http:// www.fcc.gov/oet/spectrum/table/fcctable.pdf (revised November 29, 2006). 3. M. Taguchi, T. Fujimoto, and K. Tanaka, “CPW Fed Active Dipole Antennas for Television Receivers,” Electronics Letters, vol. 30, no. 22 (October 1994): 1815–1816. 4. M. Taguchi and T. Fujimoto, “Actual Gain of CPW-Fed Active Integrated Antennas for Television Receiver,” IEICE Trans. Commun., vol. E81-B, no. 7 (July 1998): 1542–1547. 5. WIPL-D, http://www.wipl-d.com/. 6. Association of Radio Industries and Businesses, Transmission System for Digital Terrestrial Television Broadcasting, ARIB STD-B31, ver. 1.6 (November 2005; in Japanese). 7. Wikipedia, s.v. “Digital Multimedia Broadcasting,” http://en.wikipedia.org/wiki/Digital_Multimedia_ Broadcasting. 8. K. Takei and T. Sugiyama, “Antenna Module Self-Controlling Gain with Sensitivity Feedback,” Proc. of Colloquium on RF for DVB-H/DMB Mobile Broadcast: Handset and Infrastructure Challenges, London, June 2006. 9. Technical data described in this section is provided by courtesy of Central Glass Co., Ltd. 10. K. Nishikawa, “Land Vehicle Antennas,” IEICE Trans. Commun., vol. E86-B, no. 3 (March 2003): 993–1004. 11. H. Toriyama, J. Ohe, H. Kondo, and H. Yotsuya, “Development of Printed-on Glass TV Antenna System for Car,” Proc. IEEE Vehicular Tech. Conf. (1987): 334–342. 12. J. Ohe, “Development of Printed-on Quarter Glass Antenna System for Cars,” ITEJ Technical Report, vol. 13, no. 26, RE’89-15 (May 1989): 19–24. 13. J. Imai, M. Fujimoto, N. Itoh, T. Shibata, N. Suzuki, and K. Mizutani, “Mobile Reception Experiments of Digital Terrestrial Broadcasting—A Study on Beam Control for Improvement of Reception Quality,” Proc. Commun. Conf. IEICE 2001, B-5-113, September 2001. 14. S. Matsuzawa, H. Sato, and K. Nishikawa, “Study of On-glass Mobile Antennas for Digital Terrestrial Television,” IEICE Trans. Commun., vol. E88-B, no. 7 (July 2005): 3094–3096. 15. R.C. Johnson, (ed.), Antenna Engineering Handbook (New York: McGraw-Hill, 1984). 16. M. Fujimoto, K. Nishikawa, T. Shibata, N. Kikuma, and N. Inagasaki, “A Nobel Adaptive Array Using Frequency Characteristics of Multi-carrier Signals,” IEICE Trans. Commun., vol. E83-B, no. 2 (Feb. 2000): 371–379. 17. J. Imai, M. Fujimoto, T. Shibata, N. Suzuki, N. Itho, and K. Mizutani, “Experimental Results of Diversity Reception for Terrestrial Digital Broadcasting,” IEICE Trans. Commun., vol. E85-B, no. 11 (November 2002): 2527–2530. 18. H. Iizuka et al, “Modified H-Shaped Antenna for Automotive Digital Terrestrial Reception,” IEEE Trans. Antennas and Propagat., vol. 53, no. 8 (August 2005): 2542–2548. 19. K. Sanda, T. Shibata, N. Itoh, K. Ito, N. Suzuki, and J. Imai, “Adaptive Beam Steering Reception System for ISDB-T Based on Pre-FFT Diversity Technique,” Proc. IEEE Int. Conf. on Consumer Electronics (ICCE), 11.2-2, Las Vegas, January 2006.


Portable TV Antennas



Chapter 31

Reconfigurable Antennas Jennifer T. Bernhard University of Illinois at Urbana-Champaign

CONTENTS 31.1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31-2

31.2 RECONFIGURABLE APERTURES. . . . . . . . . . . . . . . . . . . . . . . . . .

31-2

31.3 RECONFIGURABLE TRAVELING WAVE ANTENNAS. . . . . . . . . . .

31-5

31.4 RECONFIGURABLE ARRAYS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31-8

31.5 RECONFIGURABLE MICROSTRIP ANTENNAS . . . . . . . . . . . . . . . 31-10 31.6 RECONFIGURABLE SLOT ANTENNAS . . . . . . . . . . . . . . . . . . . . . . 31-13 31.7 RECONFIGURABLE MONOPOLE/DIPOLE ANTENNAS . . . . . . . . 31-16 31.8 CONCLUDING REMARKS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31-19


Reconfigurable Antennas 31-2

CHAPTER THIRTY-ONE

31.1 INTRODUCTION Wireless systems are evolving toward multifunctionality. This multifunctionality provides users with options of connecting to different kinds of wireless services for different purposes at different times, whether in a city or on a battlefield. To reach their full potential, these systems require not only agile electronics and adaptive signal processing, but also reconfigurable antennas. Reconfigurable antennas can be used simply to reduce the number of antennas necessary for intended system function, but they can also be designed to serve much more complex roles. Some of these roles include use as programmable control elements with feedback to increase throughput, reduce errors and noise, and improve security, and use as reconfigurable hardware to extend the lifetime of the entire system. Examples of emerging applications include software defined radio, cognitive radio, multiple-input multiple-output (MIMO) systems, multifunction consumer wireless devices, and high performance phased arrays. As one examines the development of reconfigurable antennas, and, in fact, antennas in general, one can see that each antenna is driven by specific applications or functionality. Indeed, there is no one ultimate application that would result in a super antenna that would be able to meet the specifications of every system. In this, antenna engineers are truly fortunate—practical issues, such as cost, weight, size, and complexity, make each new system or application a new challenge. Antenna reconfigurability can help to meet these challenges in new and exciting ways. This chapter describes the work of talented engineers and researchers in their efforts to develop reconfigurable antennas, from large apertures for communication and sensing satellites to single individual elements on small portable wireless devices. These examples, as well as others, are the first steps in establishing a strong foundation of antennas for fully multifunctional systems.

A Brief Definition of Antenna Reconfigurability A reconfigurable antenna is one that has selectable or tunable fundamental characteristics, including operating frequency, impedance bandwidth, radiation pattern, and polarization. Typically, the goal is to change one or more of these characteristics independently of the others. The methods in which these characteristics are changed are electrical, mechanical, or electromechanical in nature, since fundamentally, antennas are electromechanical structures. However, given the wide range of intended applications and design specifications, engineers have developed an equally wide and varied collection of reconfiguration mechanisms. In each of the following sections, theoretical foundations and examples of different kinds of reconfigurable antennas are presented. While there are many more reconfigurable antennas than can be discussed here, the selected subset of antennas is intended to provide a broad sampling of the rich history and future potential of these tools for communication, sensing, and security.

31.2 RECONFIGURABLE APERTURES In concept, reconfigurable apertures represent the ideal functionality of reconfigurable antennas in that they allow the complete specification of radiating currents that can produce any desired frequency and radiation behavior. The challenge lies in developing real structures and control mechanisms that can support this functionality.


Reconfigurable Antennas RECONFIGURABLE ANTENNAS

31-3

Motivation for Reconfigurable Apertures As mentioned previously, each antenna is typically designed to deliver functionality for a specific application. In the case of aperture antennas, they are most often implemented for long distance, high gain applications intended for more expensive and specialized systems, such as satellites, long-haul land-based communications, and radar receivers. However, when the antenna is large in extent and is deployed in a hostile or remote location, reconfigurability plays a different role. In this case, it could lengthen the useful lifetime of the system by allowing for reprogramming/retasking, without the need for retooling or for a labor- and cost-intensive maintenance visit. The apertures discussed here implement a variety of approaches to enable this kind of reconfigurability. Examples of Reconfigurable Apertures Reflector antennas are perfect candidates for radiation pattern reconfiguration independent of frequency, since the reflective surface is physically removed and isolated from the primary feed. Mechanical deformations of the reflector (or subreflector) surface result in changes in the surface currents on the reflector, which in turn alter the aperture fields, and finally, the radiation pattern of the antenna. In one of the earlier examples of antenna reconfigurability using mechanical means, Clarricoats and his collaborators demonstrated radiation reconfiguration with a reflector antenna by changing the structure of a mesh reflector. Initially, manual adjustments in each meshed region of interest served to change the overall reflector contour, resulting in changes in beam shape and direction.1 Later, the implementation of computer-controlled stepper motors, used for pulling lines attached to specific points on the reflector mesh, allowed for automatic pattern reconfiguration via electromechanical actuation.2 Recently, another group developed and expanded the functionality of a similar system for satellite applications by changing the shape of a subreflector, with the shape of the main reflector remaining fixed.3 A thin, flexible, conductive material composes the subreflector. The subreflector’s shape is changed using piezoelectric actuators attached to its back surface. As the actuators deform the surface of the subreflector, the electromagnetic field illuminating the main reflector changes, which results in a desired radiation pattern. Figure 31-1 shows a subreflector system schematic.3 Enabling changes in the subreflector, rather than the main reflector, simplifies the antenna system fabrication and allows designers to implement main reflectors of varying geometries and deployments. This is especially important for reducing probabilities of failure during deployment or operation for space applications. One important goal of the antenna in this case3 is to achieve the desired reconfigurability with the minimum number of actuators. During the antenna system design, the positions of the actuators are determined using an iterative finite element algorithm. This algorithm determines where successive actuators should be placed in order to minimize the error between the desired and actual subreflector shape.3 In providing the ability to change the antenna’s FIGURE 31-1 Reconfigurable subreflector radiation pattern, this design approach can be system, including piezoelectric actuators and driver to deflect the surface of a Cassegrain used to effectively lengthen the useful life- subreflector (after G. Washington et al3 time of communication and relay satellites. © IEEE 2002)



CHAPTER THIRTY-ONE

Moreover, it can dramatically reduce cost by allowing antennas to be retasked via programming from the ground rather than requiring completely new satellites to enable new functions and all of the cost that this entails. The next two examples of reconfigurable antennas may stretch the definition of an aperture in the traditional sense, but one can certainly consider them to be blank palettes for establishing desired current distributions that result in desired antenna behavior. In essence, these apertures are intended to create whatever antenna is necessary to achieve specific frequency, bandwidth, radiation, and polarization characteristics. The first is an antenna that uses switches (which can be based in solid-state, MEMS, or other technologies) for connecting small non-resonant conductive pads that create a large antenna aperture.4 The number of switches required for such an aperture can easily number into the thousands. Figure 31-2 shows a conceptual drawing of the aperture.4 The desired antenna configuration is found using a genetic search algorithm, which determines the needed combination of switch states across the aperture for achieving a specified performance goal.4 For each pass through the genetic algorithm, a finite-difference time-domain electromagnetic simulation is used to evaluate candidate designs. In general, narrowband configurations achieve more gain than wideband configurations, since wideband configurations implement fewer effective radiators in the aperture, which has a fixed physical size. While this aperture approach promises tremendous flexibility in terms of both frequency and radiation performance, it does have some inherent drawbacks. High frequency operation may be degraded as a result of the pad density and switch capacitance.4 Additionally, for designs at lower frequencies, the fixed physical size of the aperture may not be able to support high gain operation. Moreover, while the large number of switches between pads provides for graceful degradation of operation with switch failure, the complexity of the required bias networks to allow the switches to be individually addressable may make the final product too costly and bulky for its intended application. The second related antenna aperture concept uses semiconductor plasmas to form antenna structures.5 Developed by researchers at the Sarnoff Corporation, this pixel approach to a reconfigurable aperture relies on high conductivity plasma islands that are formed and

FIGURE 31-2 Conceptual drawing of a reconfigurable aperture antenna based on switched links between small metallic patches (after L. N. Pringle et al4 © IEEE 2004)



controlled by DC currents that have been injected into high-resistivity silicon-based diode structures. While the fundamental concept—that of creation of conducting patches in an aperture—is identical to the last antenna discussed, this antenna aperture differs significantly in the technology necessary for achieving the desired current distributions. Figure 31-3 contains a detailed description of the structure, including the plasma injection driver configuration.5 While elegant in its conception, the embodiment is fundamentally just as complex as the previous example that uses many discrete switches. This means that many practical and technical issues exist with this approach, including the challenge of creating the necessary carrier densities in the silicon. However, it possesses one compelling feature that other reconfigurable apertures based on conductors do not have: when turned off, the aperture possesses an extremely low radar crosssection. Similar to the previous example,3 this approach to antenna reconfigurability may be more appropriate for creating intermediate apertures that are illuminated with more conventional sources. In this case, the plasma aperture could be used to reconfigure holograms for holographic antennas in which a basic source, such as a patch or horn antenna, is used to illuminate a hologram. The aperture field for the desired pattern is then produced through interaction with the plasma aperture, enabling beamsteering and beamforming without the need for expensive phase shifters. The plasma aperture could then be designed using iterative or other means and reconfigured to deliver performance comparable to traditional phased arrays.5–6

31-5

FIGURE 31-3 Depiction of a controllable plasma grid structure for a reconfigurable aperture: (a) The grid structure is fabricated on top of the silicon wafer. (b) A cross-section of the plasma grid structure shows the layer interconnection (after A. Fathy et al5 © IEEE 2003).

31.3 RECONFIGURABLE TRAVELING WAVE ANTENNAS Traveling wave antennas are well known for delivering frequency-scanned radiation patterns. Introduction of reconfigurability into these structures adds another layer of functionality that can reduce some of the system costs inherent in previous designs. Motivation for Reconfigurable Traveling Wave Antennas Traveling wave antennas have been used with great success for delivering broadband beam scanning performance through frequency sweeps, without the need for phase shifters.



CHAPTER THIRTY-ONE

However, the necessity to change operating frequency to steer beams in a particular direction places burdens on both ends of a communication link. The reconfigurable traveling wave antennas discussed here have made real gains in delivering frequency-independent beam scanning, without compromising the effective operating bandwidth of the original base antenna. Examples of Reconfigurable Traveling Wave Antennas One recently-developed reconfigurable traveling wave antenna is a dielectric waveguide antenna, which relies on the mechanical perturbation of propagation constants at millimeterwave frequencies.7 The widths of the gratings on a thin, moveable film placed over a dielectric image line are designed to gradually perturb the propagation constant along the line, as shown in Figure 31-4.7 Viewed in light of the antenna presented in the previous section, this film serves as an effective hologram for the desired pattern. Physical shifts of the grating film change the apparent grating spacings and result in a scanned beam at a single frequency. Scan angles of up to 53 degrees have been demonstrated with this design at 35 GHz, with lower scan angles achievable over an operating band between 30 and 40 GHz.7 An extension of this concept delivers scanned, dual beam performance over a similar frequency range.8 Employing the concept of the holographic antenna,5–6 its functionality could be expanded further by enabling reconfigurability of the film itself. Another interesting reconfigurable traveling wave antenna, first developed in 1960, is the electromechanically scannable trough waveguide antenna.9 Based on an asymmetrical trough waveguide intended to feed a parabolic reflector,10 two different mechanisms were investigated to change the phase velocity in the guide, resulting in scanned radiation patterns at a single frequency. The first used the rotation of artificial dielectric structures along a longitudinal axis within the trough waveguide, and the second used a mechanical variation of the height of periodic structures located on the top of the center fin of the trough waveguide.9 While elegant in concept, the authors acknowledged that both approaches presented major mechanical problems. They were solved using cam and gear solutions, which proved to be somewhat cumbersome.

FIGURE 31-4 A traveling wave antenna based on moveable grating fed by a dielectric image line, capable of beam scanning at millimeterwave frequencies (after K. Chang et al7 © IEEE 2002)



31-7

FIGURE 31-5 Trough waveguide dimensions (left), electric field lines of the dominant TE mode with equivalent structure for design and field distribution (middle), electric field lines of the TEM mode (right) (after G. H. Huff and J. T. Bernhard11 © 2005)

Huff and Bernhard11 begin with the concept of the electromechanically scanned trough waveguide9 and then propose the application of more modern perturbations and streamlined actuation mechanisms. This allows them to achieve beamsteering at a single frequency while preserving the wide bandwidth characteristics of the antenna. The fundamental structure remains the trough waveguide, however, as shown in Figure 31-5.11 In this case, beamsteering is achieved using the micro-machined cantilever structures fabricated within the trough waveguide (shown in Figure 31-6),11 to create perturbations suitable for electrostatic or magnetostatic actuation. Simulations with actuation of individual cantilevers or groups of cantilevers demonstrate beam scanning at single frequencies through broadside for this traveling wave antenna.11 Other leaky wave antennas, which use tunable-impedance, ground planes to enable beam steering, have been developed.12 Based on the same fundamental concept of controlling wave velocity changes as in the previous example, the structure12 relies on a tunableimpedance surface equipped with hundreds of diodes to affect beam changes.

FIGURE 31-6 Trough waveguide with cantilever-type perturbations, showing staggered positions of actuated and rest states of the cantilevers (after G. H. Huff and J. T. Bernhard11 © 2005)



CHAPTER THIRTY-ONE

31.4 RECONFIGURABLE ARRAYS The signals used with phased arrays of antennas are already reconfigurable in their magnitude and phase. The addition of reconfigurability at the level of the individual array element promises to provide new levels of system performance. Motivation for Reconfigurable Arrays Of course, the one antenna topology that is already well known for its flexibility is the antenna array. This flexibility is traditionally delivered through control of the signal magnitude and phase of identical antenna elements, arranged either periodically or aperiodically, in a plane or conformally on a three dimensional surface. Historically used in defenseoriented applications, arrays are now being used in a much broader set of scenarios to provide additional functionality and security. The reconfigurable arrays described here expand typical array functionality in frequency and radiation performance, providing additional degrees of freedom for system designers and ultimately, system users. Examples of Reconfigurable Arrays One persistent issue with antenna arrays is their relatively limited frequency bandwidth. One way to improve the bandwidth of such arrays is to make the individual elements frequency reconfigurable. However, in arrays that have a ground plane, the presence of the ground plane at a fixed distance from the plane of the antennas does not enable the complete frequency reconfigurability present in the elements. In one case,13 a novel layered frequency-selective surface was implemented to support the performance of broadband reconfigurable antenna arrays. A diagram of the system concept is provided in Figure 13-7.13 With careful design, the frequency selective surfaces appear as magnetic ground planes and deliver nearly constant phase over a broad frequency band. This can enhance array gain, as shown in Figure 31-8,13 which would otherwise be limited by a fixed ground plane spacing. The same authors have applied this methodology to a wideband reconfigurable slot aperture.14

FIGURE 31-7 Depiction of a reconfigurable aperture (RECAP) array over a multilayer frequency selective surface (FSS)/ frequency selective volume (FSV) structure (after Y. E. Erdemli et al13 © IEEE 2002)



31-9

FIGURE 31-8 Diagram illustrating the coherence of direct antenna radiation and reflections from the substrate that helps to provide broadband operation of reconfigurable arrays (after Y. E. Erdemli et al13 © IEEE 2002)

Using another approach, pattern reconfigurable antennas were studied in small arrays.15–18 The individual element in the array is a reconfigurable spiral microstrip antenna, also called a reconfigurable magnetic line source antenna. This antenna, through the inclusion of solid-state or RF MEMS switches, delivers broadside and endfire radiation patterns.15–18 When arrayed, these elements can extend the scan capabilities over that achievable with a traditional planar phased array.15–18 A photograph of a 4×4 array of these elements is shown in Figure 31-9.18 Measured radiation patterns for the array, with all of the elements in broadside mode with uniform amplitude and phasing, provide a typical broadside pattern. With all of the elements in endfire mode with uniform amplitudes and 180 degree progressive phasing, the array produces a dual-beam endfire pattern.18 In this case, the patterns of the individual elements are used to reinforce the array factor created in the usual way with element amplitude and phasing. This example provides only a small sample of the array capabilities that can be achieved using individually reconfigurable antenna elements in arrays. Reconfigurable arrays with electrically tuned or switched parasitic elements can provide a great deal of pattern variability while maintaining impedance match and bandwidth. Fundamentally, tuning of antenna radiation patterns in this manner relies on the mutual coupling between closely spaced driven and parasitic elements, resulting in effective array behavior from a single feed point. Therefore, changes in radiation patterns are achieved through changes in the coupling between the elements, which, in turn, change the effective source currents on both the driven and parasitic elements. One of the fundamental examples of this antenna was proposed by Harrington in 1978.19 Shown in Figure 31-10,19 the single dipole element in the center of the array is driven, while tunable reactances on the surrounding para- FIGURE 31-9 Photograph of a 4×4 array of sitic dipoles couple to the source antenna and pattern reconfigurable elements that can be conproduce a directive beam in a desired direc- figured to deliver broadside, endfire, and scanned tion. A waveguide-based reconfigurable array radiation patterns (after G. H. Huff18 © 2006)



CHAPTER THIRTY-ONE

FIGURE 31-10 A seven-element circular array of reactively loaded parasitic dipoles for reconfigurable beam steering and beamforming (after R. F. Harrington19 © IEEE 1978)

was also proposed for the same concept.20 Since then, a number of reconfigurable parasitic arrays have been proposed and studied, using both switched and reactively loaded elements, which provide a wide variety of functionality.21–24

31.5

RECONFIGURABLE MICROSTRIP ANTENNAS

Microstrip antennas are one of the most versatile types of antennas, suitable both as single elements in small devices and in large phased arrays. As a result, any current system that uses microstrip antennas has the potential to benefit from the addition of antenna reconfigurability. Motivation for Reconfigurable Microstrip Antennas In addition to their use in the large phased arrays described earlier, more and more microstrip antennas are found in portable/mobile devices for consumers. In these applications, reconfigurability can be implemented to select operating systems and reduce interference as well as to steer radiation patterns away from noise sources or users. Moreover, their well-known operational principles, illuminated through the cavity25 and transmission line models,26 lend themselves well to enable a wide range of reconfigurability. The examples here are confined to single element versions, but many of them can also be implemented in array formations as well. Examples of Reconfigurable Microstrip Antennas In work by Bhartia and Bahl,27 the operating frequency range of a microstrip patch antenna was continuously tuned by using varactor diodes (varactors) at the radiating edges of the structure. The varactors operated with a reverse bias of between 0 and 30 volts that corresponded to capacitances of 2.4-0.4 pF. With a change in bias, the capacitance induced at the radiating edge is changed, resulting in a change in the effective electrical length of the patch. This allows for the continuous tuning of the operating frequency (though not typically of the instantaneous bandwidth) over a large band, which has been shown to be 20–30% depending on the type of microstrip antenna used.27



31-11

FIGURE 31-11 Geometry of a dual-frequency microstrip patch antenna with a switchable slot (PASS) (after F. Yang and Y. RamatSamii28 © Microwave and Optical Technology Letters 2001)

Another way to affect discrete changes in a microstrip antenna’s electrical length is to change the path of radiating currents without changing the overall footprint of the antenna. An example of this approach was proposed by Yang and Rahmat-Samii for a microstrip antenna.28 Starting with a standard rectangular microstrip antenna, a slot is etched in the patch so that it is perpendicular to the direction of the main current of the patch’s first resonance, as shown in Figure 31-11.28 A PIN diode is positioned in the center of the slot to change the current paths on the patch. If the diode is off, then currents travel around the slot and the antenna operates in a lower frequency band, while when the diode is on, the effective length of the patch is shorter and the antenna operates in a higher frequency band. The frequency ratio between the upper and lower operating frequencies is controlled by the length of the slot, and as long as the slot length is not too long, the radiation pattern of the original antenna is largely preserved.28 A similar approach can also be used to reconfigure the polarization of microstrip antennas.29 While many reconfigurable designs are based on patches such as these, other shapes and modes for microstrip antennas have been used to deliver different kinds of radiation pattern reconfigurability. For instance, one design for a square spiral microstrip antenna uses switches placed at particular positions in order to deliver a broadside or 45 degree tilted beam over a common frequency bandwidth.30 The basic antenna structure consists of a single turn square spiral antenna on an electrically thin substrate, shown in Figure 31-12,30 which exhibits a broadside radiation pattern in its fundamental state. To enable reconfiguration, the antenna is equipped with two switched connections: the first between the spiral and ground, as indicated by N (not grounded) and G (grounded) and a second in a small gap in the spiral arm as indicated by S (short) and O (open).30 When the first switch connects the line to ground and the second switch opens the in-line gap, a 45° tilt from broadside in radiation pattern results. Additionally, with activation of only the second switch, this new configuration provides a broadside radiation pattern at a higher operating frequency of 6.0 GHz.30 Others have also studied similar spiral structures that provide a wealth of possibilities for switched tilts in radiation patterns31–32 as well as examined reconfigurable behavior when the antennas are integrated into packages.33 Antennas based on a similar structure have also delivered broadside to endfire radiation characteristics with RF MEMS switched connections.34 In another example, Zhang et al35 developed a microstrip pattern reconfigurable antenna based on parasitic element tuning. The antenna, shown in Figure 31-1335 is composed of



CHAPTER THIRTY-ONE

FIGURE 31-12 Radiation pattern and frequency reconfigurable square spiral microstrip antenna showing switch placements necessary for reconfiguration (after G. H. Huff et al30 © IEEE 2003)

FIGURE 31-13 Physical structure and parameters of the reconfigurable microstrip parasitic array. Switching or tuning of parasitic elements on either side of the driven element provides beam tilt capabilities over a common impedance bandwidth (after S. Zhang35 © IEEE 2004).



31-13

a single linear element with two parasitic elements positioned parallel to the driven element. Changes in the parasitic element lengths, achieved either with electronic switches (for example, diodes, FETs, and RF MEMS switches)35 or varactors,36 change the magnitudes and phases of the currents on the parasitic elements relative to the driven element. Tilts in the main beam in one plane can then be switched35 or swept36 as the lengths of the parasitic elements are changed. This antenna, as with all parasitically tuned antennas, can be analyzed theoretically using coupling and transmission line models.37 Since microstrip designs typically include a substrate material other than air, possibilities exist, in theory, for reconfiguring antenna behavior through material changes. One example38 of a frequency-tuned ferrite-based patch antenna is presented, which provides a 40% continuous tuning range with the variable static magnetic field in the plane of the substrate and perpendicular to the resonant dimension of the patch. However, the radiation performance of the design left much to be desired, with cross-polarization levels that were significantly higher than those expected from a traditional rectangular microstrip antenna.38 Others have also investigated the properties of ferrite-based microstrip antennas,39–40 with results indicating that factors including non-uniform bias fields, as well as the multiple modes and resulting field distributions excited in a bulk ferrite substrate, may preclude their use in practical applications. One exception to this could be a design for a polarizationtunable microstrip antenna. This antenna is based on static magnetic biasing of a ferrite film, which takes advantage of the frequency tuning of the cross-polarized component of the radiated field to deliver tunable changes in polarization.41 Frequency-tunable patch designs based on the use of ferroelectric substrates or thin films, which typically have much higher relative permittivities than traditional low loss substrates, also face similar practical challenges.

31.6 RECONFIGURABLE SLOT ANTENNAS Slot antennas, in their roles as the complements of dipole antennas, find applications in a number of systems that require low profile, wide bandwidths, and bi-directional radiation patterns. It is straightforward to integrate slot antennas into existing conductive structures or ground planes without additional structural support. Motivation for Reconfigurable Slot Antennas Typically, reconfigurable slot antennas are used to enable frequency agility in a system, since they are straightforward to design, simple to construct, and compatible with a number of switch topologies. Examples of reconfigurable slot antennas described next, some of which cover four or more operating bands, illustrate not only fundamental design principles but also strategies for size reduction and direct switch integration. Examples of Reconfigurable Slot Antennas Switched frequency radiating slots with a variety of geometries and radiating properties have been proposed by a number of researchers. Gupta et al42 designed a frequencyreconfigurable compact rectangular ring slot antenna that was fed with a single slotline or coplanar waveguide line. Figure 31-14 shows a photograph of the antenna.42 Reconfigured with a total of eight PIN diode switches, the lower of the two frequencies is



CHAPTER THIRTY-ONE

FIGURE 31-14 Photograph of a reconfigurable rectangular slot antenna in two different configurations that produce operation at two separate frequencies (after K.C. Gupta et al42 © IEEE 2000)

set by the perimeter of the outer loop. The upper frequency is set by the perimeter of the inner loop that was formed when shorter slot sections are activated in place of portions of the outer loop. Peroulis et al43 developed a tunable antenna using four PIN diode switches that changed the effective length of the slot so it could operate in one of four different frequency bands between 530 MHz and 890 MHz without the need for changes in the matching network. A coplanar, waveguide-fed slot antenna utilizing similar principles for frequency tuning with PIN diodes is shown in Figure 31-15.44 Often, the available switching technology can affect the choice of antenna geometry. For example,45 a hybrid coplanar waveguide slot antenna with a 22% effective bandwidth (shown in Figure 3116)45 was designed to directly accommodate a particular RF MEMS switch geometry, making final fabrication and integration more straightforward. Using tunable reactive loading, rather than switches, can result in continuously tuned slot antennas. In one case,46 a one-wavelength slot antenna loaded with two one-port reactive FET components was tuned continuously. By changing the bias voltage, the reactances of the FETs were varied and the effective length of the slot was changed, thereby changing its operating frequency. The range of tuning varied over roughly 1 GHz centered on 10 GHz, producing a 10% tuning band. For this relatively small tuning range, the radiation patterns were essentially unchanged.46 Similar tunable slot antennas loaded with varactors47–48 take advantage of higher order resonances to create tunable dual-band performance. Using a transmission line model of the loaded slot resonator, the position of the tuning elements can be determined to enable relatively independent tuning of the two bands.48 Slot antennas have also been implemented for the purpose of delivering polarization reconfigurability. In one case, the slot-ring antenna49 uses PIN diodes to reconfigure between linear and circular polarization or between two circular polarizations. The general topology

FIGURE 31-15 Frequency reconfigurable coplanar waveguide-fed slot antenna with eight diode switches and associated bias lines (after J. M. Laheurte44 © IET 2001)



31-15

FIGURE 31-16 Reconfigurable hybrid folded slot/slot dipole antenna: antenna geometry including all three switch locations (left) and photograph of the fabricated antenna using only two of the three switches (right) (after G. H. Huff and J. T. Bernhard45 © IEEE 2005)

of the antenna is shown in Figure 31-17,49 with the specific diode positions, biasing, and ground plane configurations for both designs shown in Figure 31-18.49 For the linear-circular design shown in Figure 31-18a, forward-biasing the diodes across the small discontinuities at 45° and −135° (relative to the feed point) produces linear polarization, while reverse biasing the diodes delivers circular polarization.49 The design in Figure 31-18b goes one step further and adds additional symmetric discontinuities to support switching between left- and right-handed circular polarizations.49 In both designs, care is taken to divide the ground planes carefully to support proper DC biasing for the diodes while providing RF continuity through capacitors connected between ground plane sections. This antenna is a good example of the additional factors that must be taken into account when transitioning from a fixed to a reconfigurable antenna—the fundamental structure may remain the same, but critical adjustments are required to enable proper DC connections and RF performance.

FIGURE 31-17 Basic topology of a microstrip-fed circularly polarized slot ring antenna (after M. K. Fries et al49 © IEEE 2003)



CHAPTER THIRTY-ONE

FIGURE 31-18 Two polarization reconfigurable slot ring antennas: (a) switchable between linear and left hand circular polarization and (b) switchable between left and right hand circular polarization (after M. K. Fries et al49 © IEEE 2003)

31.7 RECONFIGURABLE MONOPOLE/ DIPOLE ANTENNAS Used in everything from phased arrays to portable communication devices, monopole and dipole antennas are both elemental and ubiquitous. While the theory explaining their operation is well known and based on a simple resonance principle, their application in specific situations typically requires adjustments to their length and packaging in order to deliver operation in one particular band. Motivation for Reconfigurable Monopole/Dipole Antennas When multifunctional systems require operation at multiple frequencies, adjustments to the basic monopole/dipole design can be accomplished either by creating a multiband structure, again based on fundamental or higher order resonances, or by making the antenna frequency reconfigurable. In some cases, the multiband approach is the more sensible within cost and complexity constraints, but frequency reconfigurable monopole/dipole antennas have roles to play where size, cost, noise immunity, security, and low observability are important factors. The reconfigurable monopoles and dipoles discussed here change the effective resonant length of the antenna by adding or removing part of the antenna length through electronic, optical, mechanical, or other means. Examples of Reconfigurable Monopole/Dipole Antennas Different kinds of switches (for example, PIN diodes, FETs, optical switches, and RF MEMS) have been implemented by several groups in frequency-tunable dipole and monopole antennas for various frequency bands. Roscoe et al implemented four PIN diodes in a printed dipole antenna to deliver three operating bands between 5.2 and 5.8 GHz.50 For this specific design, the researchers limited the reconfiguration range so that a single feed network could be used for all of the operating bands. Kiriazi et al. present an example of a frequency reconfigurable antenna using RF MEMS switches.51 In this work, opening and closing a pair of RF MEMS switches reconfigures a simple dipole antenna printed on a high resistivity silicon substrate, allowing it to operate in one of two frequency bands. Others have applied the same approach to develop frequency reconfigurable Yagi antennas.52



31-17

FIGURE 31-19 Reconfigurable three dimensional fractal tree dipole antennas with four branch levels (after J. S. Petko and D. H. Werner53 © IEEE 2004)

More complicated structures based on fractal shapes, but sharing the same underlying principle, have been studied by numerous researchers. For instance, a three-dimensional fractal tree structure is proposed for use with either passive frequency traps or RF MEMS switches to deliver operation over multiple frequency bands.53 An example of the evolution of the structure is shown in Figure 31-19.53 For this antenna, the design methodology can produce operation over multiple contiguous bands, which is only limited by the practicality of implementing the necessary number of switches. Anagnostou et al. propose a reconfigurable monopole based on a Sierpinski gasket, with RF MEMS switches for connecting different sections of the antenna together in order to provide multiple operating bands.54 Subsequent work, with the direct integration of RF MEMS switches in a dipole design and shown in Figure 31-20,55 provided three separate operating bands with similar omnidirectional radiation characteristics.55 Direct fabrication of the RF MEMS switches during the fabrication of the antenna helps to minimize package parasitics that may detune or limit the frequency behavior of the antenna.

FIGURE 31-20 Photograph of a reconfigurable planar Sierpinski dipole antenna with integrated RF MEMS switches (after D. E. Anagnostou et al55 © IEEE 2006)



CHAPTER THIRTY-ONE

FIGURE 31-21 Diagram of a reconfigurable monopole antenna switchable between four frequency bands with optoelectronic switches (after J. L. Freeman et al56 © IET 1992)

Using the same concept but different switch technologies, Freeman et al changed the effective length of a monopole antenna using optical switches. This helped to eliminate many of the switch and bias line effects that can occur with other kinds of switches.56 A diagram of this antenna is shown in Figure 31-21.56 An excellent example of reconfigurable dipole antennas improving system performance is presented by Piazza and Dandekar.57 In this work, reconfigurable printed dipole arrays using PIN diodes have been implemented to enable multiple-input multiple-output (MIMO) orthogonal frequency division multiplexing (OFDM) systems.57 For this application, a set of two printed frequency reconfigurable dipoles were spaced a quarter wavelength apart at 2.45 GHz. They were then included at both the transmitter and receiver of a MIMO communication system, shown in Figure 31-22.57 In the short configuration of the antennas,

FIGURE 31-22 Diagram of an array of two reconfigurable dipoles used to achieve pattern diversity at a single operating frequency in MIMO systems (after D. Piazza and K. R. Dandekar7 © IET 2006)



31-19

the elements operated as half-wavelength dipoles. In the long configuration, the antennas were driven in a higher order mode where they were three quarter wavelengths long at the same frequency of 2.45 GHz. By changing the lengths of the antennas but by keeping the operating frequency the same, the researchers were able to achieve four different radiation patterns with each of the arrays, thereby providing a degree of pattern reconfigurability that helped to improve MIMO system capacity.57 It is worthwhile to note that implementation of switches, in both monopoles and dipoles, often prevents the antenna from being a self-supporting structure. The need for surfaces for switch and bias line connections usually requires some kind of planar or other dielectric support, which may distort radiation patterns and increase antenna loss and weight slightly.

31.8 CONCLUDING REMARKS Antenna reconfigurability itself isn’t new, but the opportunities for integrating reconfigurable antennas into systems that can fully exploit antenna capabilities are. These opportunities include software defined radio, cognitive radio, MIMO systems, multifunction consumer wireless devices, and high performance phased arrays. As all of the examples discussed in this chapter demonstrate, new antenna designs can be developed to change their frequencies, bandwidths, polarizations, and radiation patterns. They can be designed to solve specific problems now as well as to make systems more flexible and responsive to future operational requirements. However, the challenges for the future of this area of antenna engineering do not lie only in antenna design. Two other significant areas for future work include the development of appropriate control networks and the wide adoption of antenna reconfigurability as an important feature in overall system design. While initial efforts to demonstrate the benefits of reconfigurability in a system-level context have been made by researchers such as Piazza and Dandekar,57 a great deal remains to be done. As this work progresses, these new capabilities can be used effectively to deliver high throughput, high reliability, and high security wireless connections in the future.

REFERENCES 1. P. J. B. Clarricoats and H. Zhou, “The Design and Performance of a Reconfigurable Mesh Reflector Antenna,” Proc. IEE Seventh Int. Conf. on Antennas and Propagat., vol. 1 (April 1991): 322–325. 2. P. J. B. Clarricoats, H. Zhou, and A. Monk, “Electronically Controlled Reconfigurable Reflector Antenna,” Proc. IEEE Int. Symp on Antennas and Propagation, vol. 1 (June 1991): 179–181. 3. G. Washington, H. S. Yoon, M. Angelino, and W. H. Theunissen, “Design, Modeling, and Optimization of Mechanically Reconfigurable Aperture Antennas,” IEEE Trans. Antennas Propagat., vol. 50, no. 5 (May 2002): 628–637. 4. L. N. Pringle, P. H. Harms, S. P. Blalock, G. N. Kiesel, E. J. Kuster, P. G. Friederich, R. J. Prado, J. M. Morris, and G. S. Smith, “A Reconfigurable Aperture Antenna Based on Switched Links Between Electrically Small Metallic Patches,” IEEE Trans. Antennas Propagat., vol. 52, no. 6 (June 2004): 1434–1445. 5. A. Fathy, A. Rosen, H. Owen, F. McGinty, D. McGee, G. Taylor, R. Amantea, P. Swain, S. Perlow, and M. ElSherbiny, “Silicon-Based Reconfigurable Antennas—Concepts, Analysis, Implementation, and Feasibility,” IEEE Trans. Microwave Theory and Techn., vol. 51, no. 6 (June 2003): 1650–1661. 6. M. ElSherbiny, A. E. Fathy, A. Rosen, G. Ayers, and S. M. Perlow, “Holographic Antenna Concept, Analysis and Parameters,” IEEE Trans. Antennas Propag., vol. 52, no. 3 (March 2004): 830–839.



CHAPTER THIRTY-ONE

7. K. Chang, M. Li, T-Y. Yun, and C. T. Rodenbeck, “Novel Low-Cost Beam Steering Techniques,” IEEE Trans. Antennas Propag., vol. 50, no. 5 (May 2002): 618–627. 8. C. T. Rodenbeck, M. Li, and K. Chang, “Design and Analysis of a Reconfigurable Dual-Beam Grating Antenna for Low-Cost Millimeter-Wave Beam-Steering,” IEEE Trans. Antennas Propag., vol. 52, no. 4 (April 2004): 999–1006. 9. W. Rotman and A. Maestri, “An Electromechanically Scannable Trough Waveguide Array,” Proc. IRE International Convention Record, vol. 8 (March 1960): 67–83. 10. W. Rotman and A. Oliner, “Asymmetrical Trough Waveguide Antennas,” IEEE Trans. Antennas Propag., vol. 7 (April 1959): 153–162. 11. G. H. Huff and J. T. Bernhard, “Electromechanical Beam Steering of a Trough Waveguide Antenna Using Cantilever Perturbations,” Proc. 2005 Antenna Applications Symposium (September 2005): 152–165. 12. D. Sievenpiper, J. Schaffner, J. J. Lee, and S. Livingston, “A Steerable Leaky Wave Antenna Using a Tunable Impedance Ground Plane,” IEEE Antennas and Wireless Propagation Letters, vol. 1 (2002): 179–182. 13. Y. E. Erdemli, K. Sertel, R. A. Gilbert, D. E. Wright, and J. L. Volakis, “Frequency-Selective Surfaces to Enhance Performance of Broad-Band Reconfigurable Arrays,” IEEE Trans. Antennas Propag., vol. 50 (December 2002): 1716–1724. 14. Y. E. Erdemli, R. A. Gilbert, and J. L. Volakis, “A Reconfigurable Slot Aperture Design Over a Broad-Band Substrate/Feed Structure,” IEEE Trans. Antennas Propag., vol. 52 (November 2004): 2860–2870. 15. G. H. Huff, J. Feng, S. Zhang, and J. T. Bernhard, “Behavior of Pattern and/or Frequency Reconfigurable Antennas in Small Arrays,” Proc. 2003 IEEE/URSI Int. Symp. on Antennas and Propagation, URSI (June 2003): 151. 16. G. H. Huff, J. Feng, and J. T. Bernhard, “A Small Array of Broadside to Endfire Radiation Reconfigurable Antennas,” Proc. 2003 Antenna Applications Symposium (Sept. 2003): 147–161. 17. K. Hietpas, G. H. Huff, and J. T. Bernhard, “Investigation of Phased Array Beamsteering Using Reconfigurable Antennas,” Proc. IASTED Int. Conf. on Antennas, Radar, and Wave Propagation (July 2004): 41–44. 18. G. H. Huff, “A Radiation Reconfigurable Magnetic Line Source Antenna: Modeling, Integration with RF MEMS, and Applications,” Ph.D. dissertation, University of Illinois at UrbanaChampaign, 2006. 19. R. F. Harrington, “Reactively Controlled Directive Arrays,” IEEE Trans. Antennas Propag., vol. 26 (May 1978): 390–395. 20. J. Luzwick and R. Harrington, “A Reactively Loaded Aperture Antenna Array,” IEEE Trans. Antennas Propag., vol. 26, no. 4 (July 1978): 543–547. 21. D. V. Thiel, “Switched Parasitic Antennas and Controlled Reactance Parasitic Antennas: A Systems Comparison,” Proc. IEEE Antennas and Propagation Int. Symp., vol. 3 (June 2004): 3211–3214. 22. R. Schlub, D. V. Thiel, J. W. Lu, and S. G. O’Keefe, “Dual-Band Six-Element Switched Parasitic Array for Smart Antenna Cellular Communications,” Electron. Lett., vol. 36, no. 16 (August 2000): 1342–1343. 23. R. Schlub, J. Lu, and T. Ohira, “Seven-Element Ground Skirt Monopole ESPAR Antenna Design from a Genetic Algorithm and Finite Element Method,” IEEE Trans. Antennas Propag., vol. 51, no. 11 (November 2003): 3033–3039. 24. T. Ohira and K. Gyoda, “Hand-Held Microwave Direction-of-Arrival Finder Based on VaractorTuned Analog Aerial Beamforming,” Proc. IEEE Asia Pacific Conference (Taipei), vol. 2 (December 2001): 585–588. 25. Y. T. Lo, D. Solomon, and W. Richards, “Theory and Experiment on Microstrip Antennas,” IEEE Trans. Antennas Propag., vol. 27, no. 2 (March 1979): 137–145. 26. A. Derneryd, “Linearly Polarized Microstrip Antennas,” IEEE Trans. Antennas Propag., vol. 24, no. 6 (November 1976): 846–851.



31-21

27. P. Bhartia and I. J. Bahl, “Frequency Agile Microstrip Antennas,” Microwave J., vol. 25 (October 1982): 67–70. 28. F. Yang and Y. Rahmat-Samii, “Patch Antenna with Switchable Slot (PASS): Dual Frequency Operation,” Microwave and Optical Technology Letters, vol. 31 (November 2001): 165–168. 29. F. Yang and Y. Rahmat-Samii, “A Reconfigurable Patch Antenna Using Switchable Slots for Circular Polarization Diversity,” IEEE Microwave and Wireless Components Lett., vol. 12 (March 2002): 96–98. 30. G. H. Huff, J. Feng, S. Zhang, and J. T. Bernhard, “A Novel Radiation Pattern and Frequency Reconfigurable Single Turn Square Spiral Microstrip Antenna,” IEEE Microwave and Wireless Components Lett., vol. 13, no. 2 (February 2003): 57–59. 31. C. Jung, M. Lee, G. P. Li, and F. De Flaviis, “Reconfigurable Scan-Beam Single-Arm Spiral Antenna Integrated with RF-MEMS Switches,” IEEE Trans. Antennas Propag., vol. 54, no. 2 (February 2006): 455–463. 32. A. Mehta, D. Mirshekar-Syahkal, and H. Nakano, “Beam Adaptive Single Arm Rectangular Spiral Antenna with Switches,” IEE Proceedings-Microwaves, Antennas and Propagation, vol. 153, no. 1 (February 2006): 13–18. 33. G. H. Huff, J. Feng, S. Zhang, and J. T. Bernhard, “Directional Reconfigurable Antennas on Laptop Computers: Simulation, Measurement, and Evaluation of Candidate Integration Positions.” IEEE Trans. Antennas Propag., vol. 52 (December 2004): 3220–3227. 34. G. H. Huff and J. T. Bernhard, “Integration of Packaged RF MEMS Switches with Radiation Pattern Reconfigurable Square Spiral Microstrip Antennas,” IEEE Trans. Antennas Propag., vol. 54, no. 2 (February 2006): 464–469. 35. S. Zhang, G. H. Huff, J. Feng, and J. T. Bernhard, “A Pattern Reconfigurable Microstrip Parasitic Array,” IEEE Trans. Antennas Propag., vol. 52 (October 2004): 2773–2776. 36. S. Zhang, G. Huff, G. Cung, and J. T. Bernhard, “Three Variations of a Pattern Reconfigurable Microstrip Parasitic Array,” Microwave and Optical Technology Letters, vol. 45 (June 2005): 369–372. 37. S. Zhang, “A Pattern Reconfigurable Microstrip Parasitic Array: Theory, Design, and Applications,” Ph.D. dissertation, University of Illinois at Urbana-Champaign, 2005. 38. D. M. Pozar and V. Sanchez, “Magnetic Tuning of a Microstrip Antenna on a Ferrite Substrate,” Electron. Lett., vol. 24 (June 1988): 729–731. 39. R. K. Mishra, S. S. Pattnaik, and N. Das, “Tuning of Microstrip Antenna on Ferrite Substrate,” IEEE Trans. Antennas Propag., vol. 41, no. 2 (February 1993): 230–233. 40. A. D. Brown, J. L. Volakis, L. C. Kempel, and Y. Botros, “Patch Antennas on Ferromagnetic Substrates,” IEEE Trans. Antennas Propag., vol. 47 (January 1999): 26–32. 41. P. Rainville and F. Harackiewicz, “Magnetic Tuning of a Microstrip Patch Antenna Fabricated on a Ferrite Film,” IEEE Microwave and Guided Wave Lett., vol. 2 (December 1992): 483–485. 42. K. C. Gupta, J. Li, R. Ramadoss, and C. Wang, “Design of Frequency-Reconfigurable Slot Ring Antennas,” Proc. IEEE/URSI Int. Symp on Antennas and Propagation, vol. 1 (July 2000): 326. 43. D. Peroulis, K. Sarabandi, and L. P. B. Katehi, “Design of Reconfigurable Slot Antennas,” IEEE Trans. Antennas Propag., vol. 53 (February 2005): 645–654. 44. J. M. Laheurte, “Switchable CPW-fed Slot Antenna for Multifrequency Operation,” Electronics Letters, vol. 37, no. 25 (December 2001): 1498–1500. 45. G. H. Huff and J. T. Bernhard, “Frequency Reconfigurable CPW-fed Hybrid Folded Slot/ Slot Dipole Antenna,” Proc. IEEE/ACES Int. Conf. on Wireless Communications and Applied Computational Electromagnetics (April 2005): 574–577. 46. S. Kawasaki and T. Itoh, “A Slot Antenna with Electronically Tunable Length,” Proc. IEEE/URSI Int. Symp. on Antennas and Propagation, vol. 1 (June 1991): 130–133. 47. N. Behdad and K. Sarabandi, “A Varactor-Tuned Dual-Band Slot Antenna,” IEEE Trans. Antennas Propag., vol. 54, no. 2 (February 2006): 401–408. 48. N. Behdad and K. Sarabandi, “Dual-Band Reconfigurable Antenna with a Very Wide Tunability Range,” IEEE Trans. Antennas Propag., vol. 54, no. 2 (February 2006): 409–416.



CHAPTER THIRTY-ONE

49. M. K. Fries, M. Grani, and R. Vahldieck, “A Reconfigurable Slot Antenna with Switchable Polarization,” IEEE Microwave and Wireless Components Letters, vol. 13, no. 11 (November 2003): 490–492. 50. D. J. Roscoe, L. Shafai, A. Ittipiboon, M. Cuhaci, and R. Douville, “Tunable Dipole Antennas,” Proc. IEEE Int. Symp. on Antennas and Propagation, vol. 2 (July 1993): 672–675. 51. J. Kiriazi, H. Ghali, H. Radaie, and H. Haddara, “Reconfigurable Dual-Band Dipole Antenna on Silicon using Series MEMS Switches,” Proc. IEEE Int. Symp. on Antennas and Propagation, vol. 1 (June 2003): 403–406. 52. M. A. Ali and P. Wahid, “A Reconfigurable Yagi Array for Wireless Applications,” Proc. IEEE/ URSI Int. Symp on Antennas and Propagation, vol. 1 (2002): 466–468. 53. J. S. Petko and D. H. Werner, “Miniature Reconfigurable Three-Dimensional Fractal Tree Antennas,” IEEE Trans. Antennas Propag., vol. 52, no. 8 (August 2004): 1945–1956. 54. D. Anagnostou, M. T. Chryssomallis, J. C. Lyke, and C. G. Christodoulou, “Re-configurable Sierpinski Gasket Antenna using RF-MEMS Switches,” Proc. IEEE Int. Symp. on Antennas and Propagation, vol. 1 (June 2003): 375–378. 55. D. E. Anagnostou, G. Zheng, M. T. Chryssomallis, J. C. Lyke, G. E. Ponchak, J. Papapolymerou, and C. G. Christodoulou, “Design, Fabrication, and Measurements of an RF-MEMS-based SelfSimilar Reconfigurable Antenna,” IEEE Trans. Antennas Propag., vol. 54, no. 2 (February 2006) 422–432. 56. J. L. Freeman, B. J. Lamberty, and G. S. Andrews, “Optoelectronically Reconfigurable Monopole Antenna,” Electron. Lett., vol. 28, no. 16 (July 1992): 1502-1503. 57. D. Piazza and K. R. Dandekar, “Reconfigurable Antenna Solution for MIMO-OFDM Systems,” Electron. Lett., vol. 42, no. 8 (April 2006): 446–447.



Chapter 32

Active Antennas Zoya Popovic´ Nestor Lopez Sebastien Rondineau University of Colorado at Boulder

CONTENTS 32.1 INTRODUCTION AND BASIC TERMS . . . . . . . . . . . . . . . . . . . . . . .

32-2

32.2 FREQUENCY-AGILE ACTIVE ANTENNAS. . . . . . . . . . . . . . . . . . . .

32-9

32.3 OSCILLATOR ANTENNAS AND ARRAYS . . . . . . . . . . . . . . . . . . . . 32-11 32.4 AMPLIFIER ANTENNAS AND ARRAYS . . . . . . . . . . . . . . . . . . . . . . 32-17 32.5 FREQUENCY-CONVERTING ANTENNAS. . . . . . . . . . . . . . . . . . . . 32-23 32.6 RECTENNAS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32-26 32.7 OTHER ACTIVE ANTENNAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32-30


Active Antennas 32-2

CHAPTER THIRTY-TWO

32.1 INTRODUCTION AND BASIC TERMS Antennas are used either in transmission, in which case they are fed from the output of some amplifier, or in reception, in which case they provide the RF input to an LNA. Therefore, in some sense, all antennas are active. In the context of this chapter, however, an active antenna implies an antenna integrated intimately with an active circuit, including the DC bias circuit, and without an isolator or circulator between them. The absence of isolator/ circulator implies that neither the antenna nor the circuit needs to be designed in a 50-Ω environment. Figure 32-1 shows block diagrams of several types of active antennas, classified according to their functionality. Note that this chapter does not discuss in detail phased array antennas. This is a separate and large topic in itself, with excellent overviews given in the literature1,2 as well as in papers on T/R modules.3

Integrated antenna

Tuning circuit

RF power IN/OUT

Oscillator circuit

RF power IN/OUT

RF power OUT

Antenna feed DC power IN (control)

DC power IN (control)

(a)

(b) Amplifier circuit

RF power IN

RF power OUT

Input antenna

Output antenna


Low noise amplifier

Power amplifier

RF power IN

RF power IN

RF power OUT

Output antenna

Input antenna DC power IN (control)

RF power OUT

(c)


FIGURE 32-1 Schematic diagrams of a few examples of active antennas: (a) actively tuned antenna, (b) oscillator antenna, and (c) amplifier antenna with repeater, receiver, and transmitter subclasses. An active antenna implies an antenna integrated intimately with an active circuit.



ACTIVE ANTENNAS

−VGS

Radiating edges

Slot antenna DC block Microstrip line

RF block

Patch

RF block

Varactor diode

RF null RF block

RF / DC ground

VDS

DC bias control

Via to DC/RF ground

(a)

(b)

VDS

RF block

−VGS

Eout

RF block

Via to DC/RF ground Ein

Pout = Gactive Pin

Slot antennas

(c)

Frequency selective volume

f1 3f1

(d)

FIGURE 32-2 Illustration of active antennas classified by their functionality: (a) a frequency-agile, in this case continuously tunable, slot antenna with a varactor diode active device; (b) an oscillator patch antenna loaded with a transistor; (c) a dual-slot amplifier repeater antenna; (d ) a frequency-doubling slot antenna loaded with a Schottky diode

A Classification of Active Antennas Antennas are normally classified in terms of operational bandwidth (broadband and narrowband), implementation (e.g. printed, wire, or aperture), etc. In this chapter, it is appropriate to classify active antennas by the function of the active circuit. The following types are discussed at more length throughout this chapter, with an illustration of each type given in Figure 32-2: ●

Frequency-agile antennas A two- or three-terminal active device can be designed into the antenna to enable the antenna impedance to tune with frequency. Figure 32-2a shows an example of this type of active antenna: a slot microstrip feedline contains a varactor diode tuning element, similar to the demonstrations in Forman and Popovic4 and Behdad and Sarabandi.5 When the capacitance of the diode changes, the electrical length of the antenna, which in turn depends on the antenna reactance, changes and the antenna becomes resonant at a different frequency. In this case, DC power is used to provide increased bandwidth of an antenna element. These antennas find applications in multifunctional systems with multiple nonsimultaneous carriers.


Active Antennas 32-4 ●

●

●

●

CHAPTER THIRTY-TWO

Oscillator antennas A two- or three-terminal negative-resistance device can be con-

nected directly to the terminals of a single antenna element or an array of elements. In this case, DC power is converted to radiated RF power. An example of a patch antenna in the feedback loop of a transistor, similar to the one in Chang et al,6 is shown in Figure 32-2b. Oscillator antennas have been discussed for applications such as low-cost sensors,7 power combining,8,9,10 and synchronized scanning antenna arrays.11 Amplifier antennas An active device is connected to the terminals of an antenna element to provide amplification in receive mode or transmit mode. In the former case, the matching between the antenna and active element usually optimizes noise, while in the latter case, the matching optimizes power and/or efficiency. Figure 32-2c shows an example of a repeater element with two slot antennas and a prematched amplifier chip, similar to the work in Hollung et al.12 In this case, increase in gain is enabled by adding DC power to the antenna, and it becomes difficult to separate antenna gain from circuit gain. Amplifier antennas find applications in transmitters where spatial power combining can be achieved with an array, and in receivers where the feedline loss, which contributes to the total noise figure, can be eliminated by directly connecting an LNA to the receiving antenna. Frequency-conversion antennas A two- or three-terminal active device integrated with an antenna can provide direct down or up conversion of a radiated signal, at frequencies that are direct harmonics or subharmonics of a fundamental frequency (multipliers, dividers), or at frequencies with a given offset from the operating frequency (mixers). Figure 32-2d shows an example of a slot antenna with a Schottky diode, which can be used for frequency doubling since the slot is matched to the diode impedance at both the input frequency and its harmonic.13 Such antennas have applications in receivers,14 mixers with high dynamic range,15 detectors for millimeter-wave and THz receivers,16 phaseconjugating RFID type antennas,17 and high-frequency generation.18 A special case of frequency-conversion antennas is when a two- or three-terminal rectifying device is connected directly to the terminals of a receiving antenna in such a way that the received RF power is converted with optimal efficiency to DC power, while harmonic production and re-radiation is minimized. This type of active antenna is referred to as a rectenna. Such antennas have been applied to RFID tags,19 sensor powering for cases when there is no solar power and where it is difficult to replace batteries,20 directed narrow-beam array power beaming,21 and for energy recycling and/or scavenging.22 Other active antennas A number of antennas in which the polarization properties (see e.g., Peroulis et al23) or radiation pattern can be varied based on control of an active device have been demonstrated. A separate chapter in this book covers reconfigurable antennas, and in this chapter we give a brief description of a few examples of such antennas, including an example of optically controlled antennas and arrays.24

Figures of Merit for Active Antennas Most standard antenna figures of merit can be applied to active antennas, but care must be taken when defining all properties that include power, since some of the power comes from external DC or RF sources. Properties that do not include power and are defined as in the case of passive antennas are ●

●

●

Polarization (linear, circular, elliptical; described by an axial ratio) Bandwidth (usually defined through return loss, but it is important to also quantify radiation patterns and efficiency, i.e., gain vs. frequency) Normalized power radiation pattern Properties that do include power, for which careful definitions are necessary, defined in the far field are


Active Antennas ACTIVE ANTENNAS

●

Efficiency, defined as

η=

●

Pradiated , PT + Ploss

where the power loss in the antenna can actually be gain if only RF power is taken into account. Gain, defined as G (θ , φ ) = η

●

32-5

S (r ,θ , φ ) 4π r 2 S (r ,θ , φ ) =η Si (r ) PT

The transmitted power PT is not easy to determine when the antenna and active circuit are directly integrated. In such cases, the following quantity is relevant: Effective radiated power (ERP) with respect to an isotropic source (thus referred to sometimes as EIRP): ERP = PT G measured for active transmitting antenna by measuring power density in the far field and assuming it is radiated by an isotropic radiator.

●

Effective area, defined as A(θ , φ ) =

●

PR λ 2 ⋅ G (θ , φ ) = S (θ , φ ) 4π

where G could be the passive or active antenna gain. The passive gain is difficult to access in an integrated antenna, while the active gain yields an effective area that can be larger than the geometric area due to the gain of the active device, and at the expense of DC power. Power combining efficiency, relevant in the case of arrays of active transmitting and receiving antennas, and defined as PCE array =

●

●

ERP G ⋅ N ⋅ P1

where the ERP is measured for the entire array, G is found from the geometric area of the array or measured pattern, N is the number of array elements, and P1 is the power of a single isolated element. Isotropic conversion loss of a frequency-conversion active antenna has been defined first in Stephan and Itoh25 as P L Liso = IF = PRF G Gant where the measured quantities are the ERP and the down-converted power at the IF frequency, L is the conversion loss of the mixer, and Gant the gain of the antenna. The meaning of this expression is that the same mixer, impedance matched to two different antennas, will have a higher IF output power when connected to an antenna with higher gain. Isotropic conversion gain of an oscillator antenna is defined as Giso =

ERP PRF G = ηG = PDC PDC

where the measured quantities are the ERP and the input DC power. The antenna gain can be obtained by simulation or estimated from the measured radiation pattern, which then provides an estimate of the radiated RF power and efficiency of the oscillator circuit. Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.


CHAPTER THIRTY-TWO

Modeling of Active Antennas and Limitations of Current CAD Tools The most efficient way to model an active antenna is to start at the highest level of abstraction, which is a circuit simulator. The antenna can be included as a frequency-variable impedance obtained from a full-wave field simulation. This, however, does not give information about the antenna gain, polarization, or radiation pattern. The design of an active antenna involves direct matching between the active circuit and antenna. Antennas have a complex impedance with both real and imaginary parts, which vary in frequency. The impedance of active devices is also frequency-dependent. Therefore, matching both the antenna and active device to a common 50-Ω impedance is not necessarily optimal, although it is common practice due to the fact that the antenna is usually designed separately from the active circuit. An example of an amplifier antenna in which the impedance of the antenna is not 50 Ω, but is designed to have specific complex impedances at both the fundamental and second harmonic frequencies, is given in Weiss and Popovic.26 The impedance of a nonresonant microstrip-circuit-fed antenna is in this case chosen to optimize transistor efficiency by wave-shaping of the voltage and currents. Consider the Smith chart plots in Figure 32-3. The feedpoint reflection coefficient with respect to 50 Ω for two antennas is plotted over a broad frequency range. One antenna is a linearly polarized patch centered at 10 GHz and plotted from 8 to 12 GHz in Figure 32-3a. The other is a self-complementary Archimedean spiral antenna that operates above 1.7 GHz and is simulated up to 18 GHz. In an integrated active antenna, the impedance of an antenna needs to be matched to that of the active device. Figure 32-3b shows the input and output reflection coefficients into 50 Ω for a bipolar and FET transistor from 0.5 to 18 GHz, showing significant variation over frequency. Matching either antenna from Figure 32-3a to a transistor over a range of frequencies can be done in a number of ways. As an illustration, consider a unit cell, shown in Figure 32-4a, of an amplifier array.27 An off-center-fed 100-Ω second-resonant slot antenna is the output of a two-stage LNA, while a folded slot receives the input wave in a configuration corresponding to that in Figure 32-1c (top). The off-center-fed second-resonance slot and the folded slot were designed using a method of moments code, where the off-center-fed slot result is shown in Figure 32-4b. +j1.0

+j1.0 +j0.5

BJT S11

+j0.5

+j2.0

+j2.0

FET S11 BJT S22

patch S11

−j0.2

−j5.0

−j0.5

−j2.0 −j1.0

(a)

5.0

2.0

0.0

FET S22

1.0

∞

+j0.2 0.5

5.0

2.0

1.0

0.5

0.2

0.0

+j5.0

0.2

+j0.2

spiral S11

−j0.2

+j5.0

∞

−j5.0

−j0.5

−j2.0 −j1.0

(b)

FIGURE 32-3 Smith chart plots of (a) antenna feedpoint reflection coefficient (with respect to 50 Ω) for a narrowband linearly polarized (CP) patch antenna plotted from 8 to 12 GHz, and a broadband CP spiral antenna (0.5 to 18 GHz); and (b) input and output S-parameter variation of a Silicon bipolar transistor (NE6800f ) and a GaAs FET (NE7130Na) from 0.5 to 18 GHz. The BJT has gain up to around 10 GHz, while the FET operates to 18 GHz.



ACTIVE ANTENNAS

0

Return loss [dB]

−10

−20

−30

−40 WireZeus simulation Measured results −50

5

10 Frequency [GHz]

(a)

15

(b)

FIGURE 32-4 (a) An off-center-fed 100-Ω second-resonant slot antenna is the output of a two-stage LNA, while a folded slot receives the input wave in a configuration corresponding to that in Figure 32-1c. The printed slot antennas are designed using full-wave analysis. (b) Off-center-fed slot return loss.

The obtained frequency-variable impedances are then inserted into the Agilent ADS circuit simulator to check for stability. It was found that the amplifier was unstable, and a resistively loaded stabilization network that also serves as a bias network was added. Full-wave simulations were used not just for the antenna element design, but also to verify coupling inside a passive unit cell (with a thru-line instead of the transistor). Since the gain of the LNA is around 25 dB in this case, the coupling between the input and output orthogonally polarized antennas needs to be well below this value to ensure stability. Alternatively, some electromagnetic field simulators have the option of including an S-parameter multiport network in the full-wave simulation. Figure 32-5 shows an example simulation of a transistor with a patch antenna at its output port, with a general connection block between the drain of the FET and the patch antenna. First, the FET from Figure 32-3b is directly connected to the square patch antenna at the radiating edge. Ideally, the patch antenna alone patch antenna directly connected to FET 30 patch antenna matched at 50Ω and FET matched at 50Ω patched antenna matched at the FET output 25

a Patch antenna

Drain bias line

FET-Antenna connection

20 Gain [dB]

a

Vds

15 10 5 0

RF input

−5 8

(a)

9

10 Freq [GHz]

11

12

(b)

FIGURE 32-5 (a) Patch antenna matched to a transistor input impedance, with added bias line through the patch null. (b) Simulated active antenna gain for transistor directly connected to the patch at the radiating edge, both antenna and transistor matched to 50 Ω, and transistor directly matched to the antenna with a single nonoptimized matching section.



CHAPTER THIRTY-TWO

radiating-edge-fed square patch is linearly polarized with symmetric E- and H-plane patterns. The resulting active antenna gain, defined as GAA = ERP / PRF-in , is shown in Figure 32-5b. Then, both antenna and transistor are matched to 50 Ω and the gain of the cascade is shown in the same plot. This is the standard design method and gives improved gain at the expense of space used by two matching networks. Finally, the transistor is directly matched to the antenna with a single matching section, and the gain is shown to change over frequency. For this particular nonoptimal match, the gain is lower than for the 50-Ω case, and is used only as an illustration of the possible different outcomes. For these simulations, a method-of-moments analysis (Ansoft Nexxim) is performed with a two-port S-parameter (s2p) transistor “black box” for a specific bias point. A bias line is added to the field simulation through the RF null of the patch antenna, as shown in Figure 32-5a. There is no observable change in the active antenna gain, but the radiation pattern looses its symmetry in the H plane due to the addition of the bias line. The amount of pattern distortion depends on the exact position and thickness of the bias line. The E-plane pattern symmetry is affected by the transistor output port, and the polarization is slightly degraded. This type of analysis gives far-field active antenna parameters, such as gain (defined in Figure 32-5 consistent with the IEEE definition), polarization, radiation pattern, as well as near-field quantities such as current distribution. Because the gain definition only takes into account RF power quantities, the simulated gain for an active antenna is larger than the passive antenna gain as the input DC power is not included. This type of analysis is more computationally intensive than circuit-based analysis. A reasonable design procedure starts from an EM simulation of the antenna which gives a frequency-dependent impedance for inclusion into a circuit simulation. After the circuit is optimized using linear or nonlinear analysis, an equivalent multiport can be included in a final field simulation in order to validate potential effects of the circuit on radiation pattern.

Active Antenna Measurements When characterizing the performance of active antennas, it is important to take into account the active circuit properties. An example active antenna measurement for the case of a receiving amplifier antenna of the type given in Figure 32-1c (bottom left) is given here. The only quantities that are accessible are the received RF power at the output of the LNA and the DC power input to the LNA. The received RF power is a function of frequency, magnitude, incidence angle, and polarization of the input wave. For an incidence angle (q,f) and a power PT transmitted from an antenna in the far field at a distance r away, the power received at the antenna is PRF,out (θ , φ ) =

PT 2 G (θ , φ ) Aeff (θ , φ ) 1 − Γ in  Gamp   4π r 2 T

where generally the transmitted power PT and transmitter (test) antenna gain GT are known, while the effective area of the active antenna, match to the amplifier, and gain of the amplifier are not known exactly. By measuring the RF output power over the full sphere for a given polarization at a given frequency, the quantity Aeff (θ , φ ) 1− | Γ in |2  Gamp is determined. Assuming that for the given value of amplifier gain, the antenna is matched to the input of the amplifier, the above can be expressed as 2

PRF,out (θ , φ ) =

PT λ 2 Gant (θ , φ )  λ   PT GT (θ , φ ) ⋅ GAA (θ , φ ) G ( , ) Gamp =  θ φ T 4π 4π r 2  4π r 



32-9

where GAA is the gain of the active antenna, which will give Aeff > Ageom, or an aperture efficiency greater than 100 percent. In this definition, the DC power to the amplifier is not taken into account, and the efficiency of the antenna, which includes aperture and amplifier efficiency, can be found from

ηAA = ηant ηamp =

AAA PRFout Ageom PDC

where AAA = GAA λ 2 / ( 4π ) is the effective area of the active antenna.

32.2 FREQUENCY-AGILE ACTIVE ANTENNAS The feed impedance and radiation pattern of a frequency-agile active antenna is tuned within a frequency range or in multiple bands. If the electrical properties of the antenna are adjusted in a continuous manner, the antenna is said to be tunable, and if the variations are in discrete steps, the antenna is said to be reconfigurable. Varactor diodes, photo diodes, p-i-n diodes, MEMS switches, and ferrites are examples of active devices that can be integrated with an antenna in such a way that the antenna is impedance matched at different frequencies for different states of the active device. When the impedance changes, it is usually desired that the radiation pattern, polarization, isolation, efficiency, and gain remain the same. However, integration of devices that necessarily have insertion loss results in deterioration in the antenna efficiency.28 This is what makes the design of frequency-agile active antennas challenging. Note that reconfigurability can also be applied to the radiation pattern or polarization of an antenna, as overviewed in more detail in Chapter 31. Design and Characterization of Frequency-Agile Antennas The design of a continuously tunable (e.g. Al-Charchafchi and Frances29) or multiplefrequency-band (e.g., Ollikainen et al30) antenna depends on the desired bandwidth for return loss, polarization, radiation pattern, and gain (efficiency). For example, if switching between two or more bands is required to cover an up- and downlink of a half-duplex communication link,30 in addition to requiring the same return loss at the E-GSM band frequencies of 880–915 MHz and 925–960 MHz, the design goals are minimization of losses in the tuning circuit, similar radiation efficiencies, and similar patterns in the two bands. A sketch of the antenna from Ollikainen30 is shown in Figure 32-6a, while the measured dual-band return loss is shown in Figure 32-6b. The antenna element design is based on bandwidth, polarization, available real estate, and preferred circuit architecture. For continuously tunable antennas, a broadband antenna element is desired. For example, a tapered-slot Vivaldi antenna, which can typically achieve octave bandwidth, has been shown to have a 14% varactor-tuned transmit bandwidth,31 defined as a 0.8-dB variation in output power. A second-resonant slot antenna with a possible 15–20% bandwidth was tuned over a 10% bandwidth in Forman and Popovic,4 as shown in Figure 32-7. Other examples include microstrip antennas and folded-slot antennas, dipoles, etc. For frequency reconfigurability, the antenna itself does not need to be broadband. For example, resonant slots32 and dipoles33 have been demonstrated with four-band and three-band operation, respectively. Envisioned applications are frequencyhopped spread spectrum systems where a reconfigurable narrowband antenna follows a pseudo-random pattern of the frequency-hopped modulation. Fractal antennas integrated with MEMS switches have been demonstrated as well.34



CHAPTER THIRTY-TWO

(a)

(b)

FIGURE 32-6 (a) Schematic of a patch antenna with integrated pin-diode switch and (b) response in the two E-GSM bands. The radiation efficiency was comparable in the two bands, and the radiation patterns remained the same (J. Ollikainen et al30 © IEEE 2002).

The type of active element and circuit architecture are dictated by frequency of operation, and antenna feed. The most commonly used active solid-state devices in frequency-agile antennas are varactor and p-i-n diodes, for continuous and switched operation, respectively. There have also been demonstrations with photo diodes used instead of a varactor, such that the impedance changes as the optical power is varied. In addition, MEMS switches have been proposed as low-loss switching elements. Another possibility is the use of a material with DC bias tuned properties, such as in Castro-Vilaro and Rodriguez-Solis35 where a thin-film ferroelectric material, barium strontium titanate (BST), is used to produce a tunable folded-slot antenna. The DC biasing voltage varies the thin-film dielectric constant, causing a resonant frequency shift. The permittivity of the ferroelectric material can change from 400 to 1200, depending on biasing conditions. In this range the resonant frequency changes from 33.54 GHz to 31.35 GHz, giving an 11% frequency bandwidth with VSWR < 2. Characterization of frequency-agile antennas should include measurement of return loss, radiation pattern, polarization properties, and efficiency. For example, as can be seen

8.5 0

RF

8

1.55 1.54 1.53

7.5 frequency

5

−5

LO IF

−10 Vr = 0V Vr = −10V

−15

Vr = −20V

1.51

VSWR 7 0

1.52

Normalized power [dB]

1.56

VSWR

Resonant frequency [GHz]

1.57

10 15 20 Reverse Varactor voltage [V]

(a)

25

1.5 30

Vr = −30V −20

−50

0 Angle [°]

50

(b)

FIGURE 32-7 Second-resonant slot antenna tuned over a 20% bandwidth with a varactor-loaded microstrip feedline (after M. Forman and Z. Popovic4 © IEEE 1997). Measured tuning range and corresponding VSWR (a). The measured radiation patterns change as the bias (frequency) is varied (b).



ACTIVE ANTENNAS

100 90 Tuner

80 ZS

70

Amp Z, SWR

ZL Z’, SWR’

ηD [%]

60 50 40 30 20 10 0 0

1

2 3 4 5 6 Output Block ∆ Insertion Loss [dB]

7

8

FIGURE 32-8 An amplifier matched to a tuned antenna with a realistic lossy tuner circuit. The drain efficiency is plotted as a function of the insertion loss in the output network for amplifiers with efficiencies from 20 to 90%. The solid black line indicates how the hD of a 10-GHz class-E PA with 65% hD would decrease as loss in the output network increases.

in Figure 32-7b, the radiation pattern of the antenna changes over the tuning bandwidth. In this case, the varactor diode cathode is DC- and RF-shorted to the ground pane, and an extra l /4 section is added at the end of the feedline to preserve the RF open. By varying the reverse voltage of the varactor diode from 0 to 30 V, the antenna can be tuned from 7.3 to 8.1 GHz with a VSWR below 1.57, and the peak directivity varies by 2.5 dB over the tuning range. The loss and complexity added by the tuning circuit may not be justifiable. As an example, consider a transmitting active antenna where the goal of the tuning is to provide an impedance match to the amplifier output at several frequencies in a 50-Ω system (see Figure 32-8). Power is lost if either the amplifier is mismatched or there are losses in the circuit/antenna. Assuming a fairly extreme case of a high-efficiency amplifier with 65% efficiency, the question is how much tuning-circuit loss can be tolerated to justify the added complexity. Figure 32-8 shows the calculated efficiency drop as a function of tuner loss. This means that for this high-efficiency case, a tuner/antenna insertion loss of at most 0.3 dB can be tolerated. If the loss is higher, it is more reasonable to tolerate mismatch at the output of the amplifier.

32.3 OSCILLATOR ANTENNAS AND ARRAYS Antennas integrated with oscillators for power combining and sensors/receivers have been overviewed by a number of authors.36 When an antenna is closely integrated with an oscillator, there is no direct access to the antenna feed or the oscillator output. The design of the active antenna requires both circuit and field modeling. The antenna is part of the feedback loop in an oscillator antenna. It is useful to start the design by considering the antenna to be a simple resonant circuit, as shown in Figure 32-9c,27 although this is a



CHAPTER THIRTY-TWO

Re Z Full Wave

200 100 0 −100 −200

Im Z Full Wave

300

Re Z Circuit Model

Impedance [Ω]

Impedance [Ω]

300

Im Z Circuit Model

200 100 0 −100

5

10

15 20 25 Frequency [GHz]

30

−200

35

5

10

(a)

Antenna load ZA(f)

Microstrip lines

20

0 −10 −20 −30 −40 −50 −60

(c)

35

HB-Full Wave HB-Circuit Model

10

Output Spectrum [dB]

jX

30

(b)

Antenna narrow band model

−R


5

10


30

35

(d)

FIGURE 32-9 Example of simulated impedance of a patch antenna over a broad frequency range: (a) real and (b) imaginary part. (c) Simple resonant circuit model of oscillator antenna around first resonance and more realistic model of transistor oscillator antenna. (d) Harmonic balance analysis for the two antenna models, showing the importance of proper broadband modeling.

simplification valid only over a narrow bandwidth. The oscillator circuit is represented as a negative resistance with some equivalent reactance, with bias circuitry omitted. As in standard one-port oscillator models, the negative resistance needs to exceed the radiation resistance of the antenna. In this simple model, the effect of antenna behavior over a broader frequency range is not modeled accurately. Thus, harmonic levels cannot be predicted. An example of the real and imaginary parts of the broadband antenna impedance for the case of an air-patch linearly polarized antenna is shown in Figure 32-9a and b. Both full-wave and circuit models for a patch are included in the plot, showing that the first resonance is correctly modeled by the simple resonant circuit, while the higher harmonics are not. If an active oscillator antenna design is to be correct at the harmonics, the broadband impedance needs to be used in the harmonic-balance nonlinear simulations. A simplified oscillator-antenna circuit is shown in Figure 32-9c. Figure 32-9d compares the harmonic levels obtained with a harmonic balance nonlinear simulation for two circuits which differ only in the way the antenna is included: in one case as the resonant-circuit model and in the other as a set of S-parameters obtained from the full-wave simulation. Power levels at all the harmonics differ considerably in the two cases, including the power at the fundamental. This simple example illustrates the importance of correctly modeling the antenna in circuit design.



ACTIVE ANTENNAS

FETs

(b)

FETs

(a)

(c)

FIGURE 32-10 Examples of oscillator antennas: (a) a push-pull oscillator patch; (b) a Vivaldi slot VCO antenna with varactor-tuned Gunn-diode circuit; and (c) a microstrip annular ring is both the radiating element and microstrip feedback circuit for a class-E amplifier.

For a narrowband transmitter, a narrowband antenna is chosen. If radiation in the half plane is desired, a patch antenna is easily integrated with the oscillator circuit, as in the example in Figure 32-10a.14 The patch antenna has several other properties that can be advantageous for active antennas: the voltage null in the middle is a natural DC bias line, and the phasing of the voltage and current on the two halves of the patch nonradiating edge allows for multitransistor circuit designs, such as in Figure 32-10a where a push-pull FET oscillator loads a 6-GHz patch antenna giving an output ERP of around 18 dBm. Several such elements have been successfully combined in small arrays with increased ERPs. For a broadband medium-power tunable linearly polarized transmit antenna element, the Gunn-diode varactor-tuned Vivaldi antenna is chosen in Navarro et al31 (see Figure 32-10b). For millimeter-wave (155 and 252 GHz) low-power transmitters, slot antennas with high cutoff frequency transistors (HFETs) are fabricated monolithically. If unidirectional radiation is required, the slot can be backed by a ground plane, polarizer, or substrate lens. Fundamental oscillations from 115-GHz and 215-GHz monolithic oscillator antennas are reported with 10uW and 1uW estimated output powers from the oscillators.37 Classical design of oscillators based on an amplifier with appropriate feedback can also be applied to oscillator antennas. In a high-efficiency 10-GHz integrated transmitting antenna,38 an annular ring is used both as the radiating element and microstrip feedback circuit for a class-E amplifier. A maximum conversion efficiency of the DC power consumption to radiated copolarized power is 55% at 10 GHz with maximum ERP of 23.6 dBm and total radiated power of 15.5 dBm. In this case, the size of the active antenna is minimal, since the antenna is used directly as the feedback circuit, with an amplifier embedded within the antenna footprint. The annular microstrip ring antenna is operated in its TM12 mode. The wider bandwidth of the TM12 mode compared to the TM11 fundamental mode is



CHAPTER THIRTY-TWO

useful in the active antenna design, since it enables some tolerances on the active circuit. Nonlinear models of active devices have difficulty predicting behavior in switched mode of operation, which is compensated for in design by increased impedance bandwidth. Both two-port and three-port active devices have been used in oscillator antennas: Gunn diodes,31 IMPATTs,39 various types of transistors, tunnel diodes,36 and Josephson junctions.40 Transistor transmitting antennas currently dominate because of the relative ease of integration, the fact that frequency and power are dictated mostly by the circuit/antenna, and availability of devices that cover a broad range of frequencies and powers. Application Example for a Single Oscillator Element and Small Array Single oscillating antennas have been developed for compact low-cost RF sensors, while oscillator arrays have been applied to spatial power combining and phase-shifterless beamsteering. Low-cost, low-power manufacturable automotive crash sensors and RFID tag systems41 have been developed at 24 GHz for use in the ISM unlicensed band. An example of a sensor antenna is shown in Figure 32-11a. The active device at 24 GHz is an HEMT or HBT, and the circuit is implemented in microstrip on an Alumina substrate (er = 9.8) with gold metalization and is packaged in a standard package. This active antenna is designed as a three-port oscillator, where the gate microstrip resonator is the radiating element.42 In low-power sensor applications, the parameters of interest are the precise oscillation frequency, frequency stability, and phase noise. We first consider design of the oscillator antenna such that for every device, irrespective of device variations, operation within the 200-MHz band around 24 GHz is ensured. The first step is to accurately model the embedding impedances at the active device terminals with full-wave analysis. This is followed by either a linearized or nonlinear circuit analysis. In the former, the oscillator is considered as a combination of an active multiport and a passive multiport which is obtained from field simulations. The oscillation condition can be expressed as | det (ST ⋅ S p − I ) | > 0 and

Arg det (ST ⋅ S p − I ) = 0

where I is the identity matrix. This condition is plotted as a function of frequency in the complex plane and crossing of the positive real x-axis at the desired oscillation frequency (24 GHz in this case) indicates an oscillation. It is good practice to measure the S-parameters of the device for varying input power levels. The major effect under large signal conditions is the change in |S21|, and a reduction in |S21| by a factor between 0.4 and 0.6 is usually appropriate. This type of analysis usually predicts the oscillation frequency within 1% to 5%

(a)

(b)

(c)

FIGURE 32-11 (a) Photograph of the 24-GHz packaged oscillator antenna using an HEMT device; (b) photograph of four-element active array, and (c) measured spectrum of array (black) compared to freerunning spectrum of single element (gray)



32-15

of the measured value, but does not give any prediction of power or harmonic levels. To obtain the exact oscillation frequency, a varactor diode weakly inductively coupled to the gate resonator can provide 3% tuning. An important parameter for sensors that rely on small frequency modulations (e.g., Doppler shifts) is the phase noise. The phase noise can be improved by increasing the loaded Q factor of the resonator antenna, by injection locking, or by using mutual locking of a number of active antennas in a small array, and is inversely proportional to the square of the loaded Q factor of the resonator-radiator. When a longer gate resonator-radiator with doubled unloaded Q factor is inserted in the same active antenna, a decrease of 5 dB, compared to the expected 6 dB, in phase noise is measured. The different resonator will, however, affect the radiation pattern and polarization. Injection of a low-phase noise external source can reduce the phase noise further (down to –80 dBc/Hz at 100-kHz offset in this case) but this approach is not practical for low-cost small sensors because it requires a secondary source. In an array of self-locked oscillators, the signal combines coherently and the noise incoherently, resulting in improved phase noise. Mutual locking of a number of nonlinear oscillators cannot be simulated with commercial software tools. However, coupling in the passive part of the circuit can be simulated and used as a useful starting point for mutually locked array design. The coupling for a four-element array shown in Figure 32-11b can be described by a symmetric 4×4 admittance matrix. The oscillation condition can then be calculated by −Yosc,i (ω r , Ai ) =

N

∑Yij (ω r )

j =1

where Yosc,i is the frequency and amplitude-dependent transistor admittance, and Yij are the elements of the coupling matrix. Based on simulation, a four-element array was implemented, with phase noise compared to a single oscillator antenna, as shown in Figure 32-11c. The ERP increased by roughly 10 dB, while the phase noise decreased by 7 dB. Both numbers differ from the theoretical maximum when four identical elements are locked coherently without change in operating conditions. In the array, however, due to the coupling, the individual oscillators operate under different conditions than when alone, and their biases need to be adjusted for locking at 24 GHz. The expected 12-dB increase in ERP (6 dB in power and 6 dB in directivity) and 6-dB decrease in phase noise can only be design guidelines. Oscillator Antenna Arrays In addition to phase noise reduction, arrays of oscillator antennas have been implemented for power combining of low-power oscillators. In arrays with standard half-wavelength spacing, the coupling between antenna elements may not be sufficient to enable oscillator synchronization through injection locking, and additional coupling needs to be provided, e.g. through a circuit. This approach lends itself to beam-steering through frequency control. Alternatively, grid arrays with a period on the order of l /10 are inherently tightly coupled and provide the highest power densities in power-combining active apertures.27 Quantities that are directly measurable are ERP, radiation pattern, polarization properties, spectral content, and input DC power. From these quantities, the radiated RF power and conversion efficiency can be estimated. Most standard arrays with half-wavelength spacing between elements have been limited in the number of elements to 16, due to the difficulty in maintaining synchronized oscillations. In tightly coupled grid oscillators, however, the number of elements that can be coherently combined is in the hundreds. The first such planar grid oscillator with 100 MESFET elements at about l /10 spacing at 5 GHz is shown in Figure 32-12a,8 where l is the freespace wavelength. The vertical metal lines are connected to the drain and gate leads, and the



CHAPTER THIRTY-TWO

(a)

(b)

FIGURE 32-12 (a) 100-MESFET grid oscillator with measured ERP = 22W (after Z. Popovic et al8 © IEEE 1991).The grid is linearly polarized with a cross-polarization level of –30 dB at broadside. An equivalent circuit convenient for grid-oscillator analysis is shown. (b) Overmoded waveguide resonator coupled to array of waveguides combines a 2D array of Gunn-diode oscillators at millimeter-wave frequencies (after Bae et al44 © IEEE 1998).

horizontal metal lines to the source leads and VGS and VDS bias supplies. The grid is a square around l on the side, backed by a flat reflector of the same size. When the transistors are biased, oscillations start from noise. Power is radiated incoherently from the drain terminals, reflected off the mirror, and received by the gate terminals. For appropriate mirror spacing combined with the phase between gate and drain voltages, steady-state locked oscillation is reached through global coupling after several bounces inside the resonator. Analysis of this combiner is based on the assumption that the symmetry of the grid reduces it to a unit cell with electric walls on the top and bottom and magnetic walls on the sides. The passive multiport connected to the transistor terminals can be found using the EMF method8 or commercial software such as Ansoft’s HFSS. An equivalent circuit that describes the physical behavior of the grid is shown in Figure 32-12b, where the vertical leads are represented by inductances, the horizontal leads by a resonant circuit, and the currents on the leads are coupled to the plane wave (a transmission line with a 377-Ω characteristic impedance) through a center-tapped transformer. The mirror is represented by a short circuit. Circuit simulations using linear parameters give a resonant frequency within 1 to 5% of the measured ones. Subsequent grid oscillators have demonstrated operation at millimeter-wave frequencies and with up to 660W ERP and 10W of radiated power from 100 devices at X-band.43 Similar arrays at millimeter-wave frequencies where the devices are embedded in one of the mirrors of the resonator are described in, e.g., Bae et al,44 where an overmoded-waveguide resonator with an array of TE10-mode waveguides containing Gunn diodes achieved highly efficient spatial power combining at V-band (61.4 GHz). An output power of 1.5W CW was measured with an estimated efficiency of 83% with a 3×3 Gunn-diode array, Figure 32-12c. Large-scale power combining in a grid oscillator is achieved after radiation in free space. For all elements in phase, the power is combined in the broadside direction and the grid radiates like a continuous active aperture producing a radiation pattern close to (sin x /x)2. For applications where beam-steering is of interest, antenna arrays with half-wave spacing Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.


ACTIVE ANTENNAS 0 theory at broadside measurements at broadside -5 at 30º theory measurements at 30º

Amplifier

Antenna

Slave VCO

Master VCO f1

Slave VCO f2

Slave VCO f3

Slave VCO f4

Varactor tuning ports

f5

Relative power [dB]

−10 −15 −20 −25

(a)

−50

0 Angle [º]

50

(b)

FIGURE 32-13 (a) Principle of beam-steering by frequency detuning. (b) Example measured radiation patterns as the control voltages of the edge VCOs are varied to produce a frequency difference across the array, corresponding to a phase gradient (after R. A. York and T. Itoh45 © IEEE 1998).

are more appropriate. A very interesting and elegant approach is described in detail in Popovic and York27 and is illustrated in Figure 32-13. The oscillator antenna elements at the edge of the linear array are detuned in frequency, while the middle elements lock with progressive phase, corresponding to a differential frequency between the end elements. In an X-band prototype ten-element power-combining beam-scanning array of coupled voltagecontrolled oscillators (VCOs)45 and varactor-tuned patch antennas, a measured ERP of 10.5W at 8.4 GHz is achieved with a maximum scan angle of 30 degrees, corresponding to 500 degrees of phase shift across the array. A theoretical investigation of the phase noise in mutually synchronized oscillator systems indicates the necessity of using external injection locking. The steady-state phases in coupled oscillator arrays are dependent not only on the free-running frequency distribution, but also on the coupling-phase angle, as shown in the insightful theory by York.45 Beam-steering arrays with PLLs and multipliers have also been demonstrated. However, extensions to 2D arrays seem unlikely, because it is difficult to independently control and design coupling in two dimensions.

32.4 AMPLIFIER ANTENNAS AND ARRAYS Amplifier antennas fall into one of the three categories described in Figure 32-1c: receiving antenna integrated with LNA, transmitter PA with antenna load, and amplifier with antenna input and output. While the first two cases are useful both as single elements and arrays, the latter is used mainly in arrays for spatial power combining. The antenna gain is increased compared to a passive antenna at the expense of input DC and RF power. The normalized antenna radiation pattern should remain the same, although addition of circuit and bias can change the antenna far-field properties, especially polarization and sidelobes. The bandwidth of the antenna tends to increase, since effectively loss is added and the Q factor is reduced compared to the passive antenna since the amplifier efficiency is always lower than 100 percent. Single Amplifier Antenna Elements: Receiving and Transmitting In a receiver, any loss in the antenna and feedline connected to the LNA contributes to the reduction of noise figure as Ftot = Ffeed + L ( F1 − 1) , where F1 is the noise figure of the first amplifier in the receiver chain and L is the loss of the antenna feed connected to the input of the LNA. Usually, the optimal noise match at the input of the amplifier is not 50 Ω, Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.


CHAPTER THIRTY-TWO 1 Noise minimum 5 l/λ = 0.285 10

Period dipole HBT amplifier

5

1

2

5 10

−5

l/λ = 0.25 −2

−0.5

−10

l/λ = 0.145

Power gain, noise figure / dB

10 2

Dielectric substrate

6 Noise figure Theory Measurement

4 2 0 5.4

−1

Power gain

8

5.6

5.8

6

6.2

6.4

Frequency / GHz

(a)

(b)

FIGURE 32-14 (a) Input reflection coefficient of printed monopole antenna as a function of length plotted along with the optimal noise match for a SiGe HBT device. (b) Measured and calculated noise figure show increased bandwidth for the integrated amplifier antenna. The power gain is relative to a passive antenna (W. Duerr et al46 © IEEE 1997).

and is also not simultaneously an optimal input reflection coefficient match. In Duerr et al,46 the impedance of a monopole antenna for different dipole lengths was examined as a direct noise match for a SiGe HBT amplifier in the 5.8-GHz WLAN band. Figure 32-14 shows the relevant antenna and device impedances, as well as a plot of the noise figure. In Radulovic et al,47 a patch antenna is designed to present complex optimal noise figure impedance to a HEMT device at 13 GHz. The patch is fed with an indented microstrip line, which allows for design of complex input impedance by varying two geometric parameters for a given substrate. It is shown that when the patch antenna is designed to provide an optimal noise reflection coefficient of 0.28150°, the radiation pattern of the antenna remains the same, but the gain increases from 7.7 dBi for the passive antenna to 15.1 dBi for the active amplifier antenna. It is interesting to note that there have been demonstrations of broadband electrically short half-loop antenna amplifiers (50–500 MHz),48 which are cryogenically cooled and integrated with a SQUID array and buffer amplifier in a portable package. The addition of the inductive SQUID array and buffer amplifier provided the broadband matching for the small inductive antenna. It is not straightforward to measure the noise figure of the amplifier in an active receiving antenna. If the antenna temperature is TA, the noise figure of the amplifier can be found from F = 1+

T N − A G AA 290

where N is the noise spectral power density measured at the output of the LNA referred to input at room temperature of 290K (–174 dBm/Hz), and GAA is the transducer gain of the active antenna, which in turn requires measurements of a reference passive antenna.49 The ambient temperature (TA) needs to be well controlled. In a transmitter, the antenna is the output load to the amplifier. For a power amplifier, the optimal transistor output impedance is usually not 50 Ω, but is determined by the mode in which the transistor operates. For example, in the case of a class-A saturated power amplifier (PA), the load line and device and packaging parasitics determine the optimal output power impedance. In the case of a high-efficiency switched-mode PA in class-E mode of operation, a specific output impedance forces the time-domain current and voltage waveform product, i(t)v(t), to be zero during most of the period of the carrier frequency, thus minimizing the loss in the transistor. For soft switching, it can be shown that this complex impedance is, e.g., Weiss and Popovic,26 ZE =

0.28015 j 49.052° e ω Cout



ACTIVE ANTENNAS

80

70

60

50

40

30

20

10 9

(a)

(b)

Drain Eff. [%] PAE [%]

9.5

10.5 10 Frequency [GHz]

11

(c)

FIGURE 32-15 (a) Circuit side and antenna side of active class-E amplifier antenna at 10 GHz. (b) Simulated and measured nonresonant slot antenna impedance, along with optimal class-E impedances for transistor at fundamental (10 GHz) and second harmonic (20 GHz). (c) Measured efficiency of active antenna as a function of frequency.

where ω = 2π f is the fundamental operating angular frequency, and Cout is the output capacitance of the transistor. The higher harmonics are assumed to be terminated with high impedances (ideally open circuits). In Weiss and Popovic,26 a 10-GHz active class-E antenna is demonstrated with a MESFET device and a slot antenna (see Figure 32-15a). The output capacitance of the transistor is 0.107pF, resulting in an optimal output impedance of 41.7Ω, 49.05° with a resulting reflection coefficient with respect to 50 Ω equal to −6.67dB104°. Thus, a nonresonant antenna with this impedance at the fundamental 10-GHz frequency and a high impedance at the higher harmonic frequencies can be directly connected to the transistor and thus reduce the footprint of the active antenna. A slot antenna is designed using a full-wave simulator, with impedance at the fundamental and second harmonic as shown in Figure 32-15b. A passive slot was also fabricated and the measured reflection coefficient at 10 GHz, with a TRL calibration, was measured to be −6.8dB106°. The efficiency and power of the antenna are estimated from measured copolar and cross-polar radiation patterns. The radiated power of around 100 mW with 75% drain efficiency was achieved. This is higher or comparable to an amplifier designed with the same device in a microstrip circuit, since some circuit losses are avoided. Due to the direct circuit-antenna integration, the total active antenna footprint is 0.4 λ 2. Receiving Amplifier Antenna Arrays Figure 32-16 shows an active receiving antenna array where antennas are directly connected to LNAs in each element with no lossy isolators or circulators. Such antenna arrays are used in radioastronomy, either as large-area arrays or reflector feeds. The transistor amplifier produces noise waves at both of its ports, and the noise at the input will radiate, potentially coupling into the neighboring antenna elements and amplified by their LNAs thus degrading the system noise performance.50 As the phases of the phase shifters are varied and the beam is scanned, the amount of coupling between the antennas varies and this affects the reflection coefficient at the input of each of the LNAs, as well as the amount of noise coupled to neighboring elements. Referring to Figure 32-16a, it can be shown that the coupled noise at the output of all the phase shifters is ctot (θ , φ ) = c2 + Gc1Si (θ , φ )



CHAPTER THIRTY-TWO

Excellcs EPA060B-70

NEC NE325S01 FET

3

Gc1/c2 (dB)

3

2.5

2.5

2

2

1.5

1.5

1

1

0.5

0.5

0

3.5

Noise Figure (dB)

3.5

5

(a)

10 Frequency [GHz]

15

0

(b)

FIGURE 32-16 (a) Receiving LNA active array with noise coupling. (b) Plot of transistor characteristics relevant to noise coupling. The best noise figure of a transistor does not correspond to a minimal noise coupling coefficient for a given antenna array.

where c2,c1 are the noise waves at the input and output of all (identical) LNAs, G is the gain of each LNA, and Si (θ , φ ) is the active scan reflection coefficient of element i for a scan angle (θ , φ ). The noise coupling for a given scan angle can therefore be found from measured or simulated active scan reflection coefficient, amplifier gain, and amplifier input and output noise waves. The latter can be determined using standard techniques from minimal noise figure Fmin , optimal reflection coefficient Γ opt , and noise resistance Rn, parameters given in specification sheets for most transistors. If a blind (noisy) scan angle appears for a certain element type or element coupling, the noise coupling can be modified by changing the antenna type, as well as the transistor. As an example, consider a 7×7 half-wave dipole array above a ground plane with a half-wave spacing. The 49×49 full scattering matrix of the array can be found using a full-wave simulator, and the scan reflection coefficient can be found for each element. The noise coupling ratio, defined as the ratio of the uncoupled noise to the coupled noise, can be written as NCR = (Gc1Si ) / c2 and is a function of antenna array properties and amplifier properties. The amplifier quantity Gc1 / c2 is plotted for three different transistors in Figure 32-16b, while the noise figures for these amplifiers from their data sheets are shown in Figure 32-16c. A lower noise figure can un-intuitively result in higher noise coupling, reflecting the fact that minimal noise figure implies low noise power at the output, but not necessarily at the input of an LNA. An interesting figure of merit is the amount of coupling that produces 50 percent more noise power at the array output at certain scan angles, as compared to that expected without noise coupling. To find the minimal amount of inter-element antenna coupling that would give the noise ratio of –3 dB, we set 10 log

Gc1Si (θ , φ ) c2

2

2

= −3 dB

and solve for the allowable active scan coefficient for each array element, given the amplifier properties. For reasonable LNA gain, the practical scan angle for low coupled noise will be limited: low noise coupling requires low values of active scan coefficient, i.e., larger element spacing; but larger element spacing limits the scan angle due to sidelobe appearance. The design of a low-noise scanning array therefore involves choice of transistor, design of LNA and match to antennas, choice of antenna element, and design of array for minimal coupling for desired scan range.



ACTIVE ANTENNAS

Transmitting Amplifier Antenna Arrays A power amplifier integrated in each antenna element provides distributed amplification, resulting in higher achievable efficiency and improved degradation properties for a given ERP. Spatial power combining upon radiation results in an ERP that is proportional to N 2, where N is the number of amplifier antenna elements. Spatially fed combiners are fed in two ways: using a guided-wave feed network and a spatial feed. For a given combiner network loss per stage, there is a minimal number of elements N for which spatial power combining has a larger power-combining efficiency.27 An example of a circuit-fed array is shown in Figure 32-17a where 16 high-efficiency 10-GHz amplifiers are spatially combined using stacked patch antennas spaced a half wavelength apart.51 In this case, the antennas and feeds are designed to present a specific complex impedance to the output of the transistors.26 A corporate four-level Wilkinson divider feed network is used at the input of the array with a measured loss of 1.4 dB. The output powers are spatially combined through an array of multilayer patch antennas with 2:1 VSWR bandwidth of 11.6%. An average drain efficiency of 70% at 162W EIRP, or approximately 1.5W of transmitted power, is measured for an array of 16 amplifiers at 10.2 GHz. The estimated power-combining efficiency is 75%. Spatially fed, spatially combined arrays have also been demonstrated, e.g. using slot antennas at Ka-band as shown in Figure 32-18a. Input and output antennas are anti-resonant cross-polarized slots fabricated on alumina52 or aluminum-nitride53 substrates with MMIC amplifiers providing power gain between the input and output slots. Polarizers ensure unidirectional slot radiation. Several such arrays with watt-level power at Ka-band have been demonstrated, e.g. a 6×6 array gave an output power of 4W and power gain of 6 dB at 29 GHz, where a liquid-cooling test fixture removed excess heat. Aluminum nitride proved to be an adequate heat conductor for a 1W 6×6 33-GHz array. In this case, all input slot antennas are spatially fed with a dielectric-loaded (“hard”) horn antenna in the near field of the array. The amplitude and phase profile of the horn loaded with the array is not uniform and degrades the power-combining efficiency. Figure 32-18b shows the measured H-plane copolarized radiation patterns of two arrays that differed only in bias line configurations: the bias lines of Array B have extra air bridges and capacitors that short slot-mode radiating RF currents on the bias lines, which results in an improved pattern. A compact approach for combining a smaller number of elements was done in N-S. Cheng et al,54,55 where tapered-slot antennas connected to both ends of 50-Ω monolithic amplifier chips are inserted in a waveguide, achieving broadband combining from 8 to 11 GHz with over 120W of output power. In the array shown in Figure 32-19a, a total of N = 24 amplifiers is combined, with six cards inserted in the waveguide, containing four MMIC amplifiers each. This approach provides large bandwidth, good heat-sinking, and space

(a)

(b)

(c)

FIGURE 32-17 (a) Schematic of the four-element active subarray showing multiple layers. (b) Radiation patterns of the passive and active 16-element antenna arrays (S. Pajic and Z. Popovic51 © IEEE 2003).



CHAPTER THIRTY-TWO

(a)

(b)

FIGURE 32-18 (a) Layout of a Ka-band slot antenna amplifier array with prematched 50-Ω MMIC amplifiers. Polarizers are used externally to the array for unidirectional slot radiation. (b) Measured H-plane copolarized radiation patterns for two arrays that differ only in bias-line implementation (T. Marshall et al53 © IEEE 1999).

in the z-direction to add gain, and has graceful degradation as shown in Figure 32-19b. However, the architecture accommodates a limited number of cards, and the waveguide mode has a nonuniform input field profile, thus the edge amplifiers contribute less to the total output power. Another approach for large-scale combining is a grid amplifier.56 The radiating portion of a grid amplifier is similar to that of a grid oscillator (see Figure 32-12), except in this case, the input and output radiating elements are orthogonally polarized to reduce feedback between input and output waves. A differential transistor pair provides amplification in each unit cell. Figure 32-20 shows a photograph of a portion of a monolithic 37-GHz grid amplifier, with measured 5W of output power.57 Grid amplifiers have been realized as both transmission and reflection type. In DeLisio et al,58 a 31-GHz two-stage transmission

(a)

(b)

FIGURE 32-19 (a) Architecture of a broadband finline waveguide amplifier array, delivering over 120W over the entire waveguide X-band. (b) Measured gain and output power vs. number of operating amplifiers. The inset shows one card with two tapered slot antennas connected with MMIC amplifiers (N-S. Chang et al55 © IEEE 1992).



ACTIVE ANTENNAS

(a)

(b)

FIGURE 32-20 (a) Measured gain and output power of a 37-GHz monolithically integrated grid amplifier. (b) Photograph of a portion of the 256-element 1-cm-square grid-antenna amplifier chip. The input electric field is vertically polarized, and the amplified wave is radiated in the horizontal polarization. The bias lines are vertical. (Courtesy Prof. D. Rutledge, Caltech)

grid amplifier module with over 10W output power and 12-dB gain is described. An 82-GHz reflection grid amplifier with 5.5-dB gain, 400-MHz 3-dB bandwidth, and 110-mW saturated output power is presented in Cheung.59 Commercially available grid amplifiers are not plane-wave fed from the far field, but placed in an overmoded waveguide with excellent feed efficiency, in which case they arguably are no longer active antennas. Practical issues in active antenna amplifier arrays include feeding, heat management, impedance matching, and stability. Several types of spatial feeds have been demonstrated: plane-wave feeding is inefficient and practical only for characterization; Gaussian-beam feeds are low-loss but large58; dielectric-loaded horn antennas do not produce uniform amplitude and phase illumination, thus reducing power-combining efficiency60; and discrete lens antenna arrays, e.g. Hollung et al,12 are large in volume but provide additional functionality. An especially difficult problem is stability in active antennas and arrays—electromagnetic coupling is present due to radiation. Bias-line oscillations are easily excited in amplifiers where one of the loads radiates, and this can limit the gain of the active antenna. Shunt capacitors of different values placed along the bias lines to short the bias-line oscillation currents to ground can eliminate instabilities.

32.5 FREQUENCY-CONVERTING ANTENNAS The two extreme cases of frequency-converting antennas are oscillator antennas (Section 32.2) that convert DC power to power at a single RF frequency and its harmonics, and rectifier antennas that convert incident RF waves into DC power (Section 32.6). In this section, multiplier and mixer active antennas are briefly overviewed. Multiplier Antennas and Arrays Although there have been a number of active antenna frequency multipliers demonstrated in the literature, it is easy to conclude that at microwave frequencies, fundamentalfrequency oscillators will give more power and lower phase noise, which grows at best as



CHAPTER THIRTY-TWO

(a)

(b)

FIGURE 32-21 (a) A millimeter-wave active antenna quasi-optical grid multiplier; (b) measured output power at 1 THz with and without input and output tuning slabs (A. Moussessian et al62 © IEEE 1998)

20 log N, where N is the multiplication factor. At millimeter-wave frequencies, however, transistors are either not available or have low output powers, and multiplication becomes more relevant. Frequency doubler and tripler antenna arrays for millimeter-wave generation with Schottky and varactor diodes have the advantage of input power division between a large number of elements, allowing high-power generation. For example, 0.5W was generated at 66 GHz in Jou et al61 in a monolithic 760-diode grid array with around 10% conversion efficiency from a pulsed source. A similar approach was used by Moussessian et al62 where 144 diodes generated 24 mW at 1 THz from a 47W peak pulsed source with an efficiency below 1%. As shown in Figure 32-21, the input plane wave is incident on an array of antennas loaded with a nonlinear harmonic-generating element (diodes). The output wave is filtered so that only the higher harmonic is radiating in transmission. For tripler antennas, the diodes are connected in anti-parallel, eliminating the even harmonics for perfectly matched diode pairs. Mixer Antennas and Arrays Among the many interesting up/down-conversion active antennas, reviewed nicely in Mortazawi et al,36 we choose a few examples as illustrations in this chapter. Self-oscillating mixer antennas with fundamental and subharmonic oscillations have been demonstrated, and an example of a 60-GHz balanced integrated-antenna self-oscillating mixer is shown in Figure 32-22.63 This mixer employs the modal radiation characteristics of a dual-feed planar quasi-Yagi antenna to achieve RF–local oscillator (RF–LO) isolation between closely spaced frequencies. The balanced mixer is symmetric, inherently broadband, and does not need an RF balun. A 30-GHz push-pull circuit is used to generate the second harmonic, and a 30-GHz dielectric resonator is used to stabilize the fundamental oscillation frequency. This allows the possibility of building a balanced, low-cost self-contained antenna integrated receiver with low LO leakage for short-range narrowband communication. Figure 32-23b shows the measured conversion efficiency of this mixer antenna at the sum and difference IF ports. A grid mixer, similar to grid oscillators and amplifiers and attractive for millimeterwave frequencies, is demonstrated in Hacker et al64 at X-band. A planar bow-tie grid periodically loaded with 100 Schottky diodes showed an improvement in dynamic range of 3.5 dB over an equivalent single-diode mixer with the same conversion loss and noise figure. The advantage of a mixer array is that power handling and dynamic range scale as the number of the devices.



ACTIVE ANTENNAS

(a)

(b)

FIGURE 32-22 (a) Layout of a subharmonic mixer active antenna; (b) measured conversion efficiency for sum and difference IF ports as a function of the V-band RF frequency for a 30-GHz push-pull self-oscillating mixer (M. Sironen et al63 © IEEE 2001)

Different nonlinear functions can be performed by active mixer antennas. An example is four-wave mixing, or phase conjugation, as shown in Pobanz and Itoh,65 with the application to saving power in back-scatter mode communication and radar systems. A phase-conjugating active antenna is shown in Figure 32-23a. The circuit is similar to a standard single-balanced mixer with a rat-race hybrid, but with the LO and IF interchanged, so that isolation between RF and IF is accomplished via hybrid balance rather than the usual filter. In this case, the LO frequency is chosen to be twice the RF, and a dual-frequency ring hybrid is implemented. The LO is coupled into the ring through a 12-GHz bandpass filter. The filter is an open circuit at the ring port at 6 GHz, with sum

6 GHz microstrip patch antenna

RF, IF

0

RF, IF

7 GHz lowpass 12 GHz virtual ground

λ/4 at 6 GHz

diodes LO 12 GHz

12 GHz bandpass

(a)

Bistatic RCS [dB rel.]

−5 −10 −15 −20 0º −20º

−25

+45º −30

−50

0 Scatter angle [º]

50

(b)

FIGURE 32-23 (a) Layout of a phase-conjugating four-wave mixing active antenna element; (b) measured radiation patterns of an eight-element array for waves incident from broadside, –20°, and 45° (after Pobanz and Itoh65 © IEEE 1995)



CHAPTER THIRTY-TWO

and difference inputs feeding the mixer diodes at the remaining ports. Low-pass filters (LPFs) are used on the RF/IF ports, which appear as open circuits to the ring at 12 GHz. Since the LO wavelength is half that of the ring design wavelength, only a sum port can be formed to feed the diodes with equal power. By placing the LO port in the longer section of the ring at a 6-GHz quarter-wave from either port 1 or 4, two virtual ground points can be created in the matched ring. The mixer is coupled to an antenna with two orthogonal feeds, allowing this phase-conjugate scattering element to operate for any polarization of the incident wave. Measured radiation patterns for an eight-element array with a 0.8l spacing with different incidence angles of the incoming wave are shown in Figure 32-23b. In the phase-conjugating process, the spacing between the elements is effectively doubled, resulting in grating lobes even for half-wavelength spacing, which is traded off with mutual coupling between elements.

32.6 RECTENNAS An integrated antenna and rectifier is usually referred to as a rectenna, as shown in Figure 32-24a. The incident waves within a certain spectral range are received by the antenna and coupled to the rectifying device (diode in this case), and the LPF ensures that no RF is input to the power management circuit. A controller provides input to the power management circuit, which enables storage of the received energy over time, and delivery of DC power at the level and time when it is needed. Such active antennas have been primarily proposed for wireless power transmission, which dates back to Nikola Tesla’s U.S. Patent No. 685,954 (1901),66 which describes wireless transmission of energy, storage of the energy in a capacitor, and energy management in time. Rectification of microwave signals for supplying DC power by high-power beaming has been researched for several decades, and a good review of earlier work is given in Brown.67 In power beaming, the antennas have well-defined polarization, and high rectification efficiency is enabled by single-frequency high microwave power densities incident on an array of antennas and rectifying circuits. Applications for this type of power transfer have been proposed for helicopter powering, solar-powered satellite-to-ground power transmission, inter-satellite power transmission including utility power satellites, mechanical actuators for space-based telescopes, small DC motor driving, and short-range wireless power transfer.67–73 Linear, dual-linear, and circular polarization of the receiving antennas were used for demonstrations of efficiencies ranging from around 85–90% at lower microwave frequencies to around 60% at X-band and around 40% at Ka-band.

(a)

(b)

FIGURE 32-24 (a) Schematic of a rectenna and associated power management circuit; (b) an example of a dipole rectenna for 5.8-GHz narrowband operation (after J. O. McSpadden72 © IEEE 1998)



32-27

There are several distinct scenarios for wireless powering, which influence the rectenna design: ●

●

●

One or more high-directivity narrowband (single frequency) line-of-sight transmitters with well-known and fixed polarization and well-known power levels. In this approach, used for power beaming, the efficiency of the rectenna can be very high. One or more medium-power semidirectional transmitters that illuminate a range in space, with multipath present. In this case, the incident power density is known approximately, but there are multipath effects that change polarization and spatial distribution of power density. The transmitters can be single-frequency, multiple-frequency, or broadband. An application of this scenario is for powering multiple sensors whose location is not precisely known. Unknown transmitters over a range of frequencies, power levels, generally unpolarized, with varying low-level spatial power densities, such as in energy harvesting or energy scavenging.

The incident power density on the rectenna, S (θ , φ , f , t ), is a function of incident angles, and can vary over the spectrum and in time. The effective area of the antenna, Aeff (θ , φ , f ), will be different at different frequencies, for different incident polarizations and incidence angles. The average RF power over a range of frequencies at any instant in time is given by 1 PRF (t ) = fhigh − flow

fhigh 4 π

∫ ∫

flow

S (θ , φ , f , t ) Aeff (θ , φ , f ) d Ω df

0

The DC power for a single-frequency ( fi ) input RF power is given by PDC ( fi ) = PRF ( fi , t ) ⋅ η( PRF ( fi , t ), ρ, Z DC ) where h is the conversion efficiency, and depends on the impedance match ρ( PRF , f ) between the antenna and the rectifier circuit, as well as the DC load impedance. The reflection coefficient in turn is a nonlinear function of power and frequency. Therefore, to find the efficiency of the rectenna, the following measurements can be performed: ●

●

●

●

Illuminate the rectenna from a known distance and incidence angle with a known frequency and power; thus S (θ , φ , f , t ) is known. For each frequency of interest, determine the geometrical electrical area of the antenna. This is generally larger than the effective area and therefore overestimates the received RF power. Measure the DC power as a function of the DC load impedance (resistance). Calculate the estimated conversion efficiency as PRF /PDC. Since the DC power is measured directly, and the RF power is overestimated, the resulting efficiency will be an underestimate.

This process should be done at each frequency in the range of interest. However, the process is not linear and DC powers obtained in that way cannot be simply added in order to find a multifrequency efficiency.



CHAPTER THIRTY-TWO

Narrowband Rectenna Design For most rectennas and arrays reported to date, the antenna is matched to the diode around one frequency at a well-defined polarization and assuming relatively high incident power levels, e.g. in Brown.74 For example, the rectenna shown in Figure 32-24b is linearly polarized and designed to operate at 5.8 GHz with an incident power of 50mW corresponding to an incident power density of around 3.2 mW/cm2, assuming a dipole effective area of l2/8. In this case, the incident wave carries enough power to turn on the diode, and rectification efficiency can be very high (>80%). A different design is required for low incident power levels. For example, the unlicensed ISM band around 2.4 GHz is an appropriate frequency range for low-power wireless sensor powering. In this scenario, the sensor is mobile and therefore the polarization changes. Thus, the rectenna is designed to be dual-polarized with two rectifier diodes. Each diode rectifies one of the two received polarizations, and the DC signals add. The rectenna is a 19×19-mm square patch, with a 6×6-cm square ground plane on a Rogers Duroid 6010 substrate (e r = 10.2, thickness = 50 mil) chosen to reduce the antenna size. A Schottky diode is connected at each of the two centers of the two orthogonally polarized patch radiating edges (see Figure 32-25a). A via isolated from the patch ground plane terminates each diode to RF/DC ground, and the DC output is taken from the RF short in the center of the patch. Figure 32-25b shows a histogram of measured power levels for an incident power level of 170 mW/cm2 for different incident polarization combinations, which will all be present in the multipath environment envisioned in this application. This rectenna operates with incident power levels as low as 10 mW/cm2 and is capable of powering a low-power wireless sensor. Broadband Rectenna Design For low-power applications, as is the case for collected ambient energy, there is generally not enough power to drive the diode in a high-efficiency mode. Furthermore, rectification over multiple octaves requires a different approach from standard matching techniques. In a rectenna application, the antenna itself can be used as the matching mechanism instead of using a transmission-line or lumped-element matching circuit. The antenna design is therefore heavily dependent on the diode characteristics. A source-pull of the diode is performed over a given area of the Smith chart for different input powers, and the resulting DC voltage

(a)

(b)

FIGURE 32-25 (a) Measured rectified DC power as a function of the DC load for different incident power levels. A photograph of dual-polarized 2.4-GHz patch rectenna is shown in the inset. (b) Histogram of measured DC power for 170 mW/cm2 incident power density and different polarizations of the patch.



ACTIVE ANTENNAS

is quantified for each source impedance. The region of optimal source impedance is used for optimizing the antenna design. Since the harmonic balance technique calculates the response at all harmonics of the excitation frequency, in the simulation an assumption must be made for the impedance seen by the reflected harmonics, and in the presented case this impedance was set to the impedance of a broadband, self-complementary antenna, 189 Ω. Usually, the optimal source impedance moves counter-clockwise along a constant admittance circle with increasing frequency due to the junction capacitance. For maximal power transfer, the antenna impedance would match the optimal diode impedance for all frequencies. Since this is difficult to accomplish, a possible suboptimal approach is to present a constant impedance to the diode by using a frequency-independent antenna element. An equiangular spiral was chosen as the array element since it is uniplanar with convenient feedpoint for diode connection, and can be left- or right-hand circularly polarized. A single element was simulated with full-wave CAD tools (Ansoft’s Ensemble and Zeland’s IE3D) resulting in a one-port frequency-dependent impedance that becomes the diode load in the rectenna. A diode is connected at the antenna feed and the resulting rectenna element performance is shown in Figure 32-26a. The disagreement around 4 GHz is believed to be caused by the 1-cm-long unbalanced coaxial feed, which is not part of the active antenna. A 64-element array of left- and right-hand circularly polarized spiral elements, each with a rectifier diode, is shown in Figure 32-26b, as described in Hagerty et al.22 The RF powers received independently by each element are summed upon rectification as DC currents and/or voltages. The array was characterized for incident power densities ranging from tens of nW/cm2 to 0.1 mW/cm2. The rectenna array receives radiation from all directions (there is no ground plane), and to measure the re-radiated power at the harmonics, the array was placed halfway between a transmit and receive test antenna, perpendicular to the line-of-sight axis. The DC voltage is measured across a 100-Ω load. The rectification efficiency reaches the 20% range for an incident power density of 0.1mW/cm2 and arbitrary polarization. The array scaling in size is trivial, since there is no RF feed network, and only the DC output network is required. If space allows, this type of array can be made in the form of wallpaper for collecting as much power as possible. The DC collecting lines can be made to be reconfigurable, so that either current or voltage is summed, depending on the DC load.

(a)

(b)

FIGURE 32-26 (a) Nonlinear harmonic-balance simulation and measurements of the DC rectified voltage response across 60Ω for the spiral antenna in the inset, over a 4:1 bandwidth (after J. A. Hagerty et al22 © IEEE 2004); (b) photograph of 64-element dual-circularly polarized 2- to 18-GHz spiral rectenna array



CHAPTER THIRTY-TWO

(a)

(b)

(c)

FIGURE 32-27 (a) Unit cell of a 36-element T/R optically controlled active array. The optical power routes the signal between a receive-mode LNA and a transmit-mode PA. (b) Photograph of a switchless simplex T/R series-fed amplifier patch array at 24 GHz. (c) A 4-mm-long infrared imaging antenna and detector.

32.7 OTHER ACTIVE ANTENNAS The goal of this section is to briefly overview some interesting active antennas that do not fit any of the categories discussed in Sections 2 to 6, but have found useful applications or have promise for future front ends. Other design methodologies75 for antenna-circuit modules are not discussed here. Figure 32-27a shows a unit cell of a 10-GHz optically switched transmit-receive (T/R) multibeam antenna array.76 Two pairs of multimode optical fibers are held by an FR4 mount and the light from a laser diode is coupled to off-the-shelf photodetectors, which control the bias of pin-diode SPDT switches. The signal between two patch antennas in each unit cell of the amplifier array is routed either through an LNA or PA at nsec speed with only pJ of optical energy, and the optical fibers are transparent at 10 GHz and do not affect the radiation pattern. Figure 32-27b shows a single T/R series-fed 24-GHz patch array fed by a bi-directional (PALNA) chip amplifier, where the T/R path is controlled by bias, eliminating switch loss.77 Figure 32-27c shows an infrared antenna with a tunnel-diode detector, and other similar extremely high-frequency antennas have been demonstrated for millimeter-wave and THz imaging applications.78.79

REFERENCES 1. R. J. Mailloux, Phased Array Antenna Handbook, 2nd Ed. (Norwood, MA: Artech House Publishers, 2005). 2. R. C. Hansen, Phased Array Antennas (New York: John Wiley & Sons, Inc., 1998). 3. Q. Dongjiang, R. Molfino, S. M. Lardizabal, B. Pillans, P. M. Asbeck, and G. Jerinic, “An Intelligently Controlled RF Power Amplifier with a Reconfigurable MEMS-varactor Tuner,” IEEE Trans. Microwave Theory Tech., vol. 53, no. 3, part 2 (March 2005): 1089–1095. 4. M. Forman and Z. Popovic, “A Tunable Second-Resonance Cross-Slot Antenna,” 1997 IEEE AP-S Int. Symp. Digest (1997): 18–21. 5. N. Behdad and K. Sarabandi, “A Varactor-Tuned Dual-Band Slot Antenna,” IEEE Trans. Antennas and Propagation, vol. 54, no. 2 (February 2006): 401–408. 6. K. Chang, K. Hummer, and G. Gopalakrishnan, “Active Radiating Element Using FET Source Integrated with Microstrip Patch Antenna,” Electronics Letters, vol. 24, no. 21 (October 1988): 1347–1348. 7. R. H. Rasshofer and E. M. Biebl, “A Direction Sensitive, Integrated, Low Cost Doppler Radar Sensor for Automotive Applications,” 1998 IEEE MTT-S Int. Microwave Symp. Digest (1998): 1055–1058.



32-31

8. Z. Popovic, R. Weikle, M. Kim, and D. Rutledge, “A 100-MESFET Planar Grid Oscillator,” IEEE Trans. Microwave Theory Tech., vol. 39, no. 2 (February 1991): 193–200. 9. J. Birkeland and T. Itoh, “A 16 Element Quasi-Optical FET Oscillator Power Combining Array with External Injection Locking,” IEEE Trans. Microwave Theory Tech., vol. 40, no. 3 (March 1992): 475–781. 10. A. Mortazawi and B. DeLoach, “Spatial Power Combining Oscillators Based on an Extended Resonance Technique,” IEEE Trans. Microwave Theory Tech., vol. 42, no. 12 (December 1994): 2222–2228. 11. P. Liao and R. A. York, “A New Phase-Shifterless Beam-Scanning Technique Using Arrays of Coupled Oscillators,” IEEE Trans. Microwave Theory Tech., vol. 41, no. 10 (October 1993): 1810–1815. 12. S. Hollung, A. E. Cox, and Z. Popovic, “A Bi-directional Quasi-Optical Lens Amplifier,” IEEE Trans. Microwave Theory Tech., vol. 45, no. 12, part 2 (December 1997): 2352–2357. 13. S. Nam, T. Uwano, and T. Itoh, “Microstrip-Fed Planar Frequency-Multiplying Space Combiner,” IEEE Trans. Microwave Theory Tech., vol. 35, no. 12 (December 1987). 14. K. Chang, R. A. York, P. S. Hall, and T. Itoh, “Active Integrated Antennas,” IEEE Trans. Microwave Theory Tech., vol. 50, no. 3 (March 2002): 937–944. 15. J. B. Hacker, R. M. Weikle, M. Kim, M. P. De Lisio, and D. B. Rutledge, “A 100-Element Planar Schottky Diode Grid Mixer,” IEEE Trans. Microwave Theory Tech., vol. 40, no. 3 (March 1992): 557–562. 16. A. Luukanen, E. N. Grossman, A. J. Miller, P. Helisto, J. S. Penttila, H. Sipola, and H. Seppa, “An Ultra-Low Noise Superconducting Antenna-Coupled Microbolometer with a Room-Temperature Read-Out,” IEEE Microwave and Wireless Components Letters, vol. 16, no. 8 (August 2006): 464–466. 17. C. W. Pobanz and T. Itoh, “A Conformal Retrodirective Array for Radar Applications Using a Heterodyne Phased Scattering Element,” 1995 IEEE MTT-S Int. Microwave Symp. Digest, vol. 2, (1995): 905–908. 18. R. J. Hwu, C. F. Jou, W. W. Lam, U. Lieneweg, D. C. Steit, N. C. Luhmann, J. Maserjian, and D. B. Rutledge, “Watt-Level Millimeterwave Monolithic Diode-grid Frequency Multipliers,” 1988 IEEE MTT-S Int. Microwave Symp. Digest, vol. 1 (1988): 533–536. 19. B. Strassner and K. Chang, “Passive 5.8-GHz Radio-Frequency Identification Tag for Monitoring Oil Drill Pipe,” IEEE Trans. Microwave Theory Tech., vol. 51, no. 2, part 1 (February 2003): 356–363. 20. C. Walsh, S. Rondineau, M. Jankovic, G. Zhao, and Z. Popovic, “A Conformal 10-GHz Rectenna for Wireless Powering of Piezoelectric Sensor Electronics,” IEEE MTT IMS Digest (June 2005): 143–146. 21. P. Koert and J. T. Cha, “Millimeter Wave Technology for Space Power Beaming,” IEEE Trans. Microwave Theory Tech., vol. 40, no. 6 (June 1992): 1251–1258. 22. J. A. Hagerty, F. B. Helmbrecht, W. H. McCalpin, R. Zane, and Z. Popovic, “Recycling Ambient Microwave Energy with Broad-band Rectenna Arrays,” IEEE Trans. Microwave Theory Tech., vol. 52, no. 3 (March 2004): 1014–1024. 23. G. D. Boreman, C. Fumeaux, W. Herrmann, F. K. Kneubuhl, and H. Rothuizen, “Tunable Polarization Response of a Planar Asymmetricspiral Infrared Antenna,” Optics Letters, vol. 23, no. 24 (December 1998): 1912–1914. 24. J. Vian and Z. Popovic, “A Transmit/Receive Active Antenna with Fast Low-Power Optical Switching,” IEEE Trans. Microwave Theory Tech., vol. 48, no. 12 (December 2000): 2686– 2691. 25. K. D. Stephan and T. Itoh, “A Planar Quasi-Optical Subharmonically Pumped Mixer Characterized by Isotropic Conversion Los,” IEEE Trans. Microwave Theory Tech., vol. 32, no. 1 (January 1984): 97–102. 26. M. Weiss and Z. Popovic, “A 10 GHz High-Efficiency Active Antenna,” 1999 IEEE MTT-S Int. Microwave Symp. Digest, vol. 2 (1999): 663–666.



CHAPTER THIRTY-TWO

27. Z. B. Popovic and R. A. York, Active and Quasi-Optical Arrays for Solid-State Power Combining (New York: John Wiley & Sons, Inc., 1997). 28. R. Garg, P. Bhartia, I. Bahl, and A. Ittipiboon, Microstrip Antenna Design Handbook (Norwood, MA: Artech House Publishers, 2001). 29. S. H. Al-Charchafchi and M. Frances, “Electronically Tunable Microstrip Patch Antennas,” 1998 IEEE AP-S Int. Symp. Digest, vol. 1 (June 1998): 304–307. 30. J. Ollikainen, O. Kivekas, and P. Vainikainen, “Low-loss Tuning Circuits for Frequency-Tunable Small Resonant Antennas Personal, Indoor and Mobile Radio Communications, 2002,” The 13th IEEE International Symposium, vol. 4 (September 2002): 1882–1887. 31. J. A. Navarro, Y. Shu, and K. Chang, “Wideband Integrated Varactor-Tunable Active Notch Antennas and Power Combiners,” 1991 IEEE MTT-S Int. Microw. Symp. Digest, vol. 3 (1993): 1257–1260. 32. D. Peroulis, K. Sarabandi, and L. Katehi, “Design of Reconfigurable Slot Antennas,” IEEE Trans. Antennas and Propagation, vol. 53, no. 2 (February 2005): 645–654. 33. J. Roscoe, L. Shafai, A. Ittipiboon, M. Cuhaci, and R. Douville, “Tunable Dipole Antennas,” 1993 IEEE AP-S Int. Symp. Digest (1993): 672–675. 34. D. Anagnostou, G. Zheng, M. Cryssomallis, J. Lyke, G. Ponchak, J. Papapolymerou, and C. Christodoulou, “Design, Fabrication and Measurements of an RF-MEMS-Based Self-Similar Reconfigurable Antenna,” IEEE Trans. Antennas and Prop., vol. 54, no. 2 (February 2006): 422–433. 35. A. M. Castro-Vilaro and R. A. Rodriguez-Solis, “Tunable Folded-Slot Antenna with Thin Film Ferroelectric Material,” 2003 IEEE AP-S Int. Symp. Digest, vol. 2 (June 2003): 549–552. 36. A. Mortazawi, T. Itoh, and J. Harvey, Active Antennas and Quasi-Optical Arrays (New York: IEEE, Inc., 1999). 37. B. K. Kormanyos, S. E. Rosenbaum, L. P. Katehi, and G. M. Rebeiz, “Monolithic 155 GHz and 215 GHz Quasi-Optical Slot Oscillators,” 1994 IEEE MTT-S Int. Microwave Symp. Digest (1994): 835–838. 38. J. A. Hagerty and Z. Popovic, “A 10 GHz Active Annular Ring Antenna,” 2002 IEEE AP-S Int. Symp. Digest, vol. 2 (June 2002): 284–287. 39. T. Perkins, “Microstrip Patch Antenna with Embedded Impatt Diode Oscillator,” 1986 IEEE AP-S Int. Symp. Digest, vol. 24 (June 1986): 447–450. 40. M. J. Wengler, B. Guan, and E. K. Track, “190-GHz Radiation from a Quasioptical Josephson Junction Array,” IEEE Trans. Microwave Theory Tech., vol. 43, no. 4, part 1 (April 1995): 984–988. 41. M. M. Kaleja, A. J. Herb, R. H. Rasshofer, G. Friedsam, and E. M. Biebl, “Imaging RFID System at 24 GHz for Object Localization,” 1999 IEEE MTT-S Int. Microwave Symp. Digest, vol. 4 (June 1999): 1497–1500. 42. M. M. Kaleja, “Active Integrated Antennas for Sensor and Communication Applications,” Ph.D. dissertation, Technische Universität München, 2001. 43. J. B. Hacker, M. P. De Lisio, K. Moonil, L. Cheh-Ming, L. Shi-Jie, S. W. Wedge, and D. B. Rutledge, “A 10-W X-band Grid Oscillator,” 1994 IEEE MTT-S Int. Microwave Symp. Digest, vol. 2 (1994): 823–826. 44. J. Bae, T. Unou, T. Fujii, and M. Mizuno, “Spatial Power Combining of Gunn Diodes Using an Overmoded-Waveguide Resonator at Millimeter Wavelengths,” IEEE Trans. Microwave Theory Tech., vol. 46, no. 12 (December 1998): 2289–2294. 45. R. A. York and T. Itoh, “Injection- and Phase-Locking Techniques for Beam Control, IEEE Trans. Microwave Theory Tech., vol. 46, no. 11 (November 1998): 1920–1929. 46. W. Duerr, W. Menzel, and H. Schumacher, “A Low-Noise Active Receiving Antenna Using a SiGe HBT,” IEEE Microwave and Guided Wave Lett., vol. 7, no. 3 (March 1997): 63–67. 47. D. Radulovic, A. Nesic, and I. Radnovic, “Impedance of Patch Antenna for Active Antenna Structures,” IEEE AP International Symposium Digest (July 2004): 3931–3934. 48. J. Luine L. Abelson, D. Brundrett, J. Burch, E. Dantsker, K. Hummer, G. Kerber, M. Wire, K. Yokoyama, D. Bowling, M. Nee1, S.Hubbell, and K. Li, “Application of a DC SQUID Array



49.

50. 51.

52. 53. 54.

55.

56.

57. 58.

59.

60.

61.

62.

63.

64.

65.

66. 67.

32-33

Amplifier to an Electrically Small Active Antenna,” IEEE Trans. On Applied Superconductivity, vol. 9, no. 2 (June 1999): 4141–4144. H. An, B. Nauwelaers, A. Van de Capelle, and R. Bosisio, “A Novel Measurement Technique for Amplifier-type Active Antennas,” IEEE MTT-S Int. Microwave Symp. Dig. (June 1994): 1473– 1476. J. Peeters Weem and Z. Popovic, “A Method for Determining Noise Coupling in a Phased Array Antenna,” 2001 IEEE MTT IMS Digest (May 2001): 271–274. S. Pajic and Z. Popovic, “An Efficient 16-element X-band Spatial Combiner of Switchedmode Power Amplifiers,” IEEE Trans. Microwave Theory Tech., vol. 51, no. 73 (July 2003): 1863–1870. J. Hubert, J. Schoenberg, and Z. Popovic “A Ka-band Quasi-optical Amplifier,” 1995 IEEE MTT-S Int. Symp. Dig. (May 1995): 585–588. T. Marshall, M. Forman, and Z. Popovic, “Two Ka-band Quasi-Optical Amplifier Arrays,” IEEE Trans. Microwave Theory Tech., vol. 47, no. 12 (December 1999): 2568–2573. N-S. Cheng, A. Alexanian, M. Case, D. Rensch, and R. A. York, “40-W CW Broadband Spatial Power Combiner Using Dense Finline Arrays,” IEEE Trans. Microwave Theory Tech., vol. 47, no. 7 (July 1999): 1070–1076. N-S. Cheng, P. Jia, D. Rensch, and R. A. York, “A 120-W X-Band Spatially Power Combined Solid-State Amplifier,” IEEE Trans. Microwave Theory Tech., vol. 47, no. 12 (December 1999): 2557–2562. R. M. Weikle, M. Kim, J. B. Hacker, M. P. DeLisio, Z. Popovic, and D. Rutledge, “Transistor Oscillator and Amplifier Grids,” Invited paper, Proc. IEEE, vol. 80, no. 11 (November 1992): 1800–1809. B. Deckman, D. Deakin, E. Sovero, and D. Rutledge, “A 5-W, 37-GHz Monolithic Grid Amplifier,” IEEE MTT-S Int. Microwave Symp. Dig. (2000): 805–808. M. P. DeLisio, B. Deckman, C.-T. Cheung, S. C. Martin, D. P. Nakhla, E. E. Hartmann, C. J. Rollison, J. B. Pacetti, and J. Rosenberg, “A Ka-band Grid Amplifier Module with over 10 Watts Output Power,” IEEE MTT-S Int. Microwave Symp. Dig., (2004): 83–86. C.-T. Cheung, R. Tsai, R. Kagiwada, and D. B. Rutledge, “V-band Transmission and Reflection Grid Amplifier Packaged in Waveguide,” IEEE MTT-S Int. Microwave Symp. Dig. (2003): 1863– 1866. S.C. Ortiz, J. Hubert, L. Mirth, E. Schlecht, and A. Mortazawi, “A High-Power Ka-Band QuasiOptical Amplifier Array,” IEEE Trans. Microwave Theory Tech., vol. 50, no. 2 (February 2002): 487–494. C. F. Jou, W. W. Lam, H. Z. Chen, K. S. Stolt, N. C. Luhmann, Jr., and D. B. Rutledge, “Millimeterwave Diode Grid Frequency Doubler,” IEEE Trans. Microwave Theory Tech., vol. 36, no. 11 (Nov. 1988): 1507–1514. A. Moussessian, M. Wanke, Y. Li, J-C. Chiao, S. J. Allen, T. Crowe, and D. B. Rutledge, “A Terahertz Grid Frequency Doubler,” IEEE Trans. Microwave Theory Tech., vol. 46, no. 11 (November 1998): 1976–1981. M. Sironen, Y. Qian, and T. Itoh, “A Subharmonic Self-Oscillating Mixer with Integrated Antenna for 60-GHz Wireless Applications,” IEEE Trans. Microwave Theory Tech., vol. 49, no. 3 (March 2001): 442–450. J. B. Hacker, R. M. Weikle, M. Kim, M. P. De Lisio, and D. B. Rutledge, “A 100-Element Planar Schottky Diode Grid Mixer,” IEEE Trans. Microwave Theory Tech., vol. 40, no. 3 (March 1992): 557–562. C. Pobanz and T. Itoh, “A Conformal Retrodirective Array for Heterodyne Applications Using a Heterodyne Phase Scattering Element,” IEEE MTT-S International Microwave Symposium Digest, (1995): 905–908. Jim Glenn (ed.), The Complete Patents of Nikola Tesla (New York: Barnes and Noble Books, 1994): 346–360. W. C. Brown, “The History of Power Transmission by Radio Waves,” IEEE Transactions on Microwave Theory and Techniques, vol. 32, no. 9 (Sept. 1984): 1230–1242.



CHAPTER THIRTY-TWO

68. N. Shinohara and H. Matsumoto, “Experimental Study of Large Rectenna Array for Microwave Energy Transmission,” IEEE Transactions on Microwave Theory and Techniques, vol. 46, no. 3 (March 1998): 261–267. 69. T. Yoo and K. Chang, “Theoretical and Experimental Development of 10 and 35 GHz Rectennas,” IEEE Transactions on Microwave Theory and Techniques, vol. 40, no. 6 (June 1992): 1259– 1266. 70. L. W. Epp, A. R. Khan, H. K. Smith, and R. P Smith, “A Compact Dual-Polarized 8.51-GHz Rectenna for High-Voltage (50 V) Actuator Applications,” IEEE Transactions on Microwave Theory and Techniques, vol. 48, no. 1 (January 2000): 111–120. 71. Y. Fujino, T. Ito, M. Fujita, N. Kaya, H. Matsumoto, K. Kawabata, H. Sawada, and T. Onodera, “A Driving Test of a Small DC Motor with a Rectenna Array,” IEICE Trans. Commun., vol. E77-B, no. 4 (April 1994): 526–528. 72. J. O. McSpadden, R. M. Dickinson, L. Fan, and Kai Chang, “Design and Experiments of a HighConversion-Efficiency 5.8-GHz Rectenna,” IEEE MTT IMS Digest, vol. 2 (1998): 1161–1164. 73. B. Strassner and K. Chang, “A Circularly Polarized Rectifying Antenna Array for Wireless Microwave Power Transmission with over 78% Efficiency,” IEEE MTT IMS Digest (2002): 1535–1538. 74. W. C. Brown, “An Experimental Low Power Density Rectenna,” IEEE MTT-S International Microwave Symposium Digest (1991): 197–200. 75. K.C. Gupta and P. Hall (eds.), Analysis and Design of Integrated Circuit Antenna Modules (New York: John Wiley and Sons, Inc., 2000). 76. J. Vian and Z. Popovic, “A Transmit/Receive Active Antenna with Fast Low-power Optical Switching,” IEEE Trans. Microwave Theory Tech., vol. 48, no. 12 (December 2000): 2686– 2691. 77. D. Lu, D. Rutledge, M. Kovacevic, and J. Hacker “A 24-GHz Patch Array with a Power Amplifier/ Low-Noise Amplifier MMIC,” Intl. Journal of Infrared and Millimeter Waves, vol. 23 (May 2002): 693–704. 78. G. Boreman, C. Fumeaux, W. Herrman, F. Kneubuhl, and H. Rothuizen, “Tunable Polarization Response of a Planar Asymmetric-spiral Infrared Antenna,” Optics Letters, vol. 23, no. 24. (December 1998): 1912–1914. 79. C. Dietlein, J. D. Chisum, M. D. Ramírez, E. N. Grossman, A. Luukanen, and Z. Popovic, “Integrated Microbolometer Antenna Characterization from 109–650 GHz,” 2007 IEEE MTT IMS Digest (2007).



Chapter 33

Fractal Antennas Douglas H. Werner Joshua S. Petko Thomas G. Spence The Pennsylvania State University


33-2

33.2 USEFUL FRACTAL ANTENNA GEOMETRIES . . . . . . . . . . . . . . . .

33-2

33.3 ITERATED FUNCTION SYSTEMS . . . . . . . . . . . . . . . . . . . . . . . . . .

33-6

33.4 FRACTAL ANTENNA ELEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . .

33-9

33.5 FRACTAL ANTENNA ARRAYS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33-13 33.6 ANTENNA ARRAYS BASED ON FRACTAL AND APERIODIC TILINGS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33-21


Fractal Antennas 33-2

CHAPTER THIRTY-THREE

33.1 INTRODUCTION The term nature-based antenna design was recently coined in Neiss et al1 to describe the process of designing complex antenna systems based on concepts inspired by nature. Examples of nature-based techniques that have been recently applied to solve advanced antenna engineering design problems include fractal geometry, neural networks, fuzzy logic, genetic algorithms, particle swarm optimization, and ant colony optimization. In some cases, even two or more of these nature-based design techniques have been combined together to effectively solve a specific antenna design problem. An example of this can be found in Petko and Werner,2 where a powerful nature-based antenna array synthesis technique was introduced that incorporates aspects of both fractal geometry and genetic algorithms. Traditional approaches to the analysis and design of antenna systems have their foundation in Euclidean geometry. In this chapter, however, we focus on nature-based antenna design concepts that employ fractal geometry. The term fractal, meaning broken or irregular fragments, was originally used by Mandelbrot3 to describe a family of complex shapes that possess an inherent self-similarity or self-affinity in their geometrical structure. The original inspiration for the development of fractal geometry came largely from an indepth study of the patterns of nature. Fractals are abundant in nature, with a few examples of natural fractals being snowflakes, ferns, trees, coastlines, mountain ranges and even galaxies.4 Several book chapters and review articles have been published that provide a comprehensive overview of research in the area of fractal antenna engineering.5–9 This research can be divided into two categories, the first of which deals with the study of fractal-shaped antenna elements, and the second of which concerns the use of fractals in the design of antenna arrays. The next section presents a brief summary of some of the most popular types of fractal geometries found to be useful in antenna engineering applications. Section 33.3 then provides an introduction to the theory of iterated function systems (IFSs) along with a discussion of the important role they play in the mathematical description of fractals. Sections 33.4 and 33.5 summarize approaches that employ fractal concepts in the design of antenna elements and arrays, respectively. Finally, Section 33.6 is devoted to design methodologies for antenna arrays that are based on fractal and aperiodic tilings.

33.2 USEFUL FRACTAL ANTENNA GEOMETRIES This section presents a brief introduction to some of the more common fractal geometries that have been found to be useful in antenna engineering applications. This includes applications involving fractal-shaped antenna elements as well as conventional Euclidean antenna elements placed in fractal array configurations. The first and perhaps the most recognizable fractal geometry that will be considered is known as the Sierpinski gasket.4 Figure 33-1 illustrates the first few stages in the construction of the Sierpinski gasket fractal. The procedure for geometrically constructing this fractal begins with an equilateral triangle contained in the plane, shown in black as stage 1 of Figure 33-1. The next step in the construction process is to remove the central triangle whose vertices are located at the midpoints of the sides of the original stage 1 triangle, which leads to the geometry shown in stage 2 of Figure 33-1. This process is then repeated for the three remaining black triangles, with the result shown in stage 3 of Figure 33-1. The next stage (stage 4) in the construction of the Sierpinski gasket is also shown in Figure 33-1. The Sierpinski gasket fractal is



FRACTAL ANTENNAS

Stage 1

Stage 2

Stage 3

Stage 4

FIGURE 33-1 The first four stages in the generation of a Sierpinski gasket fractal

generated by carrying out this iterative process an infinite number of times. It follows from this definition that the Sierpinski gasket is an example of a self-similar fractal; i.e., it is composed from small but exact copies of itself. This self-similar property of the Sierpinski gasket fractal is illustrated in Figure 33-2. Another useful fractal is commonly referred to as the Koch snowflake.4 This fractal also starts out as an equilateral triangle in the plane, as depicted by the solid black triangle in stage 1 of Figure 33-3. However, unlike the Sierpinski gasket, which was formed by systematically removing smaller and smaller triangles from the original structure, the Koch snowflake is constructed by adding smaller and smaller triangles to the structure in an iterative fashion. The first four stages in the iterative process of constructing the Koch snowflake are illustrated in Figure 33-3. A number of structures based on purely deterministic or random fractal trees have also proven to be extremely useful in developing new design methodologies for antenna elements and array configurations. An example of a deterministic fractal tree is shown in Figure 33-4a. A ternary (three-branch) generator is used for the first three stages of growth. A fractal-random tree, on the other hand, is a type of fractal tree which uses multiple generators selected in random order to form the tree structure. Figure 33-4b shows the first three stages of growth of a random fractal tree. In this case the tree is grown by randomly selecting between two generators, one with three branches and the other with two branches. Fractal-random trees are more representative of natural trees or plants than deterministic fractal trees, which appear too ordered. The examples presented in Figure 33-4 are of two-dimensional fractal trees. However, the concept of a fractal tree can be easily generalized to three-dimensions. Figure 33-5 shows the first four stages in the growth of a three-dimensional deterministic fractal tree.8,10

FIGURE 33-2 Illustration of the self-similar geometry of the Sierpinski gasket fractal



Stage 1


Stage 2

Stage 3

Stage 4

FIGURE 33-3 The first four stages in the generation of a Koch fractal snowflake

FIGURE 33-4 Example of (a) deterministic fractal tree and (b) fractal-random tree

Stage 1

Stage 2

Stage 3

Stage 4

FIGURE 33-5 First four stages of growth of a three-dimensional fractal tree with a four-branch generator (after Petko and Werner10 © IEEE 2004)



FRACTAL ANTENNAS

Stage 1

Stage 2

Stage 3

Stage 4

FIGURE 33-6 The first four stages of growth for a space-filling Hilbert curve

The space-filling properties of the Hilbert curve and related curves make them attractive candidates for use in the design of fractal antennas. The first four stages in the construction of the Hilbert curve are shown in Figure 33-6.4,11 The Hilbert curve is an example of a space-filling fractal curve that is self-avoiding (i.e., has no intersection points). Another important example of a space-filling curve with applications in antenna engineering is the Peano-Gosper curve.3,11 Figure 33-7 illustrates the first three stages in the construction of the Peano-Gosper curve. The initiator is shown in Figure 33-7a as the dashed-line segment superimposed on the stage 1 generator. A copy of the stage 1 generator is shown again in Figure 33-7b as the dashed curve. Also shown in Figure 33-7b is the stage 2 Peano-Gosper curve, which is obtained by replacing each of the seven segments in the stage 1 generator by an appropriately scaled copy of itself. The next iteration (stage 3) of the Peano-Gosper curve is shown in Figure 33-7c. At this stage, the space-filling property of the curve starts to become more apparent. Fractal tiles, or fractiles, represent another important category of fractal objects that has found application in antenna theory and design.3,12 Fractiles are tiles or islands that possess fractal boundaries. They are a unique subset of all possible tile geometries that can be used to cover the plane without overlapping or leaving any gaps in between the tiles. We consider here two specific types of fractiles, known as fudgeflakes and Gosper islands. Both fudgeflakes and Gosper islands can be used to cover the plane via a tiling.3,12 The first through sixth iterations of the fudgeflake are shown in Figure 33-8, while the first through

(a) Stage 1

(b) Stage 2

(c) Stage 3

FIGURE 33-7 The first three stages of the space-filling Peano-Gosper curve. The initiator is shown as the dashed-line segment superimposed on the stage 1 generator, which is shown as the dashed curve superimposed on the stage 2 generator.




1st

2nd

3rd

4th

5th

6th

FIGURE 33-8 Geometry of the first through sixth iterations of the fudgeflake fractile

fifth iterations of the Gosper island are shown in Figure 33-9. Finally, Figure 33-10 demonstrates how fudgeflake and Gosper island fractiles can be used to provide a perfect covering of the plane (i.e. no gaps or overlaps between tiles).

33.3

ITERATED FUNCTION SYSTEMS

Iterated function systems (IFSs) are the mathematical language of fractals. They provide a unified approach to the theory of fractal geometry and represent an extremely versatile tool for conveniently generating a wide variety of useful fractal structures.4,13 These IFSs are based on a collection of contractions achieved through the application of a series of affine transformations w defined as  x  a w  =   y  c

b  x  e     +  d  y  f 

(33-1)

or equivalently as w( x, y) = (ax + by + e, cx + dy + f )

(33-2)

where a, b, c, d, e, and f are real numbers. Hence, the affine transformation w is represented by six parameters, which may be expressed using the compact notation a  c

b

e  f

d

(33-3)

such that a, b, c, and d control rotation and scaling, while e and f govern linear translation. Now if we let w1, w2, …, wN be a set of affine linear transformations and A be the initial geometry, then a new geometry can be formed by applying the set of transformations to

1st

2nd

3rd

4rh

5th

FIGURE 33-9 Geometry of the first through fifth iterations of the Gosper island fractile



FRACTAL ANTENNAS

(a)

(b)

FIGURE 33-10 A covering of the plane by (a) fudgeflake fractiles and (b) Gosper island fractiles

the original geometry A and collecting together the results from w1(A), w2(A), …, wN (A) in the following way: N

W ( A) = ∪ wn ( A)

(33-4)

n =1

where W is known as the Hutchinson operator.4 A fractal geometry can be obtained by repeatedly applying W to the previous geometry in an iterative fashion. For example, if the set A0 represents the initial geometry, then this iterative process would yield a sequence of Hutchinson operators given by A1 = W ( A0 ), A2 = W ( A1 ), …, Ak +1 = W ( Ak )

(33-5)

An IFS generates a sequence that converges to a final image A∞ in such a way that W ( A∞) = A∞

(33-6)

This image is called the attractor of the IFS and represents a “fixed point” of W, where the “points” in this case are actually defined as sets. The IFS procedure for generating the well-known Koch fractal curve is demonstrated in Figure 33-11. In this case, the initial set A0 is the line interval of unit length such that A0 = {x: x ∈[0,1]}. Four affine linear transformations (i.e., N = 4) are then applied to A0 as depicted in Figure 33-11. Next, the results of these four linear transformations are combined together using Eq. 33-4 to form the first iteration of the Koch curve denoted by A1. The second iteration of the Koch curve A2 may then be obtained by applying the same four affine transformations to A1 and again using Eq. 33-4 to combine the results. Higher-order versions of the Koch curve are generated by simply repeating the iterative process until the desired resolution is achieved. The first four iterations of the Koch curve are shown in Figure 33-12. This sequence of curves eventually converges to the actual Koch fractal (represented by A∞) as the number of iterations approaches infinity. IFSs have proven to be a very powerful tool in the analysis and design of fractal antennas. This is primarily because they provide a general framework for the description, classification, and manipulation of fractals.13 To further illustrate this important point, an IFS approach was used to produce the Sierpinski gasket shown in Figure 33-13a and the treelike structure shown in Figure 33-13b, two very different fractal objects. The IFS codes for generating these two objects have also been provided in the figure.4




FIGURE 33-11 The standard Koch fractal curve as an IFS (after Werner and ganguly9 © IEEE 2003)

Iteration 1

Iteration 2

Iteration 3

Iteration 4

FIGURE 33-12 The first four stages in the generation of the standard Koch fractal curve via an IFS approach

FIGURE 33-13 The IFS code for a Sierpinski gasket and a fractal tree (after Werner and Ganguly9 © IEEE 2003)



FRACTAL ANTENNAS

33.4

FRACTAL ANTENNA ELEMENTS

Some of the more common fractal geometries that have found applications in antenna engineering are depicted in Figure 33-14. The Koch snowflakes and islands have been primarily used to develop new designs for miniaturized loop as well as microstrip patch antennas.8,9,14,15 Moreover, fractile concepts have also been employed in the design of fudgeflake and Gosper island microstrip antennas.16 The self-similar structure of Sierpinski gaskets and carpets has been exploited to develop multiband and broadband antenna elements.7,9,17,18,19,20 Finally, novel designs for miniaturized dipole and monopole antennas have been developed based on a variety of Koch curves and fractal trees.7,8,9,10,21,22,23 The self-similar properties of fractal trees have been utilized to create antenna elements with multiband characteristics.24–27 In addition, fractal tree geometries have been used as an effective means to miniaturize antenna designs. A set of dipole antennas has been studied that utilizes three-dimensional fractal tree structures as end-loads in Gianvittorio and Rahmat-Samii.8 This set of antennas, illustrated in Figure 33-5, has been shown to exhibit lower resonant frequencies than standard dipole antennas of comparable length while not significantly affecting the bandwidth or the gain. An investigation of performance trends of various three-dimensional end-loaded fractal tree dipoles has also been carried out in Petko and Werner.10 It has been shown that denser fractal tree end-loads exhibit greater reduction in resonant frequency than less dense structures. The influence of branch elevation angles of the fractal tree generator was also studied in Petko and Werner.10 Figure 33-15 plots the value of the VSWR that corresponds to the resonant frequency of various four-branch fractal tree antenna designs with different branch elevation angles. The results show that fractal tree antennas with small branch elevation angles have a lower VSWR but a higher

(a)

(b)

(c)

FIGURE 33-14 Some fractal geometries commonly used in antenna applications: (a) Koch curves and fractal trees have been used to create miniaturized dipole and monopole antenna elements. (b) Koch snowflakes and islands have been used to create loop and patch antennas. (c) Sierpinski gaskets and carpets have been used to create multiband antenna elements.




FIGURE 33-15 VSWR vs. elevation angle resonant frequency for a set of stage 3, four-branch fractal tree antennas with elevation angles ranging from 10° to 90°. The VSWR was calculated with respect to a 50-Ω line (after Petko and Werner10 © IEEE 2004).

resonant frequency than those with large elevation angles. This trend continues until the resonant frequency reaches a minimum for designs with elevation angles of approximately 50°. Designs with elevation angles beyond 50° are not optimal because both the resonant frequency and VSWR increase with the increase in elevation angle. These trends have been used to design more effective miniature fractal tree dipoles, such as one created from a six-branch generator with branch elevation angles alternating between 30° and 50°. Figure 33-16 illustrates the geometry of the end-loads for the first three iterations of this dense fractal tree antenna. In Figure 33-17, the reflection properties (i.e., S11 and VSWR) of these dense fractal tree antennas are compared with the reflection properties of the four-branch fractal trees illustrated in Figure 33-5. The second-stage, sixbranch 50°–30° antenna is shown to have a resonant frequency of 920 MHz, which is identical to the resonant frequency of the third iteration for the four-branch fractal tree antenna. The resonant frequency for the third-iteration, six-branch 50°–30° antenna is shown to be 790 MHz. This resonance occurs at a frequency 57 percent lower than the 1820-MHz resonance of an equivalently sized conventional Euclidian dipole antenna, and is 70 MHz lower than the fourth iteration of the four-branch fractal tree antenna. One of the key advantages of end-loading a dipole antenna with a fractal-tree geometry is evident when LC resonators or RF switches are incorporated throughout the end-load structure. Resonators and switches are often used to make conventional antennas either multiband or reconfigurable; however, in conventional structures, it is difficult to create

Stage 1

Stage 2

Stage 3

FIGURE 33-16 Fractal tree end-loads for the three-dimensional six-branch 30°–50° fractal tree dipole antenna (after Petko and Werner10 © IEEE 2004)


Fractal Antennas FRACTAL ANTENNAS

33-11

FIGURE 33-17 S11 versus frequency for a stage 1, stage 2, and stage 3 three-dimensional, six-branch 50°–30° fractal tree antenna. Also shown for comparison is the S11 for the first four stages of a standard fourbranch fractal tree antenna and an equivalently sized conventional dipole. The S11 was calculated with respect to a 50-Ω line (after Petko and Werner10 © IEEE 2004).

resonances that are close to each other in frequency. This difficulty arises primarily because in order to achieve this type of behavior, the reactive loads and switches would have to be placed in close physical proximity to each other on the antenna. Moreover, the reactive loads or switches are usually distributed in a series fashion along the antenna, which can compound the losses associated with the use of these devices. In a fractal tree end-load, however, these problems do not exist because reactive loads and switches can be spread throughout the fractal structure, placing the loads into a parallel configuration and separating the required physical distance between them. These reactive loads and switches essentially change the configuration, and hence the size, of the fractal tree endload for specific frequencies. In one study, RF switches are strategically placed throughout the end-load of a thirdstage, six-branch 50°–30° fractal tree dipole antenna (see Figure 33-16) to make it reconfigurable (i.e., tunable) over a bandwidth of 68 percent. The design uses 102 individual switches placed in each end-load of the antenna, as shown in Figure 33-18, to produce 20 reconfigurable states. The resulting dipole antenna was found to be reconfigurable from 770 MHz to 1570 MHz for a bandwidth of 800 MHz with a VSWR under 3:1 and reconfigurable from 970 MHz to 1570 MHz for a bandwidth of 560 MHz with a VSWR below 2:1. Also, since the design uses the fractal tree end-loads, the lowest resonance of this antenna occurs at a frequency 57 percent below that of a conventional linear dipole of equivalent length. In Figure 33-19 each of the 20 reconfigurable states is represented by a separate S11 curve (indicated by light-gray lines) with the lowest resonant frequency representing the state with all the switches closed and the highest resonant frequency representing the state




Side View

Top View

FIGURE 33-18 Switch layout for the reconfigurable six-branch, 50°–30° fractal tree dipole antenna (after Petko and Werner10 © IEEE 2004)

with all the switches open. The remaining states are achieved by opening the switches progressively from the top to the bottom. In addition, for three of the reconfigurable states the antenna effectively operates as a 50°–30° fractal tree dipole with fractal stages 3, 2, or 1. The S11 curves for these three special cases are indicated on the graph by thick dark-gray lines. Finally, the solid-black line represents the overall minimum S11 the antenna can be configured to for a particular frequency over the entire operating range of the antenna.

FIGURE 33-19 S11 versus frequency for the reconfigurable six-branch, 50°–30° fractal tree antenna. The light-gray curves represent each of the 20 states the antenna can be configured to operate. The dark-gray curves represent reconfigured states that operate as stage 1, stage 2, and stage 3 fractal tree antennas. The black line represents the overall minimum S11 the antenna can operate over the entire band. The S11 was calculated with respect to a 50-Ω line (after Petko and Werner10 © IEEE 2004)



33.5

33-13

FRACTAL ANTENNA ARRAYS

The advantages of fractal geometries are most evident when applied to antenna arrays. Their nature-based structure provides interesting radiation properties that are not found in conventional arrays. The radiation characteristics of deterministic fractal array configurations, such as Cantor linear arrays, Cantor ring arrays, and Sierpinski carpet planar arrays, have been considered in the literature.6,7,9,28,29 In some cases, these arrays were found to possess lower sidelobe levels than their conventional Euclidean counterparts. In addition, the fact that these fractal-based array geometries can be generated in an iterative fashion (i.e., via successive stages of growth starting from a simple generating array) has been exploited in order to develop rapid algorithms for use in efficient radiation pattern computations.6,7,9 Fractal concepts were also employed in Puente and Pous28 to design linear arrays that exhibit multiband radiation characteristics. Moreover, a fractal radiation pattern synthesis technique has been developed in Werner et al30 for the design of reconfigurable multiband linear and planar arrays. In addition to deterministic fractal arrays, properties of random fractals have been used to develop a design methodology for quasi-random arrays, called fractal-random arrays.7,9,31 Inspiration for fractal-random geometries are found in the natural world, where many objects closely resemble fractals but are also not completely deterministic. Figure 33-4 illustrates the process of introducing randomness into a fractal-random object. Each time a generator is applied to construct a fractal structure, it is randomly selected from a set of multiple generators. In this manner, random fractals are used to generate array configurations that are somewhere between completely ordered (i.e., periodic) and completely disordered (i.e., random). An example of a fractal-random array layout is shown in Figure 33-20a. The main advantage of this technique is that it yields sparse arrays that possess relatively low sidelobe levels, which is a property indicative of periodic arrays, but over a range of bandwidths that are comparable to random arrays. More recently, a robust and versatile design methodology for large size arrays has been developed that is based in part on the concept of an “optimal” fractal-random array.2 This technique incorporates a genetic algorithm with a specialized subset of fractal-random arrays called polyfractal arrays. A genetic algorithm is a nature-based design tool that optimizes a problem globally using the Darwinian notions of natural selection and survival

(a) 49 Element Fractal-Random Array

(b) 46 Element Polyfractal Array

FIGURE 33-20 Fractal-random tree analogy for the construction of linear fractal-random and polyfractal arrays. The connection factors in (b) are illustrated by the numbers above the branches of the four-branch and the three-branch generators. Antenna elements are positioned at the ends of the topmost branches (after Petko and Werner2 © IEEE 2005)




of the fittest. Useful overviews on the topic of genetic algorithms can be found in the literature.32,33,34 Although easily visualized, fractal-random geometries cannot exactly be reproduced from a small set of parameters and are not truly recursive when applied in an IFS. The subset of polyfractal arrays overcomes these limitations while retaining many of the desirable properties of fractal-random arrays. To construct a polyfractal array, the IFS technique introduced in Section 33.3 must first be modified to handle multiple generators. Polyfractal arrays are constructed from multiple generators, 1,2, …, M, each of which has a corresponding Hutchinson operator W1, W2, …,WM. Each Hutchinson operator Wm in turn contains Nm affine linear transformations, wm,1, wm,2 ,…, wm,Nm. These affine linear transformations, wm,n, are simplified from the more general six-parameter affine linear transformations introduced in Section 33.3 to a similitude approach consisting of three local parameters, rm,n, jm,n, ym,n, and one globalscale parameter sf that is applied throughout the polyfractal structure. This affine linear transformation can be represented by x s f cos(ϕ m,n + ψ m,n ) − s f sin(ϕ m,n + ψ m,n ) x rm,n cos ϕ m,n  ω m ,n   =    +  y s f sin(ϕ m,n + ψ m,n ) s f cos(ϕ m,n + ψ m,n )  y rm,n sin ϕ m,n 

(33-7)

In addition to these three local parameters, a fourth local parameter, km,n, is associated with each affine linear transformation. This parameter, referred to as the connection factor, is an integer value ranging from 1 to M, the number of generators used to construct the polyfractal array. In this generalized iterated function system approach, a Hutchinson operator, Wm, is used to construct a stage  + 1 polyfractal array from the set of possible stage  polyfractal arrays, F. Each affine linear transformation, wm,n, can only be performed on stage  polyfractal arrays where the generator applied at stage  matches the connection factor, km,n. Because the connection factors dictate how the affine linear transformations are applied, only one unique polyfractal array geometry can be associated with each Hutchinson operator. Therefore, the set of stage  polyfractal arrays, F, can be expressed by the notation F = {F,1 , F,2 ,…, F,M }

(33-8)

where the first subscript defines the level of the polyfractal array and the second subscript defines the generator employed at that level. Therefore, a polyfractal array of stage  + 1 constructed by generator m can be represented by F+1,m = Wm ({F,1 , F,2 ,…, F,M }) =

∪ ω m,n (F,κ n =1 Nm

m ,n

)

(33-9)

After the polyfractal structure is formed, the overall structure is adjusted by another globalscale parameter sg. The global-scale factor sf can be factored out of the Hutchinson operators to create the final normalized IFS construction procedure for a stage L polyfractal array: FL,m = sg ( s f ) L −  Wm

({ F

L −1,1

})

Nm

(

, FL−1,2 ,…, FL−1,M = sg ( s f ) L −  ∪ ω m,n FL−1,κ m ,n n =1

)

(33-10)

This process is illustrated for a linear polyfractal array in Figure 33-20b using the fractaltree analogy. In this figure, numbers representing the connection factors are attached to the ends of the generator branches. The branches ending with number 1 have generator 1 connected to their ends. Likewise, the branches ending with number 2 have generator 2 connected to their ends. The ends of the topmost branches represent the positions of the antenna elements. Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.


FRACTAL ANTENNAS

Recursive beamforming algorithms can be created for both fractal and polyfractal arrays. Many fractal array recursive beamforming algorithms are based on the pattern multiplication approach. The radiation pattern of a stage  fractal array is equal to the product of the radiation pattern of a stage -1 fractal subarray and the array factor of the appropriately scaled fractal generator. In other words, the stage  fractal array can be thought of as an array of stage -1 fractal subarrays. An equation expressing the stage  fractal subarray radiation pattern, FRL, is derived using the fractal similitude wn. The similitude is based on Eq. 33-7, using the local parameters, rn, jn, and yn and scale parameters sg and sf (the m is omitted because there is only one generator). To perform pattern multiplication, all subarray radiation patterns must be identical and oriented in the same direction. Therefore, the sum of jn and yn is required to be equal to a multiple of 2π, making the axes of symmetry of the subarrays parallel. The equation for a recursive beamforming algorithm based on pattern multiplication can be written as N

(

)

FRL (θ ,ϕ ) = FRL−1 (θ ,ϕ ) ∑ exp j  k sg ( s f ) L −  rn sin θ cos(ϕ − ϕ n )   n =1

(33-11)

Using isotropic sources as the initial subarray radiation pattern, the final stage L fractal array factor can be written in a similar manner as has been done in Werner et al6: AFLL (θ ,ϕ ) =

L

N

∑ exp j k Sδ −1rn sinθ cos(ϕ − ϕ n ) ∏ =1 n = 1

(33-12)

where d = 1/sf and S = sg (sf)L-1. Typically, the values of rn are scaled such that S can be set equal to one. The unique scaling procedure and connection factor–based construction allow the rapid recursive beamforming algorithms associated with fractal arrays to be generalized to handle polyfractal arrays. The fractal array recursive beamforming operation discussed above requires all subarrays to have the same radiation pattern and be oriented in the same direction. In that way, pattern multiplication can be employed. In the more general polyfractal array, there are multiple types of subarrays that do not necessarily point in the same direction. Therefore, these subarray patterns cannot be factored out of the sum and the resulting expression for the stage , generator m subarray pattern is given by FRL,m (θ , ϕ ) =

∑ (FRL−1,κ

Nm

n =1

m ,n

)

(

)

(θ , ϕ − ϕ m,n − ψ m,n ) exp j  k sg ( s f ) L −  rm,n sin θ cos(ϕ − ϕ m,n )  

(33-13)

This subarray radiation pattern is based on the set of stage -1 fractal subarray patterns. The final radiation pattern can be determined by using isotropic sources for the initial subarray radiation patterns and recursively applying the expression until the stage L radiation pattern is obtained. Figure 33-21 illustrates this process for the example array shown in Figure 33-20b. One of the main advantages of the recursive beamforming approaches associated with fractal and polyfractal arrays is that they can be exploited to considerably speed up the convergence of the genetic algorithm (GA). This allows the possibility of optimizing much larger size arrays than has previously been possible using other approaches. While conventional genetic algorithm techniques can be applied to polyfractal arrays of the same size, it is advantageous to generalize the optimization process to include polyfractal arrays of varying sizes. To accomplish this, the genetic algorithm crossover routine, where parameters of two parent arrays are combined to create new offspring, must be modified so that the parent chromosomes of the polyfractal arrays are broken apart into simpler pieces. This process, discussed in more detail in Petko and Werner,2 essentially combines




FIGURE 33-21 Rapid recursive beamforming algorithm for a two-generator polyfractal array (after Petko and Werner2 © IEEE 2005)

various affine linear transformations from each parent array and performs crossover at that level. The resulting transformations are then combined into the offspring arrays. The genetic algorithm mutation routine can also be generalized for polyfractal arrays of varying sizes. Genetic mutation typically consists of changing a single parameter; however, polyfractal arrays offer a wide range of flexibility. Mutations can be performed on any of the global parameters as well as on any transformation. In addition, mutation can also remove or add a gene to a generator. Mutation can also be used to switch the order of the generators. This process is also discussed in more detail in Petko and Werner.2 By allowing the algorithm to handle more general polyfractal array geometries, an added degree of flexibility is incorporated into the optimization process. When the optimization appears to reach premature convergence for a population of polyfractal arrays, a unique mutation process, called generator autopolyploidization, can be used to stimulate the evolution process.35 The generator autopolyploidization process divides each fractal-random generator into two identical parts. The connection factors that were used to select the previous generator are uniformly divided to choose between the two new generators. In this way the arrays are exactly the same; however, they are described using twice the number of local parameters, adding new flexibility for genetic evolution. After the generator duplication, each chromosome parameter is mutated a small amount through a perturbation process. This perturbation adds a degree of genetic diversity to the population that can aid in the overall evolution procedure. Finally, the population is ready for another period, or epoch, of genetic evolution. An example of this generator autopolyploidy process is illustrated in Figure 33-22 for the transformation of a one-generator fractal array into a two-generator polyfractal array. It is important to note how generator autopolyploidization can benefit the optimization process. Initially, the genetic algorithm is used to evolve simple solutions consisting of a small number of generators. The limited number of generators restricts the initial antenna



33-17

FIGURE 33-22 Generator autopolyploidization process for the conversion of a one-generator fractal array to a two-generator polyfractal array (after Petko and Werner35 © IEEE 2007)

geometries to a small area of the search space; however, the time required to evaluate the performance of each antenna is very small. The optimization takes advantage of this short evaluation time until convergence is reached. At this point the algorithm can get no further using this limited set of generators; therefore, the generator autopolyploidization process is applied to expand the number of generators, which in turn expands the search space. This expansion comes at the cost of increasing the evaluation time; however, the optimization now has a good idea of where to search for more complex solutions based on the simpler designs found during the previous epoch. On the contrary, if the genetic algorithm began by optimizing the more complex solutions, not only would it initially take much longer to evaluate each array, the large number of unfocused input parameters could overwhelm the optimization process. This cycle of autopolyploidization can be repeated numerous times, producing more and more complex solutions with each optimization epoch. If the evolutionary progress is summarized for each generation, the resulting diagram would exhibit a staircase pattern. Thus, generator autopolyploidization offers polyfractal-based genetic algorithms an efficient “path” to complex design solutions, first optimizing simple designs very quickly and then only adding the increasing levels of complexity when they are finally needed. Next, an example is presented to illustrate the GA design methodology for evolving an optimal uniformly excited linear polyfractal array. The design considered here is an example of a large-N array, i.e., an array that has a relatively large number of elements. In this case, an initial population of arrays is optimized to minimize the peak sidelobe level and maintain a narrow beamwidth for a polyfractal array with a 0.5l minimum interelement spacing. The initial population was based on 500, 2401-element periodic arrays with 0.5l spacing. These initial arrays were created using a polyfractal structure consisting of one generator. Three generator autopolyploidy operations were performed: one initially and two triggered after the population did not improve over 30 generations. Therefore the optimization process is divided into three epochs of evolution. The evolutionary diagram, shown in Figure 33-23, exhibits the characteristic staircase pattern resulting from the three generator autopolyploidy operations. During the first epoch, the polyfractal arrays had two generators. This epoch ended at generation 217, where the number of generators was doubled to four for the second epoch. Two hundred generations later, the third epoch began with each polyfractal array now having eight generators. The final 1616-element design was found after 700 generations and had a sidelobe level of −24.30 dB and a beamwidth of 0.056°. The recursive beamforming algorithm on average calculated radiation patterns 20 times faster for arrays with two generators, 15 times faster for arrays with four generators, and 10 times faster for arrays with eight generators when compared to a conventional discrete Fourier transform (DFT) approach. Table 33-1 summarizes these speed increases for each epoch. The array factor and geometrical layout of this antenna array are shown in Figure 33-24 with the corresponding performance properties of the antenna listed in Table 33-2.




Number of Generators and Radiation Pattern Evaluation Speed Increase per Epoch for the 1616-element Polyfractal Array Using the Recursive Beamforming Algorithm vs. a Conventional DFT Approach (Radiation patterns have been sampled at 72,000 points.) TABLE 33-1

Number of Generators Evaluation Speed Increase

TABLE 33-2

Epoch 2

Epoch 3

2

4

8

1937%

1431%

892%

Performance Properties of a 1616-element Genetically Optimized Polyfractal Array

Number of Elements 1616

Epoch 1

SLL (dB)

HPBW

Minimum Spacing

Average Spacing

−24.30

0.056°

0.5l

1.01l

FIGURE 33-23 Evolutionary diagram of a 1616-element genetically optimized polyfractal array (after Petko and Werner35 © IEEE 2007)

FIGURE 33-24 Radiation pattern and array layout for a 1616-element genetically optimized polyfractal array (after Petko and Werner35 © IEEE 2007)



FRACTAL ANTENNAS

FIGURE 33-25 Layout for a 1230-element, uniformly excited, Pareto optimized polyfractal array.

More complex design problems can have multiple criteria that must be optimized. In these cases, it is desirable to compare the solutions that lie upon the Pareto front (i.e., the set of all nondominated solutions). Several different types of genetic algorithms have been developed to search for this Pareto front.36− 43 An example of a uniformly excited polyfractal array evolved from one of these Pareto genetic algorithm optimization techniques39,40 is discussed here. The goal of the optimization is to minimize the peak sidelobe level at two frequencies: one in which the minimum spacing is 0.75l and another where the minimum spacing is 2.0l. The optimization evolved arrays from an initial population of fractal arrays based on a 1296-element periodic array. Three generator autopolyploidy processes were performed: one initially, one at generation 150, and one at generation 300. The optimization was terminated after 450 generations and the resulting Pareto front consists of arrays with between 1186 and 1230 elements. One solution with 1230 elements was chosen from the Pareto front for discussion. This example has a −19.3-dB sidelobe level at both the 0.75l and at the 2.0l minimum spacings when the mainbeam is pointed at broadside. The layout of this design is shown in Figure 33-25, while Table 33-3 summarizes the properties of this array at the two different design frequencies. While the array was optimized only at these two frequencies, the fractal-random properties of polyfractal arrays allow for low sidelobe levels over fairly wide ranges of bandwidth without the appearance of grating lobes. Figure 33-26 plots the sidelobe level performance of the polyfractal array shown in Figure 33-25 as a function of the minimum separation between elements in wavelengths and compares them to a 1296-element periodic array. When steered to broadside, the sidelobe level of the periodic array is −13.26 dB until the electrical spacing between elements increases to about 1l, when grating lobes first appear. In contrast, the polyfractal array maintains sidelobe levels less than −14 dB even up to 20l minimum spacings. This array was not optimized at a 20l minimum spacing; nevertheless, the robust physical properties of the polyfractal arrays in combination with the optimization goals at 0.75l and 2.0l minimum spacings have allowed the low sidelobe performance to be maintained far beyond the specified separation distance of 2.0l. This array also has a sidelobe level under −16 dB for up to 16l minimum spacing, and a sidelobe level under −17 dB up to 7l minimum spacing when the mainbeam is directed to broadside. In addition, one can see how the goals of the optimizer at the 0.75l and 2.0l minimum spacings have evolved an array with sidelobe levels under −19 dB for these lower frequencies.

Performance at the Two Optimization Target Frequencies for the 1230-element Polyfractal Array TABLE 33-3

0.75λ Minimum Spacing Number of Elements 1230

2.0λ Minimum Spacing

SLL

HPBW

Average Spacing

SLL

HPBW

Average Spacing

−19.31 dB

0.0078º

5.60λ

−19.31 dB

0.0029º

14.96λ




FIGURE 33-26 Sidelobe level performance of a uniformly excited, 1230-element, Pareto optimized polyfractal array over a range of bandwidth compared to the performance of a conventional uniformly excited periodic linear array. The polyfractal array was optimized at broadside for simultaneous minimum element spacings of 0.75l and 2.0l.

Finally, these arrays also exhibit their wideband properties when they are steered away from broadside. Figure 33-26 also plots the peak sidelobe level vs. minimum spacing when the mainbeam is steered 60 degrees from broadside. The grating lobes for the periodic array when steered to 60 degrees from broadside first appear at interelement spacings of 0.56l. On the other hand, the peak sidelobe level performance of the polyfractal array is still less than −14 dB even up to 20l minimum spacing. Figure 33-27 shows the radiation pattern for this optimized polyfractal array when the minimum spacing between elements is 8l and the beam is steered 60 degrees from broadside.

FIGURE 33-27 Radiation pattern for a uniformly excited, 1230-element, Pareto optimized polyfractal array with its mainbeam steered 60 degrees from broadside and an 8l minimum spacing between elements



33-21

33.6 ANTENNA ARRAYS BASED ON FRACTAL AND APERIODIC TILINGS This section is devoted to design methodologies for antenna arrays that are based on fractal and aperiodic tilings. Although a wide variety of tilings ranging from simple to complex patterns have been known since early civilizations (an example of an early tiling is shown in Figure 33-28), the formal mathematical theory of tilings has only been in existence for about a century.44 Tiling theory has found applications in many fields, such as crystallography, biology, and communication theory. More recently, tiling theory has been successfully applied to the field of electromagnetics, in particular, antenna array design using fractal tilings45,46,47 and aperiodic tilings.48 The unique geometry of these tilings can be exploited to generate antenna arrays that exhibit low sidelobe levels and suppressed grating lobes for minimum element spacings of at least one-wavelength. Moreover, it has been shown that robust element perturbation schemes based on the GA can be used to greatly improve the performance of these arrays.49,50,51 A modular design approach based on the theory of fractile geometry has been introduced in Werner et al45 for designing low-sidelobe arrays with suppressed grating lobes. A fractile array is defined as an array composed of a tiling of self-similar subarrays with fractal boundaries that covers the plane or a portion of the plane without any overlaps or gaps. A variety of different fractile array configurations have been investigated including Peano-Gosper, terdragon, six-terdragon, and fudgeflakes.45 Another important property of fractile arrays is that their self-similar tile geometry can be exploited to develop a rapid iterative procedure for calculating far-field radiation patterns that, for sufficiently large arrays, can be considerably faster than using a conventional DFT. One specific type of fractile array that has been reported on extensively in the literature is based on the Peano-Gosper family of space-filling curves (see Figure 33-7) and is therefore known as the Peano-Gosper fractile array (PGFA).45,46,47,49,50 The elements of the array are uniformly distributed along a Peano-Gosper curve as shown in Figure 33-29, which leads to a planar array configuration with an equilateral triangular (i.e., hexagonal) lattice on the interior that is bounded by an irregular closed Koch fractal curve around its perimeter. The Koch fractal boundary plus its interior form a Gosper island that can be used to cover the plane

FIGURE 33-28 Photograph of a tiling from the Alhambra in Granada, Spain (Courtesy of Douglas H. Werner)




Stage 1

Stage 2

FIGURE 33-29 Element distribution along stage 1 and stage 2 of the Peano-Gosper curve. The initial uniformly spaced element locations are denoted by a +. Examples of GA perturbed element locations are denoted by an o.

via a tiling.45,46 Furthermore, Gosper island tiles as illustrated in Figure 33-9 are self-similar and can be divided into seven smaller tiles, each representing a scaled copy of the original. Finally, the PGFA may be iteratively constructed to any arbitrary stage of growth based on a set formula for shifting, scaling, and rotating the generating array defined at stage 1. These unique properties were exploited in Werner et al45 and Bogard et al47 to develop a design methodology for deterministic arrays that possess several desirable properties, including a uniform current distribution, low sidelobe levels, relatively broad bandwidth, modular architecture, and the ability to perform rapid beamforming. Moreover, due to its equilateral triangular lattice, the PGFA will have no grating lobes for broadside operation even when the minimum spacing between elements is increased to one wavelength. Because the PGFA possesses a fractal boundary contour, its sidelobe levels are lower than an equivalent size rectangular array with an equilateral triangular (i.e., hexagonal) grid of elements on its interior. Figure 33-30 shows a comparison of the array factor for a 352-element rectangular-hexagonal array, a 344-element stage 3 PGFA, and a 361-element (19×19) square periodic array.47 All of the arrays have uniform element excitation and a minimum element spacing of one-wavelength. At this spacing the PGFA exhibits lower overall sidelobes than either its conventional square periodic or rectangular-hexagonal array counterparts. In the design of phased arrays (especially broadband phased arrays), care must be taken to ensure that grating lobes do not appear when the main beam is to be steered away from broadside. For example, when the element spacing of a PGFA is increased to one wavelength, no grating lobes occur in the entire visible region for broadside operation. However, when the minimum element spacing is one wavelength or greater and the main beam is scanned away from broadside, grating lobes occur in the far-field radiation pattern of the array. It has been shown in Bogard and Werner49 and in Bogard et al50 that a GA technique can be used to perturb the element locations in an optimal way on the interior of the PGFA to eliminate the grating lobes and provide acceptable sidelobe levels during scanning. This procedure results in a modular or tiled phased-array architecture that has an irregular fractal distribution of elements around its periphery together with an aperiodic arrangement of elements on its interior. During the design process, the element locations of the stage 1 generator are varied and subsequently used to generate higher-order stages of the PGFA through an efficient iterative procedure while, at each stage, the array maintains its broadband



33-23

FIGURE 33-30 Comparison of the normalized array factor versus q for a 361-element (19×19) square periodic array, a 352-element rectangular-hexagonal array, and a 344-element stage 3 PGFA at j = 0. In all three cases the array elements are uniformly excited and they have a minimum element spacing of one wavelength.

characteristics within a specified scan volume. Examples of GA perturbed element locations along the Peano-Gosper curve are shown in Figure 33-29. In Bogard and Werner49 and in Bogard et al,50 a stage 3 PGFA was optimized and its performance was compared with that of a uniform PGFA. The array had an initial element spacing of 2l and was optimized for scanning up to q = 30°. During scanning, grating lobes are suppressed over the entire visible region of the GA optimized design. Figure 33-31

FIGURE 33-31 Normalized array factor cut for the stage 3 PGFA with an initial element spacing of 2l. The main beam of the array was steered to q = 30°, j = 180°.




shows a q = 30° cut of the array factor for the optimized PGFA and the uniform PGFA when the main beam is steered to q = 30° and j = 180°. In this plane cut, a grating lobe is present in the array factor of the original uniform PGFA, whereas the optimized PGFA has complete grating lobe suppression over the same scanning range. Analogous to fractal-based tilings, aperiodic tilings can be used to generate antenna arrays that exhibit properties that are not typically associated with conventional periodic arrays. Unlike fractile arrays, which consist of a collection of elements that are placed a uniform distance apart along a space-filling curve (e.g. a Peano-Gosper curve), aperiodic tiling–based arrays consist of elements that are located at the vertices of an aperiodic tiling lattice.48 Arrays that are generated in this manner tend to have geometrical structures that contain local order and rotational symmetry, but are devoid of any translational symmetries. A common method that is employed to generate these tilings is based on a decomposition process,44 which is similar in some ways to the IFS techniques used to generate fractal geometries. In the process, tiles are decomposed into a collection of smaller tiles, which are either scaled copies of the original tiles or other tiles. This iterative process is continued until a large tiling is created. The Danzer tiling is an example of an aperiodic tiling that can be used to generate broadband low-sidelobe planar antenna arrays that are naturally thinned.48,51 Moreover, the Danzer tiling is a type of aperiodic tiling that comprises a collection of three specific triangular-shaped prototiles. The tiling is formed by covering the plane with the triangles while maintaining specific matching rules throughout. A portion of a Danzer tiling is shown in Figure 33-32 along with its three prototiles.44 An example of a Danzer tiling–based array is shown in Figure 33-33. The peak sidelobe level versus the minimum element spacing in terms of wavelengths for the array is shown in Figure 33-34. The tiling for the array has been scaled such that its physical dimensions correspond to a minimum element spacing, d, of l /2 at the lowest intended operating frequency, and it has been truncated to have a circular aperture with a 12l radius. In Figure 33-34, the performance of the Danzer array is compared with a conventional periodic array that has the same circular aperture and minimum element spacing. It is clear that the Danzer array outperforms the periodic array in terms of grating lobe suppression for large element spacings. Additionally, to fit within the same aperture size, the conventional periodic array requires approximately 1793 elements, while the naturally thinned (i.e., its mean interelement spacing is greater than l /2) Danzer array only requires 811 elements. Given the intrinsic properties of this class of arrays, they are attractive candidates for use as the baseline in various antenna designs, such as generating very sparse arrays51 and arrays with very wide bandwidth. One design methodology that is capable of producing

FIGURE 33-32 Truncated region of a Danzer tiling and its three prototiles



FRACTAL ANTENNAS

Original Danzer Array

Optimized Danzer Array

FIGURE 33-33 Geometry of the initial Danzer array and the GA optimized Danzer array

wideband planar arrays is based on a simple perturbation of the basic aperiodic array generation process. In this technique, an additional point (or points) is placed within the boundary of the prototiles that are used to generate the aperiodic tiling. The locations of these points are preserved within the prototiles as the overall aperiodic tiling is generated via the decomposition process. The result of this process is the formation of an aperiodic tiling that contains an additional point within each of its constituent tiles. Converting this to an antenna array yields fundamental elements at the vertices of the tiling along with elements at each of the additional points. The array can be scaled to have a specific minimum element spacing and then truncated to fit within a desired aperture. By adjusting the position of the point within each base tile, it is possible to greatly vary the radiation properties of the modified aperiodic tile–based array.

FIGURE 33-34 Sidelobe level performance of the initial Danzer array, GA optimized Danzer array, and periodic array. All of the arrays have a circular aperture with a radius of 12l and a minimum element spacing of d = 0.5l at the lowest operating frequency.




(a)

(b)

FIGURE 33-35 (a) Perturbation points of the Danzer prototiles and (b) an example of a truncated Danzer tiling with its corresponding perturbed array element locations

A GA-based optimizer was combined with the perturbation technique with the objective of designing a Danzer array that has the lowest possible sidelobe levels at d /l = 5. Since the Danzer array comprises a collection of three prototiles, the GA was only required to optimize the planar coordinates of the three additional perturbation points, as illustrated in Figure 33-35. During the optimization, all of the generated arrays were scaled to have a minimum element spacing of l /2 and were truncated to have a circular aperture with a radius of 12l at the lowest operating frequency. The base tiling for the optimization was the same one that was used in the formation of the array shown in Figure 33-33. The design that results from this optimization process is shown in Figure 33-33 and has a normalized radiation pattern with a maximum sidelobe level of −10.05 dB at a frequency corresponding to d /l = 5. The peak sidelobe level versus frequency for the array is shown in Figure 33-34. For this particular example, the simple perturbation scheme is capable of greatly extending the bandwidth (peak sidelobe level ≤ −10 dB with no grating lobes) of the basic Danzer array.

REFERENCES 1. K. M. Neiss, D. H. Werner, M. G. Bray, and S. Mummaready, “Nature-Based Antenna Design: Interpolating the Input Impedance of Fractal Dipole Antennas via a Genetic Algorithm Trained Neural Network,” Proc. USNC/URSI Nat. Radio Science Meeting (June 2002): 374. 2. J. S. Petko and D. H. Werner, “The Evolution of Optimal Linear Polyfractal Arrays Using Genetic Algorithms,” IEEE Transactions on Antennas and Propagation, vol. 53, no. 11 (November 2005): 3604−3615. 3. B. B. Mandelbrot, The Fractal Geometry of Nature (New York: W. H. Freeman, 1983). 4. H. O. Peitgen, H. Jurgens, and D. Saupe, Chaos and Fractals: New Frontiers of Science (New York: Springer-Verlag, Inc., 1992). 5. D. L. Jaggard, “Fractal Electrodynamics: From Super Antennas to Superlattices,” Fractals in Engineering, J. L. Vehl, E. Lutton, and C. Tricot (eds.) (New York: Springer-Verlag, 1997). 6. D. H. Werner, R. L. Haupt, and P. L. Werner, “Fractal Antenna Engineering: The Theory and Design of Fractal Antenna Arrays,” IEEE Antennas and Propagation Magazine, vol. 41, no. 5 (October 1999): 37–59. 7. D. H. Werner and R. Mittra (eds.), Frontiers in Electromagnetics, Chaps. 2 and 3 (Piscataway, New Jersey: IEEE Press, 2000). 8. J. P. Gianvittorio and Y. Rahmat-Samii, “Fractal Antennas: A Novel Antenna Miniaturization Technique, and Applications,” IEEE Antennas and Propagation Magazine, vol. 44, no. 1 (February 2002): 20–36.



33-27

9. D. H. Werner and S. Ganguly, “An Overview of Fractal Antenna Engineering Research,” IEEE Antennas and Propagation Magazine, vol. 45, no. 1 (February 2003): 38–57. 10. J. S. Petko and D. H. Werner, “Miniature Reconfigurable Three-Dimensional Fractal Tree Antennas,” IEEE Transactions on Antennas and Propagation, vol. 52, no. 8 (August 2004): 1945–1956. 11. H. Sagan, Space-Filling Curves (New York: Springer-Verlag, 1994). 12. G. A. Edgar, Measure, Topology, and Fractal Geometry (New York: Springer-Verlag, 1990). 13. M. F. Barnsley, Fractals Everywhere, 2nd Ed. (New York: Academic Press Professional, 1993). 14. N. Cohen, “Fractal Antennas: Part 1,” Communications Quarterly (Summer 1995): 7–22. 15. C. Borja and J. Romeu, “On the Behavior of Koch Island Fractal Boundary Microstrip Patch Antenna,” IEEE Transactions on Antennas and Propagation, vol. 51, no. 6 (June 2003): 1281–1291. 16. T. G. Spence and D. H. Werner, “Genetically Optimized Fractile Microstrip Patch Antennas,” Proc. 2004 IEEE Antennas and Propagation International Symposium, vol. IV (June 21–26, 2004, Monterey, CA): 4424–4427. 17. C. Puente, J. Romeu, R. Pous, X. Garcia, and F. Benitez, “Fractal Multiband Antenna Based on the Sierpinski Gasket,” IEE Electronics Letters, vol. 32, no. 1 (January 1996): 1–2. 18. C. Puente, J. Romeu, R. Pous, and A. Cardama, “On the Behavior of the Sierpinski Multiband Fractal Antenna,” IEEE Transactions on Antennas and Propagation, vol. 46, no. 4 (April 1998): 517–524. 19. G. J. Walker and J. R. James, “Fractal Volume Antennas,” IEE Electronics Letters, vol. 34, no. 16 (August 1998): 1536–1537. 20. C. T. P. Song, P. S. Hall, H. Ghafouri-Shiraz, and D. Wake, “Fractal Stacked Monopole with Very Wide Bandwidth,” IEE Electronics Letters, vol. 35, no. 12 (June 1999): 945–946. 21. N. Cohen, “Fractal and Shaped Dipoles,” Communications Quarterly (Spring 1996): 25–36. 22. C. Puente, J. Romeu, R. Pous, J. Ramis, and A. Hijazo, “Small But Long Koch Fractal Monopole,” IEE Electronics Letters, vol. 34, no. 1 (January 1998): 9–10. 23. C. P. Baliarda, J. Romeu, and A. Cardama, “The Koch Monopole: A Small Fractal Antenna,” IEEE Transactions on Antennas and Propagation, vol. 48, no. 11 (November 2000): 1773–1781. 24. C. Puente, J. Claret, F. Sagues, J. Romeu, M. Q. Lopez-Salvans, and R. Pous, “Multiband Properties of a Fractal Tree Antenna Generated by Electrochemical Deposition,” Electronics Letters, vol. 32, no. 25 (December 1996): 2298–2299. 25. M. Sindou, G. Ablart, and C. Sourdois, “Multiband and Wideband Properties of Printed Fractal Branched Antennas,” Electronics Letters, vol. 35, no. 3 (February 1999): 181–182. 26. D. H. Werner, A. Rubio Bretones, and B. R. Long, “Radiation Characteristics of Thin-wire Ternary Fractal Trees,” Electronics Letters, vol. 35, no. 8 (April 1999): 609–610. 27. K. J. Vinoy, J. K. Abraham, and V. K. Varadan, “Fractal Dimension and Frequency Response of Fractal Shaped Antennas,” 2003 IEEE Antennas and Propagation Society International Symposium, vol. 4 (June 22–27, 2003): 222–225. 28. C. Puente and R. Pous, “Fractal Design of Multiband and Low Side-lobe Arrays,” IEEE Transactions on Antennas and Propagation, vol. 44, no. 5 (May 1996): 730–739. 29. D. L. Jaggard and A. D. Jaggard, “Cantor Ring Arrays,” Microwave and Optical Technology Letters, vol. 19 (1998): 121–125. 30. D. H. Werner, M. A. Gingrich, and P. L. Werner, “A Self-Similar Fractal Radiation Pattern Synthesis Technique for Reconfigurable Multiband Arrays,” IEEE Transactions on Antennas and Propagation, vol. 51, no. 7 (July 2003): 1486–1498. 31. Y. Kim and D. L. Jaggard, “The Fractal Random Array,” Proc. IEEE, vol. 74, no. 9 (1986): 1278–1280. 32. D. E. Goldberg, Genetic Algorithms in Search, Optimization & Machine Learning (Reading, MA: Addison-Wiley Publishing Company, Inc., 1989). 33. R. L. Haupt and S. E. Haupt, Practical Genetic Algorithms (New York: John Wiley & Sons, Inc., 1998).




34. Y. Rahmat-Samii and E. Michielssen (eds.), Electromagnetic Optimization by Genetic Algorithms (New York: John Wiley & Sons, Inc., 1999). 35. J. S. Petko and D. H. Werner, “An Autopolyploidy-based Genetic Algorithm for Enhanced Evolution of Linear Polyfractal Arrays,” IEEE Transactions on Antennas and Propagation, vol. 55, no. 3, part I (March 2007): 583–593. 36. J. Horn, N. Nafpliotis, and D. E. Goldberg, “A Niched Pareto Genetic Algorithm for Multiobjective Optimization,” Proceedings of the First IEEE Conference on Evolutionary Computation, IEEE World Congress on Computational Intelligence, vol. 1 (June 1994): 82–87. 37. N. Srinivas and K. Deb, “Multiobjective Optimization Using Nondominated Sorting in Genetic Algorithms,” Evolutionary Computation, vol. 2, no. 3 (Fall 1994): 221–248. 38. C. Fonesca and P. Fleming, “An Overview of Evolutionary Algorithms in Multiobjective Optimization,” Evolutionary Computation, vol. 3, no. 1 (Spring 1995): 1–16. 39. E. Zitzler and L. Thiele, “An Evolutionary Algorithm for Multiobjective Optimization: The Strength Pareto Approach,” TIK-Report (Zurich, Switzerland: Swiss Federal Institute of Technology, May 1998). 40. E. Zitzler and L. Thiele, “Multiobjective Evolutionary Algorithms: A Comparative Case Study and the Strength Pareto Approach,” IEEE Transactions on Evolutionary Computation, vol. 3, no. 4 (November 1999): 257–271. 41. J. Knowles and D. Corne, “The Pareto Archived Evolution Strategy: A New Baseline Algorithm for Pareto Multiobjective Optimisation,” 1999 Congress on Evolutionary Computation (July 1999): 98–105. 42. D. A. Van Veldhuizen and G. B. Lamont, “Multiobjective Evolutionary Algorithms: Analyzing the State-of-the-Art,” Evolutionary Computation, vol. 8, no. 2 (2000) 125–127. 43. M. Laumanns, G. Rudolph, and H. Schwefel, “A Spatial Predator-Prey Approach to MultiObjective Optimization: A Preliminary Study,” Parallel Problem Solving from Nature—PPSN V, A. E. Eiben, M. Schoenauer, and H. P. Schwefel (eds). (Amsterdam, Holland: Springer-Verlag, 1998): 241–249. 44. B. Grünbaum and G. C. Shephard, Tilings and Patterns (New York: W. H. Freeman and Company, 1987). 45. D. H. Werner, W. Kuhirun, and P. L. Werner, “Fractile Arrays: A New Class of Tiled Arrays with Fractal Boundaries,” IEEE Transactions on Antennas and Propagation, vol. 52, no. 8 (August 2004): 2008–2018. 46. D. H. Werner, W. Kuhirun, and P. L. Werner, “The Peano-Gosper Fractal Array,” IEEE Transactions on Antennas and Propagation, vol. 51, no. 8 (August 2003): 2063–2072. 47. J. N. Bogard, D. H. Werner, and P. L. Werner, “A Comparison of the Peano-Gosper Fractile Array with the Regular Hexagonal Array,” Microwave and Optical Technology Letters, vol. 3, no. 6 (December 2004): 524–526. 48. V. Pierro, V. Galdi, G. Castaldi, I. M. Pinto, and L. B. Felsen, “Radiation Properties of Planar Antenna Arrays Based on Certain Categories of Aperiodic Tilings,” IEEE Transactions on Antennas and Propagation, vol. 53, no. 2 (February 2005): 635–644. 49. J. N. Bogard and D. H. Werner, “Optimization of Peano-Gosper Fractile Arrays Using Genetic Algorithms to Reduce Grating Lobes During Scanning,” Proc. IEEE International Radar Conference, (May 2005): 905–909. 50. J. N. Bogard, D. H. Werner, and P. L. Werner, “Optimization of Peano-Gosper Fractile Arrays for Broadband Performance Using Genetic Algorithms to Eliminate Grating Lobes During Scanning,” Proc. IEEE AP-S Int. Symp., vol. 1B (July 2005): 755–758. 51. T. G. Spence and D. H. Werner, “Thinning of Aperiodic Antenna Arrays for Low Side-Lobe Levels and Broadband Operation Using Genetic Algorithms,” Proc. 2006 IEEE AP-S Int. Symp., vol. 3 (July 2006): 2059–2062.



Chapter 34

Low Profile Antenna Performance Enhancement Utilizing Engineered Electromagnetic Materials Fan Yang University of Mississippi

Hossein Mosallaei Northeastern University

Yahya Rahmat-Samii University of California, Los Angeles CONTENTS 34.1 INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

34-2

34.2 CHARACTERIZATIONS AND DESIGNS OF ELECTROMAGNETIC BAND GAP (EBG) STRUCTURES. . . . .

34-3

34.3 LOW PROFILE WIRE ANTENNAS ON EBG GROUND PLANE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

34-7

34.4 PATCH ANTENNAS WITH ENHANCED PERFORMANCE USING EBG STRUCTURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34-11 34.5 SURFACE WAVE ANTENNAS WITH A MONOPOLE-LIKE RADIATION PATTERN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34-12 34.6 IMPEDANCE AND MAGNETO-DIELECTRIC SUBSTRATES FOR SMALL ANTENNA DESIGNS . . . . . . . . . . . . . . . . . . . . . . . . . 34-15 34-1 Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.

Low Profile Antenna Performance Enhancement Utilizing Engineered Electromagnetic Materials 34-2

CHAPTER THIRTY-FOUR

34.1 INTRODUCTION The growing number of wireless communication systems continuously challenges antenna engineers to create novel antenna structures and improve existing antenna designs. Due to the advancements in computational electromagnetics and fabrication technologies, antenna engineers nowadays are capable of exploiting complex engineered electromagnetic materials in antenna designs.1–4 Surveying the literature, one observes that various engineered materials have been investigated and numerous antenna applications have been proposed. For example, Figure 34-1 illustrates some representative engineered electromagnetic materials with unique electromagnetic properties.5 Frequency selective surfaces (FSS) are widely used in radome and reflector antenna designs as wave filters.6,7 Double negative (DNG) material refers to those materials with effective negative permittivity and permeability, which results in properties such as lefthanded (LH) wave propagation and negative index of refraction (NIR).8–10 Periodic structures that prohibit the propagation of electromagnetic waves in a certain frequency band for certain arrival angles and polarization senses are classified as electromagnetic band gap (EBG) structures.11–13 Another important category consists of ground planes that exhibit unique reflection characteristics other than conventional PEC, known as complex artificial ground planes.14,15 This chapter summarizes several typical engineered electromagnetic materials and illustrates their applications in antenna engineering. It is demonstrated that they not only improve the performance of conventional antennas such as gain, bandwidth, and efficiency, but also lead to novel radiator concepts and structures like surface wave antennas and reconfigurable antennas.

Multilayer Dielectric Structures

Double Concentric Square Loop FSS

High Q Dipole FSS

FSS Artificial Ground Plane

Dichroic Plate FSS

Sierpinski Fractal FSS

Double Negative Material

Multilayer Tripod EBG

Mushroom-like EBG

PolarizationDependent EBG

RIS Metasubstrate

Embedded-Circuit Metamaterial

FIGURE 34-1 Different classes of engineered electromagnetic materials



LOW PROFILE ANTENNA PERFORMANCE ENHANCEMENT

34.2 CHARACTERIZATIONS AND DESIGNS OF ELECTROMAGNETIC BAND GAP (EBG) STRUCTURES Among various engineered electromagnetic materials, mushroom-like EBG structure has attracted increasing attention.12 It consists of four parts: a ground plane, a dielectric substrate, periodic metal patches, and connecting vias, as shown in Figure 34-2. This structure is easy to fabricate using the printed circuit board (PCB) technique. Frequency Band Gap and In-phase Reflection Coefficient The mushroom-like EBG structure exhibits distinctive electromagnetic properties with respect to incident electromagnetic waves: ●

●

+ ≥ the EBG structure shows a frequency band gap through which the surface wave cannot propagate for any incident angles and any polarization states, resulting in an ideal isolator for electromagnetic waves. When the incident wave is a plane wave (kx2 + ky2 < k02), the EBG structure has an in-phase reflection coefficient of +1 at a certain frequency, which resembles an ideal perfect magnetic conductor (PMC) that does not exist in nature. When the incident wave is a surface wave ( kx2

ky2

k02 ),

y x

Top View

z x

Cross View

FIGURE 34-2 Geometry of a mushroom-like electromagnetic band gap (EBG) structure (after Yang et al62 © Microwave Optical and Technology Letters 2005)

In the above equations, kx and ky denote the wave numbers in the horizontal directions, while k0 is the free-space wave number. The operation mechanism of this EBG structure can be explained by an LC filter array: the inductor L results from the current flowing through the vias, and the capacitor C is due to the gap effect between the adjacent patches. Some empirical formulas for the inductor L and the capacitor C are presented in Rahman and Stuchly.16 To accurately characterize the frequency band gap property and the in-phase reflection coefficient, the finite difference time domain (FDTD) method17–20 is used to analyze the EBG structure. A single unit of the EBG structure is simulated with periodic boundary conditions (PBC) incorporated on four sides to model an infinite periodic structure. The dimensions of the analyzed EBG structure are W = 0.10l, g = 0.02l, h = 0.04l, er = 2.94

(34-1)

where W is the width of the square patch, g is the gap width, h is the substrate thickness, and er is the dielectric constant of the substrate. The vias’ radius in the EBG structure is 0.005l. The free-space wavelength at 4 GHz, l = 75 mm, is used as a reference length to define the physical dimensions of the EBG structure. It is worthwhile to point out that the periodicity of the EBG structure (0.12l) is much smaller than the wavelength. In addition, these dimensions are readily scaled to other frequencies of interest for different applications. Figure 34-3a shows the w -b diagram of the EBG structure, where the vertical axis is the frequency and the horizontal axis represents the values of the horizontal wave numbers (kx, ky) in the Brillion zone. Each point in the dispersion diagram represents a certain surface wave mode. It is observed that in the frequency range from 3.5 GHz to 5.9 GHz,



CHAPTER THIRTY-FOUR

10 9

Light line

200

Light line

150

M

Frequency (GHz)

7

Γ

Reflection phase (Degree)

8

X

6 5

Band Gap 4 3 2

0 Γ

Second mode X

Wavenumber

(a)

M

50 0 −50 −100 −150

First mode 1

100

−200 Γ

0

2

4

6 Freq. (GHz)

8

10

12

(b)

FIGURE 34-3 Electromagnetic properties of an EBG structure characterized using FDTD method with periodic boundary conditions (PBC): (a) the w -b diagram and (after Yang et al62 Microwave Optical and Technology Letters 2005) (b) the reflection phase. The EBG structure exhibits a frequency band gap for surface waves and an in-phase reflection coefficient for plane-wave incidence.

no surface waves can exist regardless of propagation angles or polarization states. Thus, this frequency region is defined as a surface wave band gap of the EBG structure. Figure 34-3b shows the reflection phase curve for a normally incident plane wave. The reflection phase of a surface is defined as the phase of the reflected E field normalized to the phase of the incident E field at the reflecting surface. It is known that a perfect electric conductor (PEC) has an 180° reflection phase and a perfect magnetic conductor (PMC) has a 0° reflection phase. In contrast, the reflection phase of the EBG surface decreases continuously from 180° to −180° as frequency increases. For example, the EBG surface exhibits a 90° reflection phase around 4.6 GHz and a 0° reflection phase around 5.8 GHz. It is important to note that the reflection phase varies with incident angles and polarization states. Parametric Studies The electromagnetic properties of the EBG structure are determined by its physical dimensions, namely, patch width W, gap width g, substrate permittivity er, and substrate thickness h. Hence, in-depth studies on these parameters are required to develop engineering design guidelines for the EBG structure.21 The reference EBG parameters are listed below: W = 0.12l12 GHz, g = 0.02l12 GHz, h = 0.04l12 GHz, er = 2.20

(34-2)

The vias’ radius is 0.005l12 GHz. Figure 34-4 presents the effects of the preceding four parameters on the reflection phase of the EBG structure. During each study, only the parameter of interest changes, while the other three parameters remain the same as Eq. 34-2. Two important properties on the reflection phase curve are concerned: the resonant frequency where the reflection phase equals to zero and the slope of the curve that corresponds to the frequency bandwidth of the EBG structure. Following are observations that can be obtained from Figure 34-4: ●

●

When the patch width W increases, the resonant frequency decreases and the slope of the curve becomes steep. When the gap width g increases, the resonant frequency increases and the slope of the curve becomes flat.



LOW PROFILE ANTENNA PERFORMANCE ENHANCEMENT 200

200 150

0.04

100



150

0.08

50

0.12

0 0.16

−50

0.20

−100

0.12

50

0.08 0.04

0 −50

0.02 0.01

−100 −150

−150 −200

100

5

10

15 20 Freq (GHz)

25

−200

30

5

10

(a)

30

20

25

200

150

150

0.01

100



25

(b)

200

0.02

50

0.04

0

0.06

−50

0.08

−100 −150 −200

15 20 Freq (GHz)

1.0

100

2.2

50 3.27

0 −50

6.0

−100

9.2 10.2

−150

5

10

15 20 Freq (GHz)

25

30

(c)

−200

0

5

10 15 Freq (GHz)

(d)

FIGURE 34-4 Parametric studies on the EBG dimensions: (a) patch width W effect, (b) gap width g effect, (c) substrate thickness h effect, and (d) substrate permittivity er effect

●

●

When the substrate thickness h increases, the resonant frequency decreases but the slope of the curve becomes flat. When the dielectric constant er increases, the resonant frequency decreases and the slope of the curve becomes steep.

Although the preceding observations focus on the reflection phase feature of the EBG structure, it is noticed that similar parameter effects also apply in the surface wave band gap property. Polarization-Dependent EBG Designs The aforementioned mushroom-like EBG structure has a symmetric square unit cell so that its reflection phase for the normal incidence is independent of the polarization states. When the unit geometry is modified, EBG structures with polarization-dependent reflection phases can be obtained.22–24 Figure 34-5a shows a polarization-dependent EBG design using rectangular patch units. The patch length L is 0.24l3 GHz and the width W is 0.16l3 GHz. The gap width is 0.02l3 GHz and the vias’ radius is 0.0025l3 GHz. The substrate thickness is 0.04l3 GHz and the dielectric constant is 2.20. As a reference, a square patch EBG surface (0.16l3 GHz × 0.16l3 GHz) is also studied. Due to different values of L and W, the reflection phase of the EBG surface Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.


CHAPTER THIRTY-FOUR

200

Square EBG Rect. EBG, Y pol Rect. EBG, X pol


150

Y X

Top view

100 50 0

−50 −100 −150 −200

Cross view

2

2.5

(a)

3 3.5 Freq. (GHz)

4

4.5

(b)

FIGURE 34-5 (a) A rectangular-patch EBG surface and (b) the reflection phases of the EBG structure with respect to different polarizations of the incident wave (after Yang and Rahmat-Samii23 © Microwave Opt. Technol. Lett. 2004)

becomes dependent on the x- or y-polarization state of the incident plane wave, as depicted in Figure 34-5b. When the incident plane wave is y-polarized, the rectangular patch EBG surface has the same reflection phase as the square patch EBG surface because the patch widths are the same. For the x-polarized incident plane wave, the patch length L plays a dominant role in determining the reflection phase. Since the length L is longer than the width W, the reflection phase curve shifts down to lower frequencies. It is noticed that near 3 GHz, the EBG surface shows a −90° reflection phase for the x-polarized wave and a + 90° reflection phase for the y-polarized wave. Besides the rectangular patch unit, various other approaches have also been used to realize the polarization-dependent feature. For example, Figure 34-6 shows two alternative designs:

Y

Y X

X

Top view

Top view

Cross view

Cross view

(a)

(b)

FIGURE 34-6 Polarization-dependent EBG structures: (a) slot-loaded design and (b) offset vias design (after Yang and Rahmat-Samii23 © Microwave Opt. Technol. Lett. 2004)




a slot-loaded EBG and an EBG with offset vias. In Figure 34-6a, a pair of y-oriented slots are symmetrically incorporated into the patch of a square EBG surface. The slots affect electric currents flowing along the x direction, resulting in a longer current path. Thus, the reflection phase of the x-polarized wave decreases to a lower frequency. In contrast, the reflection phase of the y-polarized wave remains the same as the square patch EBG surface because the slots have little effect on the electric currents flowing along the y direction. In Figure 34-6b the vias are offset along the x direction while they are still centered along the y direction. Therefore, the reflection phase for the y-polarized wave remains unchanged whereas the reflection phase for the x-polarized wave varies with the vias’ position. When the vias are located in the center of the patch, only one resonant frequency is observed. When the vias are offset, dual resonance behaviors are observed with one resonant frequency higher than the original frequency and the other lower. The different frequencies correspond to the different widths of the left and right regions of the patch with respect to the vias. The left part is narrower and thus corresponds to the higher resonant frequency. The right part is wider and is related to the lower resonant frequency.

34.3 LOW PROFILE WIRE ANTENNAS ON EBG GROUND PLANE Because of the attractive electromagnetic properties, the EBG structure has been applied in various antenna applications. In this section, the in-phase reflection coefficient feature is exploited to increase the radiation efficiency of low profile wire antennas.21,25–27 The low profile design usually refers to an antenna structure whose overall height is less than one-tenth of the operating wavelength, which is desirable in many mobile communication systems. Dipole Antenna on an EBG Ground Plane To appreciate the advantages of using an EBG ground plane, we start with a dipole antenna, the simplest radiating structure. As shown in Figure 34-7a, a dipole antenna is horizontally positioned above an EBG ground plane to obtain a low profile configuration.28 The dipole length is 0.40l12 GHz and its radius is 0.005l12 GHz. The height of the dipole over the top surface of the EBG ground plane is only 0.02l12 GHz. A finite EBG ground plane with a 1l12 GHz × 1l12 GHz size is used in the analysis, and the dimensions of the EBG structure are the same as those given in Eq. 34-2. 0 PEC −5

PMC

S11 (dB)

−10 −15

EBG

−20 −25 −30

10

(a)

12

14 Freq (GHz)

16

18

(b)

FIGURE 34-7 Dipole antenna near an EBG ground plane: (a) geometry and (b) FDTD simulated return loss (after Yang and Rahmat-Samii21© IEEE 2003)



CHAPTER THIRTY-FOUR

Figure 34-7b shows the FDTD simulated return loss of the dipole antenna. For comparison purposes, the performances of dipole antennas on PEC and PMC ground planes are also simulated and plotted. When the dipole is located above the PEC ground plane, the return loss is only −3.5 dB. The PEC surface has a 180° reflection phase so that the direction of the image current is opposite to that of the original dipole. The radiations from the image current and the original dipole cancel each other, resulting in a very poor return loss. When the PMC surface, which has a reflection phase of 0°, is used as the ground plane, the dipole has a return loss of −7.2 dB. The return loss is improved; however, it suffers from a strong mutual coupling between the image current and the dipole due to their close proximity.21,29 The best return loss of −27 dB is achieved by the dipole antenna over the EBG ground plane. The reflection phase of the EBG varies with frequency from 180° to −180°. In a certain frequency range, the EBG surface successfully serves as the ground plane for a low profile dipole, resulting in a high radiation efficiency. It is clear from this comparison that the EBG surface is a good ground plane candidate for low profile wire antenna designs. A further question is how to identify the operational frequency band of an EBG ground plane. To this end, the parameters of the EBG surface are fixed and the length of the dipole is varied to resonate at different frequencies. By observing the return loss value and radiation patterns of the dipole at different frequencies, one can find a useful operational frequency band of the EBG ground plane for low profile wire antenna designs. Figure 34-8a shows the return loss results of a dipole with its length varying from 0.26 l12 GHz to 0.60 l12 GHz. It is observed that the dipole shows a return loss better than −10 dB in a frequency range from 11.5 to 16.6 GHz. From a computational efficiency viewpoint, it would be interesting to know if one could directly use the reflection phase curve of the EBG structure to identify the operational frequency band. Thus, the reflection phase of the EBG surface is evaluated, and the result is presented in Figure 34-8b. In contrast to the 180° reflection phase of a PEC surface or the 0° reflection phase of a PMC surface, if one chooses the 90° ± 45° reflection phases as the criterion for the EBG ground plane, a frequency region from 11.3 to 16 GHz is identified, which is close to the frequency region obtained in the dipole model. It is revealed from this comparison that the operational frequency band of an EBG ground plane is the frequency region inside which the EBG surface shows a quadratic reflection phase (90° ± 45°). This quadratic reflection phase criterion has been further demonstrated by various numerical and experimental results.21 0

200 Reflection phase (Degree)

−5

S11 (dB)

−10 −15

0.60

−20 −25 −30

0.26

0.54 0.32 0.48

0.42

0.36

−35 −40

10

12

14 Freq (GHz)

(a)

16

18

150 100 50 0 −50

10

12

14 Freq (GHz)

16

18

(b)

FIGURE 34-8 Identification of the operational frequency band of an EBG ground plane: (a) return loss of a nearby dipole with length varying from 0.26 to 0.60 l12 GHz and (b) reflection phase of the EBG surface (after Yang and Rahmat-Samii21 © IEEE 2003)



10

10

0

0

−10

−10

Directivity (dB)

Directivity (dB)


−20 −30 −40 −50

−150

−100

−50

0 θ (Degree)

0.48 , f=12GHz 0.36 , f=13.6GHz 0.32 , f=15.3GHz 50 100 150

(a)

0.48 , f=12GHz 0.36 , f=13.6GHz 0.32 , f=15.3GHz

−20 −30 −40 −50

−150

−100

−50

0 50 θ (Degree)

100

150

(b)

FIGURE 34-9 Radiation patterns of three dipoles at their resonant frequencies: (a) E-plane patterns and (b) H-plane patterns (after Yang and Rahmat-Samii21 © IEEE 2003)

The radiation patterns of dipole antennas above the EBG ground plane are also calculated to verify the radiation efficiency. Figure 34-9 displays both the E- and H-plane patterns of three dipole antennas at their resonant frequencies: (1) 0.48 l12 GHz dipole resonating at 12 GHz, (2) 0.36 l12 GHz dipole resonating at 13.6 GHz, and (3) 0.32 l12 GHz dipole resonating at 15.3 GHz. It is observed that all three dipoles radiate efficiently with directivities around 8 dB. Circularly Polarized Designs Circularly polarized (CP) antennas are desired in many communication systems such as the Global Positioning System (GPS) and satellite links. The EBG ground plane is also implemented in low profile CP wire antenna designs. The first approach is to replace the traditional PEC ground plane of a curl antenna with an EBG ground plane, as shown in Figure 34-10a. The circular polarization is generated by the traveling current along the curl30 and the EBG ground plane helps to improve the radiation efficiency of the curl in a low profile configuration.31,32 It is noticed that the overall antenna height (0.07l)

(a)

(b)

FIGURE 34-10 Photographs of low profile CP wire antennas on EBG ground plane: (a) a curl on a square patch EBG ground plane (after Yang and Rahmat-Samii31 © Microwave Opt. Technol. Lett. 2001) and (b) a dipole on a rectangular patch EBG ground plane (after Yang and Rahmat-Samii33 © IEEE 2005)



CHAPTER THIRTY-FOUR

is much smaller than that of a conventional curl design on a PEC ground plane (0.25l). Figure 34-10b presents the second approach for low profile CP wire antenna design where a linearly polarized dipole is used instead of the CP curl.33 The circular polarization pattern of the overall antenna is realized by the superposition of the directly radiating wave from the dipole and the reflected wave from the EBG ground plane. A polarization-dependent EBG ground plane is carefully designed and implemented here so that the reflected wave has a perpendicular polarization direction to the directly radiating wave as well as a 90° phase shift.23 The function of the polarization-dependent EBG ground plane is similar to a meandering line polarizer.6 Reconfigurable Antenna with Pattern Diversity Another interesting application of the EBG ground plane is to build a low profile reconfigurable wire antenna with radiation pattern diversity.34 Reconfigurable antennas are desirable in modern wireless communication systems35–37 because they can provide more functionalities than ordinary antennas by reconfiguring their radiation performance, such as the operating frequencies,38,39 polarizations,40–42 and radiation patterns.43–47 Figure 34-11a shows the geometry of a reconfigurable wire antenna design, where a feeding probe is connected to two metal strips through two switches. When the left switch is ON and the right switch is OFF, the probe has an electrical connection to the left strip, resulting in a bent monopole oriented along the −x direction. When the left switch is OFF and the right switch is ON, the probe has an electrical connection to the right strip, resulting in a bent monopole oriented along the +x direction. As a consequence, the direction of the antenna beam can be switched in space and the diversity in the radiation pattern is realized. A reconfigurable antenna prototype is built to demonstrate the operational mechanism, and the measured diversity patterns are shown in Figure 34-11b. It is observed that the antenna beam is switched between ±26°.

0° −30°

30°

0 dB −5 dB

°

°

−60

60 −10 dB −15 dB

°

90°

−90

y x

Top View Switches

z x

Cross View (a)

Wire EBG surface

−120°

x oriented +x oriented −150°

°

120

150° 180°

(b)

FIGURE 34-11 A low profile reconfigurable wire antenna on the EBG ground plane with radiation pattern diversity: (a) antenna geometry and (b) measured radiation patterns




34.4 PATCH ANTENNAS WITH ENHANCED PERFORMANCE USING EBG STRUCTURES The EBG structure has two main features, and the in-phase reflection coefficient feature is used in the preceding section to improve the radiation efficiency of low profile wire antennas. This section discusses how the surface wave band gap feature is used to enhance the performance of microstrip antennas and arrays. Microstrip antennas are widely used in wireless communications due to their numerous advantages, including low profile configuration, light weight, low fabrication cost, and conformability with RF circuitry. In typical microstrip antenna designs, unwanted surface waves are excited in the substrate. These surface waves degrade the antenna performance, such as decreasing the antenna gain, increasing the back lobe, and increasing the mutual coupling. This problem becomes more severe when high dielectric constant substrates are used to design miniaturized microstrip antennas. Various methods have been proposed to solve this problem,48–51 and the EBG structures demonstrate excellent capability in reducing the surface waves.52–61 In this section, the mushroom-like EBG structure is integrated into microstrip antennas and arrays, and the applications of metasubstrate in microstrip antennas will be discussed in Section 34.6. Patch Antenna Element The EBG structure is first integrated with a microstrip patch antenna element to increase the antenna gain and suppress the back radiation.54 The basic idea is to properly design the EBG structure such that the resonant frequency of the patch antenna falls inside the band gap of the EBG structure, hence inhibiting the surface waves. Figure 34-12a shows a patch antenna surrounded by four rows of EBG cells. The patch antenna is designed to resonant at 5.8 GHz and is fabricated on a 2.54-mm-thick RT/duroid 6010 (er = 10.2) substrate with a finite ground plane of 52 × 52 mm (1l × 1l at 5.8 GHz). The EBG patch size is 2.5 × 2.5 mm and the gap width between adjacent patches is 0.5 mm. It is noticed that the EBG structure is compact in this design because a thick substrate with a high dielectric constant is used. For comparison purposes, a conventional patch antenna on the same substrate but without EBG cells is built as a reference. It is noticed that both antennas have similar return loss,

0

EBG

Pattern (dB)

−5

No EBG

−10 −15 −20 −25 −30

(a)

−150

−100

−50

0 50 Angle (Degree)

100

150

(b)

FIGURE 34-12 (a) Photo of a microstrip patch antenna surrounded by an EBG structure and (b) measured E-plane radiation patterns for patch antennas with and without EBG structure



CHAPTER THIRTY-FOUR

0 −5

S (dB)

−10

S11

−15

No EBG With EBG

−20

S12

−25 −30

(a)

5.2

5.4

5.6 5.8 Freq (GHz)

6

6.2

6.4

(b)

FIGURE 34-13 (a) Photos of a normal microstrip antenna array (above) and a design that integrates the EBG structure into the array (below); (b) measured scattering coefficients of the microstrip antenna arrays (after Yang and Rahmat-Samii55 © IEEE 2003)

but the EBG case shows an improved radiation pattern, as plotted in Figure 34-12b. With the surrounding EBG cells, the antenna gain is successfully increased by 3.2 dB while the back lobe is effectively reduced by more than 15 dB. Patch Antenna Array The EBG structure also can be integrated with a microstrip antenna array to reduce the mutual coupling between antenna elements.55 As shown in Figure 34-13a, two pairs of microstrip antennas with and without the EBG structure are fabricated on 1.92-mm-thick Roger RT/duroid 6010 substrates. The antenna’s size is 6.8 × 5 mm, and they are fabricated on a finite ground plane of 100 × 50 mm with the patch distance of 38.5 mm (0.75 l5.8 GHz). The EBG patch width is 3 mm and the gap width is 0.5 mm. The measured results are shown in Figure 34-13b, where both antennas resonate at 5.86 GHz with a return loss better than −10 dB. For the antenna array without the EBG structure, the mutual coupling at 5.86 GHz is as high as −16.8 dB. In contrast, the mutual coupling of the antennas with the EBG structure is only −24.6 dB. An 8-dB reduction of mutual coupling is achieved at the resonant frequency because of the effective suppression of surface waves by the EBG structure. This low mutual coupling design is potentially useful for various phase array applications, such as eliminating blind angles in radar systems.

34.5 SURFACE WAVE ANTENNAS WITH A MONOPOLE-LIKE RADIATION PATTERN The mushroom-like EBG structure exhibits both a surface wave band gap feature and an inphase reflection coefficient. However, these two characteristics are not necessarily associated with each other.62 For example, this section discusses another engineered electromagnetic material that only has an in-phase reflection coefficient but has no band gap for surface waves. This new engineered material is used to design surface wave antennas that realize a monopolelike radiation pattern with a low profile configuration. Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.



Dipole-Fed Surface Wave Antenna Figure 34-14a shows the geometry of a horizontal dipole antenna above a patch-loaded grounded slab. The dipole is simply fed by a 50-Ω coaxial cable: one arm of the dipole is connected to the center probe of the cable, and the other arm is connected to the outside conductor of the cable, which is soldered to the lower PEC of the artificial surface. Compared to Figure 34-7a, this design uses a patch-loaded grounded slab as the ground plane, which is similar to a mushroom-like EBG structure except that the vertical vias are all removed. It is revealed that removing the vertical vias has little effect on the in-phase reflection feature for a normally incident plane wave; however, it completely changes the surface wave characteristic of the engineered ground plane. Figure 34-14b presents the dispersion diagram of the patch-loaded grounded slab. The dimensions of this artificial ground plane are the same as those given in Eq. 34-1. It is clear that when the vertical vias are removed, the surface wave band gap disappears. Therefore, the surface waves can exist over the entire frequency band. Due to the different properties of the engineered ground plane, the radiation performance of the antenna is vastly different from the ones discussed in Section 34.3. The return loss of the antenna with various dipole lengths is shown in Figure 34-15a. It is noticed that the low profile antenna also obtains a good return loss in a frequency region where the artificial surface exhibits a quadratic reflection phase. When the dipole length is 0.26l4 GHz, the antenna achieves a good return loss of −30 dB at 4.05 GHz with a 7.1 percent impedance bandwidth. It is worthwhile to point out that the length of the dipole is much smaller than the half wavelength at the operating frequency. In contrast, if a dipole is located near an EBG ground plane and resonates at the same frequency, the length of the dipole is close to a half wavelength. Figure 34-15b shows the radiation patterns of the antenna at the resonant frequency of 4.05 GHz. Several interesting observations can be made from this figure. First, the antenna shows a small radiation power in the broadside direction (q = 0°). The main beam of this antenna points to the q = 50° direction with a directivity of 5 dBi. This is different from the

10 9

Light line

Light line M

8

Frequency (GHz)

7

y x

Top View Dipole

z

Γ

X

6 5

No Band Gap 4 3 2

First mode 1

x

Artifical ground planve

Cross View (a)

0 Γ

Second mode X

Wavenumber

M

Γ

(b)

FIGURE 34-14 (a) A dipole antenna near a patch-loaded grounded slab and (b) the dispersion diagram of the patch-loaded grounded slab. No surface wave band gap exists for this engineered structure (after Yang et al62 © Microwave Optical and Technology Letters 2005).



CHAPTER THIRTY-FOUR 10

0

0.06 λ

−5

0

−15

0.10 λ 0.38 λ

−20

0.14 λ

0.34 λ

−25

0.18 λ

0.30 λ

−10

−20 φ=0, E

0.22 λ

−30

θ

φ=90, E

−30

0.26 λ

−35 −40

Directivity (dB)

S11 (dB)

−10

φ=0, Eφ

θ

φ=90, E 3

4

Freq. (GHz)

5

(a)

6

−40

φ

−150

−100

−50

0 50 θ (Degree)

100

150

(b)

FIGURE 34-15 Radiation characteristics of the dipole-fed surface wave antenna: (a) FDTD simulated return loss of the antenna with different dipole lengths (after Yang et al62 © Microwave Optical and Technology Letters 2005) and (b) radiation patterns of the antenna at 4.05 GHz with a dipole length of 0.26l

dipole antenna on an EBG ground plane whose main beam points to the broadside direction, as shown in Figure 34-9. Second, Eq is the copolarized field in both the xz (f = 0°) and yz planes (f = 90°). Thus, both the xz and yz planes are E planes. In contrast, if the dipole is near an EBG ground plane, the xz plane is the E plane but the yz plane is the H plane. In summary, this antenna has entirely different radiation characteristics compared to a horizontal dipole on an EBG ground plane. The radiation characteristics in Figure 34-15 can be explained by the propagation and diffraction of surface waves in the engineered ground plane. When a dipole is positioned near a patch-loaded grounded slab, strong surface waves are excited to propagate along the ground plane. The dipole works more like a transducer than a radiator. Therefore, the optimal length of the dipole is not necessarily equal to a half wavelength. It is also noticed that the surface waves are dominated by the TMz mode and the electric field is vertically polarized. When the surface waves diffract at the edge of the ground plane, a monopolelike radiation pattern is generated. For example, since the diffraction at the edge is hard boundary diffraction, two diffracted rays from opposite edges will cancel in the broadside direction, resulting in a radiation null. The hard diffraction also determines the polarization of the radiating field along the q direction. Therefore, this antenna is properly identified as a surface wave antenna (SWA)63–67 because of this operational mechanism. The attractive feature of this SWA design is the low profile configuration. The height of the horizontal dipole over the artificial ground plane is only 0.02l4 GHz whereas the height of a typical monopole antenna is 0.22l4 GHz. The dipole height is less than 10 percent of the conventional monopole antenna. Therefore, this low profile SWA design has a promising potential in mobile communication systems such as vehicle radio systems. Patch-Fed Surface Wave Antenna A deficiency in the previous design is the relatively high cross-polarization attributed to the direct radiating field from the dipole. To solve this problem and obtain a better monopole-like radiation pattern, a patch-fed surface wave antenna (PFSWA) is sketched in Figure 34-16a. The same patch-loaded grounded slab is used in the antenna design while a circular patch is inserted in the middle of the substrate to excite surface waves.68 Investigation shows that the energy can be effectively coupled from the cavity fields underneath the circular patch to the surface waves propagating along the engineered ground plane. Hence, good radiation




W

g

10

φ=0, co-pol φ=0, X-pol φ=45, co-pol φ=45, X-pol

5 0 −5

x

z

Gain (dB)

y

−15

Top View

−20

Periodic loading

εr x

Patch feed Coaxial probe

Cross View (a)

−10

−25 −30

−150

−100

−50

0 θ (Degree)

50

100

150

(b)

FIGURE 34-16 A low profile patch-fed SWA with a monopole-like radiation pattern: (a) antenna geometry and (b) measured radiation patterns

efficiency can be obtained. Parametric studies reveal that 8 × 8 periodic patches are enough to launch the surface waves. In order to get a symmetric diffraction pattern, the ground plane is truncated into a circular shape. Figure 34-16b shows measured radiation patterns of a PFSWA prototype, which is very similar to a typical monopole pattern. Both the xz plane (f = 0°) and diagonal plane (f = 45°) patterns are presented, which are almost identical to each other. It has a deep null in the broadside direction and the antenna beam is at q = 47° direction with a gain of 5.6 dBi. The co-polarization is along the q direction, and the cross-polarization is −25 dB lower than the co-polarization in the front side. A noticeable cross-polarization is observed in the backside of the antenna due to diffractions from the supporting posts and feeding cable.

34.6 IMPEDANCE AND MAGNETO-DIELECTRIC SUBSTRATES FOR SMALL ANTENNA DESIGNS The growing number of wireless applications has presented RF engineers with a continuing demand for low-cost, power-efficient, and small-size system designs. Depending on the application at hand and required system characteristics, such as data rate, environment, range, etc., the system parameters, such as operating frequency, transmitter power, and modulation scheme, may vary widely. However, independent of the application, compactness, wide bandwidth, high efficiency, ease of fabrication and integration, and low cost are always sought in wireless systems. One of the most important components of wireless systems is their antenna. The substrate of the antenna has a significant role in successfully providing the above engineering requirements. In this section, three novel substrates—namely, reactive impedance, layered magneto-dielectric, and embedded-circuit metasubstrates— are presented, and their advantages and challenges in the design of small antennas with improved bandwidth performance are addressed.



CHAPTER THIRTY-FOUR

Reactive Impedance Surface Design of antenna elements with significant front-to-back radiation ratio is usually accomplished through the use of metal-backed substrates.69 However, printed antennas on metalbacked substrates have limited bandwidth and efficiency. As demonstrated earlier, this problem stems from the fact that the radiated field from the image of the antenna’s electric current, which is placed in close proximity and parallel to a PEC, tends to cancel out the radiated field from the antenna current itself. In this case, matching the antenna input impedance is rather difficult, and if a matching condition can be achieved, it would be over a relatively narrow bandwidth. To circumvent this difficulty, a reactive impedance surface (RIS), as proposed in Mosallaei and Sarabandi,29 can be used for the antenna substrate. The RIS has three major features: ●

●

●

It provides a total reflection power that enhances the antenna front-to-back ratio. The image of a point source located above the RIS is a spatially distributed current element that has the minimum interaction with the point source. This has the significant advantage of reducing the mutual coupling between the antenna and its substrate, resulting in the impedance bandwidth enhancement. Mutual coupling F between an infinitesimal dipole located at 0.02l0 above the RIS and its image as a function of the normalized impedance n/h0 is demonstrated in Figure 34-17. It shows the minimum coupling occurs for a surface impedance of n = 0.33h0, an impedance between the PEC and PMC behaviors, as pointed out before. RIS has the ability to store magnetic (or electric) energy that can be properly used to compensate for the near-field electric (or magnetic) energy of the radiating structure resulting in the antenna size reduction.

To realize an RIS with surface impedance h = jn, a periodic configuration of smallsize patch elements printed on a dielectric material backed by a PEC can be used (see Figure 34-18a). The gap capacitances between the patches provide an equivalent capacitor that is in parallel with an inductor obtained from the PEC after the distance d through the dielectric material. A parallel LC circuit offers a reactive impedance behavior satisfying the required condition. The FDTD is applied to demonstrate the surface impedance property of the RIS and the results for normal and oblique incident waves are plotted in Figure 34-18b. Since the dielectric constant of substrate is relatively large, 1

200 PMC

PEC

150 Phase F (deg.)

Amplitude F

0.8

0.6

0.4

50

0.2

0 −2 10

100

10−1

ν/ηο

(a)

100

101

0 −2 10

10−1

ν/ηο

100

101

(b)

FIGURE 34-17 Mutual coupling between an infinitesimal dipole, located above the RIS, and its image: (a) amplitude and (b) phase (after Mosallaei and Sarabandi29 © IEEE 2004)




o

Λz

5

PEC

Normalized Surface Impedance ν/η (η=jν)

a

ε Λy d

L

C

ε r = 25.0, µ r = 1.0, tgδ l = 0.001 Λ y = Λ z = 1.20 cm, d = 0.40 cm

3 2 1 0 −1 −2 −3 −4 −5

a = 0.80 cm

Normal Oblique, νTE Oblique, νTM

4

1

1.2

1.4

1.6

(a)

1.8 2 2.2 2.4 Frequency (GHz)

2.6

2.8

3

(b)

FIGURE 34-18 An RIS constructed from the periodic array of patches printed on a dielectric material backed by PEC: (a) the geometry and its equivalent circuit model and (b) normalized surface impedance (after Mosallaei and Sarabandi29 © IEEE 2004)

according to Snell’s law, the transmission angle is almost independent of the incident angle. As a result, the surface performance is almost insensitive to the incident angle (see Figure 34-18b). It must be mentioned that in the proposed design, no connecting vias between the patch elements and ground plane are constructed. In fact, the vias have nothing to do with the reflection phase and impedance performance of the structure. The major application of vias is when undesirable surface waves can propagate through the substrate, where in this case the vias can be used to successfully forbid the propagation of these waves. This has been clearly demonstrated in Section 34.4. Figure 34-19a shows the geometry of a patch antenna printed on a relatively low dielectric material with er = 6 and located above the RIS substrate. The FDTD is applied to characterize the structure, and the return loss is determined in Figure 34-19b, where a resonant frequency of f0 = 1.86 GHz is obtained. The inductive property of RIS below the resonant frequency is properly combined with the capacitive behavior of patch in this spectral range and successfully shifts down the resonance frequency of antenna, resulting in the antenna ls

5

εr=6.0

0

w

0.40 cm

l

εr=25.0

RIS

l = 1.60 cm, w = 2.0 cm l s = ws = 4.80 cm (a)

−5

ws

|S11| (dB)

0.20 cm

BW

−10 −15 −20

Conventional patch RIS

−25 −30

1.4

1.5

1.6

1.7 1.8 1.9 Frequency (GHz)

2

2.1

2.2

(b)

FIGURE 34-19 (a) Patch antenna printed on RIS and (b) its return loss (after Mosallaei and Sarabandi29 © IEEE 2004)



CHAPTER THIRTY-FOUR

(a)

(b)

FIGURE 34-20 Fabrication of patch on the RIS substrate: (a) magnesium silicate and magnesium calcium titanate blocks metalized and etched and (b) assembled patch antenna over the RIS substrate (after Mosallaei and Sarabandi29 © IEEE 2004)

miniaturization by a factor of about 5(l0/10 antenna size). In addition, since the RIS has a very small coupling with the antenna, a relatively wide bandwidth of about BW = 5.0% is achieved. For comparison, the performance of the patch printed on a dielectric material with er = 21.0 (providing the same resonant frequency) and thickness t = 0.2 cm is also plotted in Figure 34-19b. Patch radiator printed on the conventional substrate provides a very narrow bandwidth of about BW = 0.63%. A thicker conventional dielectric substrate may provide a wider bandwidth; however, the impedance matching is more difficult. Fabrication of the patch antenna printed over the RIS substrate is illustrated in Figure 34-20. Two independent layers are fabricated separately on high-quality ceramic substrates. The dielectric material used for the patch antenna substrate is Trans-Tech D-6 magnesium silicate (er = 6), commonly known as Forsterite. The RIS substrate is made using Trans-Tech MCT-25 magnesium calcium titanate composition (er = 25). Using a thick film silver paste and Trans-Tech’s screen printing process, the array of square patches, corresponding metal backing, and patch are generated. Both substrates are heat treated to form an intimate bond of the silver to the dielectric material. This intermediary stage is diagrammed in Figure 34-20a. Using a two-part, low-loss dielectric adhesive, the substrates are assembled in a fixture to ensure alignment. The return loss of the fabricated patch antenna is measured and plotted in Figure 34-21a. The antenna resonance is found to be at f0 = 1.92 GHz and it exhibits an impedance match with better than −25-dB return loss. The measured relative bandwidth is BW = 6.71%. The radiation patterns are measured in the anechoic chamber of the University of Michigan Radiation Laboratory and are shown in Figure 34-21b. The gain and front-to-back ratio are, respectively, measured to be G = 4.5 dBi and 5.6 dB. This measured gain corresponds to an excellent radiation efficiency of er = 90%. To our knowledge, this is the highest reported gain and bandwidth for such a small planar antenna. Magneto-Dielectric Substrate In this section, the design of a layered magneto-dielectric substrate for antenna miniaturization and bandwidth enhancement is presented. A common approach for antenna miniaturizing is to print the radiator on a high-permittivity substrate. However, the strong electric energy stored inside the high dielectric substrate beneath the patch increases the radiation quality factor of antenna (Q) resulting in a narrow impedance bandwidth. To overcome this problem, one can use a magneto-dielectric material with moderate values of mr and er



LOW PROFILE ANTENNA PERFORMANCE ENHANCEMENT Z 0°

5

30°

30° 5

0

0

60°

60°

-5

−5

|S11| (dB)

-10

X,Y

0°

−10

90°

−15

120°

−20

150°

−25 −30

1.4

120° 150° 180°

1.5

1.6

1.7 1.8 1.9 Frequency (GHz)

2

2.1

E-plane H-plane

2.2

(a)

|E| dB

(b)

FIGURE 34-21 (a) Measured return loss and (b) radiation patterns of the patch antenna fabricated over the RIS substrate (after Mosallaei and Sarabandi29 © IEEE 2004)

that provides the same miniaturization factor n = µr ε r , while a wider bandwidth can be achieved.70 It has been shown by Hansen and Burke71 that the zero-order bandwidth for an antenna over a magneto-dielectric substrate with thickness t can be approximated by BW ≈

96 µr ε r t λ0 2 [4 + 17 µr ε r ]

⋅

(34-3)

Thus for a given miniaturization factor (constant µr ε r ), the antenna bandwidth can be enhanced by increasing mr/er (mr > er). Figure 34-22a shows a patch antenna with size 10 × 8 cm printed on a four-layer dielectric and hexaferrite materials with thickness 2 cm. The size of the ground plane is 20 × 20 cm. The hexaferrite is a Z-type hexagonal material fabricated in Trans-Tech. It has the permittivity of around 16 and permeability shown in Figure 34-22b. The dielectric material has permittivity er = 2.2 and loss tangent tgdl = 0.001. Notice that to offer a low-loss radiation performance, the antenna must operate away from the material resonant frequency. 50

a

L

l = 10 cm, w = 8 cm t = 2 cm (a = 0.5 cm), L = 20 cm (a)

Relative Permeability (µr=µ’rjµ’’r)

w t

Resonance is due to sample

µ’r

l

40

µ’’r

30 Rangeofaccuracy 20

10

0 7 10

108 109 Frequency (Hz)

1010

(b)

FIGURE 34-22 Patch antenna printed on four-layer dielectric and hexaferrite substrate: (a) the geometry and (b) hexaferrite performance (after Mosallaei and Sarabandi70 © IEEE 2004)



CHAPTER THIRTY-FOUR Z 0° 30°

30° 5 0

0

60°

60°

-5 -10

−10

90°

0°

|S11| (dB)

BW

X,Y

−20

120°

−30

−40

250

120° 150°

Dielectric substrate Engineered substrate (εe,µo) 260

270

280

150° 180°

290

Frequency (MHz)

(a)

300

E-plane H-plane

|E| dB

(b)

FIGURE 34-23 The performance of patch antenna printed on the magneto-dielectric substrate: (a) return loss and (b) radiation patterns (after Mosallaei and Sarabandi70 © IEEE 2004)

As discussed above, the larger mr/er provides the larger bandwidth. The layered design allows us to effectively accomplish a large mer/eer, although the hexaferrite itself has almost the same permittivity and permeability below the resonance. Since the patch antenna supports a dominant TEM wave with electric and magnetic fields polarized vertically and horizontally, respectively, they experience the effective dielectric and magnetic materials of eer = 3.84 − j0.004 and mer = 8.61 − j0.16.70 This provides a miniaturization factor greater than 5 with considerably enhanced antenna bandwidth. The FDTD is applied to obtain the return loss and radiation patterns of antenna, as shown in Figure 34-23. The antenna resonance is at f0 = 277 MHz and it provides a wide bandwidth of about BW = 3.2%. The size of the antenna is around 0.09l0 with a miniaturization factor of 5.4. The directivity of the antenna is D0 = 2.9 dB and it has a front-to-back ratio 1.3 dB (ground plane size is 0.18l0 × 0.18l0). The calculated antenna efficiency is about er = 67%. Notice that to achieve the same miniaturization factor utilizing only a dielectric material (mr = 1) one must use er = 23.7. This reduces the bandwidth to about BW = 0.5% as shown in Figure 34-23a. The efficiency in this case for a dielectric loss tangent of 0.001 is about er = 64%. Therefore, utilizing the magneto-dielectric substrate, one can offer a miniaturized wideband planar antenna with relatively high efficiency. The antenna bandwidth for the proposed magneto-dielectric substrate is about six times higher than that of the dielectric substrate. However, the major challenge is to practically achieve a low-loss magnetic material at any frequency of interest (e.g., GHz range). Embedded-Circuit Metasubstrate A metasubstrate constructed from loop circuits embedded in a low dielectric host medium can realize a magneto-dielectric substrate at any frequency of interest.72–75 Figure 34-24a shows the geometry of a periodic configuration of loop circuits. The building block unit cell of the structure is constructed from a metallic loop terminated to a large capacitor, modeling a resonant LC circuit. The loops are very small in size realizing artificial material molecules. They are oriented along the y-direction to offer the desired effective permeability Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.


4.8 mm 0.6 mm

1.8 mm

1.2 mm

εr=30

z 3 mm

y

x

6.6 mm

Effective Constitutive Parameters (εeff,r µeff,r)

LOW PROFILE ANTENNA PERFORMANCE ENHANCEMENT 15 10 5 0 −5

εeff,r µeff,r

−10 −15

2.4 mm

1

1.5 2 Frequency (GHz)

(a)

2.5

(b)

FIGURE 34-24 An embedded-circuit metamaterial with e-m parameters: (a) the geometry and (b) permeability performance (after H. Mosallaei and K. Sarabandi75 © IEEE 2007)

in this direction. The gap coupling between the loops provides the required effective permittivity. The designed embedded-circuit medium has constant effective permittivity of eeff = 7.0e0, and the effective resonant permeability given by Mosallaei and Sarabandi75 of   1 µeff = µ0 1 − ( 0.52 )2 1 − ( 2.0 )2 f 2 (GHz) 

(34-4)

These parameters are plotted in Figure 34-24b. Note that the resonant frequency of permeability, and thus the m, can be properly tuned at any frequency of interest by simply changing the loop capacitors. A metasubstrate is made from the embedded-circuit inclusions, as shown in Figure 34-25a. The operating mode of the patch antenna is a TM wave with electric and magnetic fields polarized along the z- and x-directions, respectively. Hence, the fields experience the effective parameters of the designed metamaterial. The FDTD is applied to characterize the antenna printed on the embedded-circuit metasubstrate, and the result for return loss is shown in Figure 34-25b. A resonant frequency of f0 = 1.74 GHz (miniaturization factor of about 4) and bandwidth of about BW = 1% are determined. The patch antenna printed on a dielectric substrate with er = 13.92 (keeping the same resonant frequency) has bandwidth of about BW = 0.6%. The metasubstrate offers a relatively wider bandwidth. It must be highlighted that the dispersion behavior of the permeability function plays an important role in degrading the impedance bandwidth performance. Both the dielectric and embedded-circuit substrates provide similar radiation patterns, as shown in Figure 34-26. 5 0

2.16 cm

−5

1.80 cm |S11| (dB)

−10

3.84 cm

−15

0.3 cm −20

3.96 cm Dielectric Metasubstrate

−25 −30

1.70

(a)

1.72

1.74 1.76 Frequency (GHz)

1.78

1.80

(b)

FIGURE 34-25 Patch antenna printed on an embedded-circuit metasubstrate: (a) the geometry and (b) return loss (after H. Mosallaei and K. Sarabandi75 © IEEE 2007)



CHAPTER THIRTY-FOUR Z

Z 0°

0° 30°

5

0

60°

0

60°

60°

-10

X,Y

0°

90°

120°

120° 150°

60°

-5

-5

E-plane H-plane

30°

30°

30° 5

150°

-10

X,Y 90°

0°

120°

120° 150°

150°

180°

180°

|E| dB

|E| dB

(a)

(b)

FIGURE 34-26 Radiation patterns of antenna printed on (a) dielectric and (b) metasubstrate (after H. Mosallaei and K. Sarabandi75 © IEEE 2007)

To fabricate the metasubstrate, the first step is to realize an optimal design for the resonant loop circuits. A metallic loop terminated to an interdigitated capacitor seems like a reasonable choice for realizing LC inclusions. However, this is a nonoptimal use of the unit cell area because the interdigitation consumes too much valuable space inside the inductive loop and thereby inhibits optimal coupling to the incident magnetic field. It has been demonstrated in Buell et al76 that a spiral loop is much more preferable because it uses less area to provide equivalent capacitance while simultaneously providing additional inductance, and hence additional permeability. A metasubstrate is fabricated utilizing the spiral inclusions illustrated in Figure 34-27.76 It has a cell size of ∆x = ∆z = 2 cm, ∆y = 3.028 mm (120 mils). The substrate is fabricated on 120-mil-thick Rogers RO-4003 dielectric. The spiral resonators are etched from 1/2 -oz-thick copper (0.017 mm) with a line-width (w) and spacing (s) of 0.127 mm (5 mils). In this design, lx = lx = 16 mm. To reduce substrate mass, 33/64-in-diameter air holes are drilled along the y-axis into the center of each spiral resonator cell. The final substrate mass is reduced by a factor of approximately 1/3, which is significant for a 2-cm-thick substrate. The substrate has a total size of 24 × 24 × 2 cm and a weight of approximately 3.5 lb. It provides an effective permittivity of ereff = 13.13,

FIGURE 34-27 Fabricated metasubstrate utilizing spiral loop circuits (after Buell et al76 © IEEE 2006)


Low Profile Antenna Performance Enhancement Utilizing Engineered Electromagnetic Materials LOW PROFILE ANTENNA PERFORMANCE ENHANCEMENT

(a)

34-23

(b)

FIGURE 34-28 (a) Measured return loss and (b) radiation patterns of the patch antenna printed on the fabricated metasubstrate (after Buell et al76 © IEEE 2006)

and an effective permeability with resonant frequency obtained around 285 MHz. Although here a design for the operation in UHF band is presented, the metamaterial unit cell can be properly redesigned for operation at any frequency of interest. A probe-fed microstrip patch antenna resonant at 250 MHz is built on the metasubstrate (see Figure 34-27). To resonant at 250 MHz, the patch dimensions are found to be 9.3 × 9.3 cm. The measured return loss and radiation patterns are shown in Figure 34-28. A miniaturization factor of about 6.4 and efficiency around er = 20% are measured. A better efficiency can be achieved if a thicker metallization is used. This will improve the magnetic loss tangent by a factor of 2. An additional source of loss is the dielectric loss tangent of the host dielectric material. Decreasing the host material dielectric loss tangent will significantly improve the efficiency of the dielectric medium, but not as strongly as improvements to the metallization would. To estimate the impact of changes in material loss tangent on antenna efficiency, the performance of a 250-MHz patch printed on a substrate with material parameters of er = 9.8, mr = 3.1, and tand = tande = tan dm is simulated in Figure 34-29. Obviously, the better material loss offers the better antenna radiation efficiency.

FIGURE 34-29 Radiation efficiency of a patch antenna at 250 MHz versus substrate material loss tangent on modeled metamaterial substrate (after Buell et al76 © IEEE 2006)



CHAPTER THIRTY-FOUR

In summary, it must be mentioned that although metasurfaces and metasubstrates theoretically present promising advantages; in practice, achieving low-loss wideband metamaterial would be a challenge. It is anticipated that the development of all-dielectric metamaterials to address some of these engineering concerns might help.

REFERENCES 1. J. D. Joannopoulos, R. D. Meade, and J. N. Winn, Photonic Crystals (Princeton, NJ: Princeton Univ. Press, 1995). 2. IEEE Trans. Microwave Theory Tech., Special Issue on Electromagnetic Crystal Structures, Designs, Synthesis, and Applications, vol. 47, no. 11 (November 1999). 3. IEEE Trans. Antennas Propagat., Special Issue on Meta-materials, vol. 51, no. 10 (October 2003). 4. IEEE Trans. Antennas Propagat., Special Issue on Artificial Magnetic Conductors, Soft/Hard Surfaces and other Complex Surfaces, vol. 53, no. 1 (January 2005). 5. Y. Rahmat-Samii and H. Mosallaei, “Electromagnetic Band-gap Structures: Classification, Characterization and Applications,” IEE-ICAP Symposium (April 2001): 560–564. 6. B. A. Munk, Frequency Selective Surfaces, Theory and Design (New York: John Wiley & Sons, Inc., 2000). 7. B. A. Munk, Finite Antenna Arrays and FSS (New York: John Wiley & Sons, Inc. 2003). 8. V. G. Veselago, “The Electrodynamics of Substances with Simultaneous Negative Values of e and m,” Sov. Phys. Usp., vol. 10, no. 4 (1996): 509–514. 9. J. B. Pendry, “Negative Refraction Makes a Perfect Lens,” Phys. Rev. Lett., vol. 85, no. 18 (October 2000): 3966–3969. 10. R. A. Shelby, D. R. Smith, and S. Schultz, “Experimental Verification of a Negative Refractive Index of Refraction,” Science, vol. 292 (April 2002): 77–79. 11. F. -R. Yang, K. -P. Ma, Y. Qian, and T. Itoh, “A Uniplanar Compact Photonic-Bandgap (UC-PBG) Structures and Its Applications for Microwave Circuit,” IEEE Trans. Microwave Theory Tech., vol. 47, no. 8 (August 1999): 1509–1514. 12. D. Sievenpiper, L. Zhang, R. F. J. Broas, N. G. Alexopoulos, and E. Yablonovitch, “HighImpedance Electromagnetic Surfaces with a Forbidden Frequency Band,” IEEE Trans. Microwave Theory Tech., vol. 47, no. 11 (November 1999): 2059–2074. 13. A. S. Barlevy and Y. Rahmat-Samii, “Characterization of Electromagnetic Band-Gaps Composed of Multiple Periodic Tripods with Interconnecting Vias: Concept, Analysis, and Design,” IEEE Trans. Antennas Propagat., vol. 49, no. 3 (March 2001): 343–353. 14. P. -S. Kildal, “Artificial Soft and Hard Surfaces in Electromagnetics,” IEEE Trans. Antennas Propagat., vol. 38, no. 10 (October 1990): 1537–1544. 15. Rahmat-Sammi and F. Yang, “Development of Complex Artificial Ground Planes in Antenna Engineering,” Metamaterials: Physics and Engineering Explorations, N. Engheta and R. Ziolkowski (eds.) (New York: John Wiley & Sons Inc., 2006). 16. M. Rahman and M. A. Stuchly, “Transmission Line-Periodic Circuit Representation of Planar Microwave Photonic Bandgap Structures,” Microwave Opt. Technol. Lett., vol. 30, no. 1 (July 2001): 15–19. 17. A. Taflove and S. Hagness, Computational Electromagnetics: The Finite-Difference Time-Domain Method, 2nd Ed. (Boston: Artech House, 2000). 18. M. A. Jensen and Y. Rahmat-Samii, “Performance Analysis of Antennas for Hand-held Transceiver Using FDTD,” IEEE Trans. Antennas Propagat., vol. 42 (August 1994): 1106–1113. 19. H. Mosallaei and Y. Rahmat-Samii, “Periodic Bandgap and Effective Dielectric Materials in Electromagnetics: Characterization and Applications in Nanocavities and Waveguides,” IEEE Trans. Antennas Propagat., vol. 51, no. 3 (March 2003): 549–563. 20. A. Aminian and Y. Rahmat-Samii, “Spectral FDTD: A Novel Computational Technique for the Analysis of Periodic Structures,” 2004 IEEE APS International Symposium, Monterey, CA, June 20–26.



34-25

21. F. Yang and Y. Rahmat-Samii, “Reflection Phase Characterizations of the EBG Ground Plane for Low Profile Wire Antenna Applications,” IEEE Trans. Antennas Propagat., vol. 51 (October 2003): 2691–2703. 22. F. Yang and Y. Rahmat-Samii, “Polarization Dependent Electromagnetic Band-Gap Surfaces: Characterization, Designs, and Applications,” 2003 IEEE AP-S Digest, vol. 3 (June 2003): 339–342. 23. F. Yang and Y. Rahmat-Samii, “Polarization Dependent Electromagnetic Band Gap (PDEBG) Structures: Designs and Applications,” Microwave Opt. Technol. Lett., vol. 41, no. 6 (July 2004): 439–444. 24. Y. Q. Fu and N. C. Yuan, “Reflection Phase and Frequency Bandgap Characteristics of EBG Structures with Anisotropic Periodicity,” J. of Electromagnetics Waves and Applications, vol. 19, no. 14 (2005): 1897–1905. 25. S. Clavijo, R. E. Diaz, and W. E. McKinzie III, “Design Methodology for Sievenpiper HighImpedance Surfaces: An Artificial Magnetic Conductor for Positive Gain Electrically Small Antennas,” IEEE Transactions on Antennas and Propagation, vol. 51, no. 10 (October 2003): 2678–2690. 26. D. J. Kern et al, “The Design Synthesis of Multiband Artificial Magnetic Conductors Using High Impedance Frequency Selective Surfaces,” IEEE Trans. Antennas Propagat., vol. 53 (January 2005): 8–17. 27. M. F. Abedin and M. Ali, “Effects of EBG Reflection Phase Profiles on the Input Impedance and Bandwidth of Ultrathin Directional Dipoles,” IEEE Trans. Antennas Propagat., vol. 53 (November 2005): 3664–3672. 28. Z. Li and Y. Rahmat-Samii, “PBG, PMC and PEC Ground Planes: A Case Study for Dipole Antenna,” 2000 IEEE APS International Symposium Digest, vol. 4 (July 16–21, 2000): 2258–2261. 29. H. Mosallaei and K. Sarabandi, “Antenna Miniaturization and Bandwidth Enhancement Using a Reactive Impedance Substrate,” IEEE Trans. Antennas Propagat., vol. 52 (September 2004): 2403–2414. 30. H. Nakano, S. Okuzawa, K. Ohishi, H. Mimaki, and J. Yamauchi, “A Curl Antenna,” IEEE Trans. Antennas Propagat., vol. 41, no. 11 (November 1993): 1570–1575. 31. F. Yang and Y. Rahmat-Samii, “A Low Profile Circularly Polarized Curl Antenna over Electromagnetic Band-Gap (EBG) Surface,” Microwave Opt. Technol. Lett., vol. 31, no. 3 (2001): 165–168. 32. H. Nakano et al, “Effects on the Radiation Characteristics of Using a Corrugated Reflector with a Helical Antenna and an Electromagnetic Band-Gap Reflector with a Spiral Antenna,” IEEE Trans. Antennas Propagat., vol. 53 (January 2005): 191–199. 33. F. Yang and Y. Rahmat-Samii, “A Low Profile Single Dipole Antenna Radiating Circularly Polarized Waves,” IEEE Trans. Antennas Propagat., vol. 53, no. 9 (September. 2005): 3083–3086. 34. F. Yang and Y. Rahmat-Samii, “Bent Monopole Antennas on EBG Ground Plane with Reconfigurable Radiation Patterns,” 2004 IEEE APS International Symposium Digest, vol. 2, (June 20–26, 2004): 1819–1822. 35. Y. Qian and T. Itoh, “Progress in Active Integrated Antennas and Their Applications,” IEEE Trans. Microwave Theory Tech., vol. 46, no. 11 (November 1998): 1891–1900. 36. J. T. Bernhard, “Reconfigurable Antennas and Apertures: State-of-the-Art and Future Outlook,” Proc. SPIE Conf. on Smart Electronics, MEMS, BioMEMS, and Nanotechnology, vol. 5055 (March 2003): 1–9. 37. F. Yang and Y. Rahmat-Samii, “Patch Antennas with Switchable Slots (PASS) in Wireless Communications: Concepts, Designs, and Applications,” IEEE Antennas and Propagation Magazine, vol. 47, no. 2 (April 2005): 13–29. 38. D. H. Shaubert, F. G. Farrar, A. Sindoris, and S. T. Hayes, “Microstrip Antennas with Frequency Agility and Polarization Diversity,” IEEE Trans. Antennas Propagat., vol. 29, no. 1 (January 1981): 118–123. 39. F. Yang and Y. Rahmat-Samii, “Patch Antenna with Switchable Slot (PASS): Dual Frequency Operation,” Microwave Optical and Technology Letters, vol. 31, no. 3 (November 2001): 165–168.



CHAPTER THIRTY-FOUR

40. M. Boti, L. Dussopt, and J. M. Laheurte, “Circularly Polarized Antenna with Switchable Polarization Sense,” Electronic Letters, vol. 36, no. 18 (August 2000): 1518–1519. 41. F. Yang and Y. Rahmat-Samii, “A CP Polarization Diversity Patch Antenna Using Switchable Slots,” IEEE Microwave and Wireless Components Letters, vol. 12, no. 3 (March 2002): 96–98. 42. W. E. Mckinzie III and R. Fahr, “A Low Profile Polarization Diversity Antenna Built on an Artificial Magnetic Conductor,” 2002 IEEE AP-s International Symposium Digest, vol. 1 (June 2002): 762–765. 43. K. Chang, M. Li, T. -Y. Yun, and C. T. Rodenbeck, “Novel Low Cost Beam Steering Technique,” IEEE Trans. Antennas Propagat., vol. 50, no. 5 (May 2002): 618–627. 44. K. W. Lee and R. G. Rojas, “Novel Scheme for Design of Adaptive Printed Antenna Element,” IEEE AP-S Digest, vol. 3 (July 2000): 1252–1255. 45. D. Sievenpiper, J. Schaffner, R. Loo, G. Tangonan, S. Ontiveros, and R. Harold, “A Tunable Impedance Surface Performing as a Reconfigurable Beam Steering Reflector,” IEEE Trans. Antennas Propagat., vol. 50, no. 3 (March 2003): 384–390. 46. J. Sor, C. -C. Chang, Y. Qian, and T. Itoh, “A Reconfigurable Leaky-Wave/Patch Microstrip Aperture for Phased-Array Applications,” IEEE Trans. Microwave Theory Tech., vol. 50 (August 2002): 1877–1884. 47. G. H. Huff, J. Feng, S. Zhang, and T. Bernhard, “A Novel Radiation Pattern and Frequency Reconfigurable Single Turn Square Spiral Microstrip Antenna,” IEEE Microwave and Wireless Component Letter, vol. 13, no. 2 (February 2003): pp. 57–59. 48. G. P. Gauthier, A. Courtay, and G. H. Rebeiz, “Microstrip Antennas on Synthesized Low DielectricConstant Substrate,” IEEE Trans. Antennas Propagat., vol. 45, no. 8 (August 1997): 1310–1314. 49. I. Papapolymerou, R. F. Drayton, and L. P. B. Katehi, “Micromachined Patch Antennas,” IEEE Trans. Antennas Propagat., vol. 46 (February 1998): 275–283. 50. J. S. Colburn and Y. Rahmat-Samii, “Patch Antennas on Externally Perforated High Dielectric Constant Substrates,” IEEE Trans. Antennas Propagat., vol. 47 (December 1999): 1785–1794. 51. D. R. Jackson, J. T. Williams, A. K. Bhattacharyya, R. L. Smith, S. J. Buchheit, and S. A. Long, “Microstrip Patch Designs that Do Not Excite Surface Waves,” IEEE Trans. Antennas Propagat., vol. 41, no. 8 (August 1993): 1026–1037. 52. R. Gonzalo, P. de Maagt, and M. Sorolla, “Enhanced Patch Antenna Performance by Suppressing Surface Waves Using Photonic-Bandgap Substrates,” IEEE Trans. Microwave Theory Tech., vol. 47 (November 1999): 2131–2138. 53. R. Cocciolo, F. R. Yang, K. P. Ma, and T. Itoh, “Aperture Coupled Patch Antenna on UC-PBG Substrate,” IEEE Trans. Microwave Theory Tech., vol. 47 (Nov. 1999): 2123–2130. 54. F. Yang, C. -S. Kim, and Y. Rahmat-Samii, “Step-like Structure and EBG Structure to Improve the Performance of Patch Antennas on High Dielectric Substrate,” 2001 IEEE AP-S Digest, vol. 2 (July 2001): 482–485. 55. F. Yang and Y. Rahmat-Samii, “Microstrip Antennas Integrated with Electromagnetic BandGap (EBG) Structures: A Low Mutual Coupling Design for Array Applications,” IEEE Trans. Antennas Propagat., vol. 51, no. 10 (October 2003): 2936–2946. 56. Z. Iluz, R. Shavit, and R. Bauer, “Microstrip Antenna Phased Array with Electromagnetic Bandgap Substrate,” IEEE Trans. Antennas Propagat., vol. 52, no. 6 (June 2004): 1446–1453. 57. L. Zhang, J. A. Castaneda, and N. G. Alexopoulos, “Scan Blindness Free Phased Array Design Using PBG Materials,” IEEE Trans. Antennas Propagat., vol. 52, no. 8 (August 2004): 2000–2007. 58. L. Yang et al, “A Novel Compact Electromagnetic-Bandgap (EBG) Structure and Its Applications for Microwave Circuits,” IEEE Trans. Microwave Theory Tech., vol. 53, no. 1 (January 2005): 183–190. 59. H. Boutayeb et al, “Analysis and Design of a Cylindrical EBG-based Directive Antenna,” IEEE Trans. Antennas Propagat., vol. 54, no. 1 (January 2006): 211–219. 60. A. Neto et al, “On the Optimal Radiation Bandwidth of Printed Slot Antennas Surrounded by EBGs,” IEEE Trans. Antennas Propagat., vol. 54, no. 4 (April 2006): 1074–1083. 61. Y. Q. Fu et al, “Mutual Coupling Reduction Between Large Antenna Arrays Using Electromagnetic Bandgap (EBG) Structures,” J. of Electromagnetic Waves and Applications, vol. 20, no. 6 (2006): 819–825.



34-27

62. F. Yang, A. Aminian, and Y. Rahmat-Samii, “A Novel Surface Wave Antenna Design Using a Thin Periodically Loaded Ground Plane,” Microwave Optical and Technology Letters, vol. 47, no. 3 (November 2005): 240–245. 63. Francis J. Zucker, Antenna Engineering Handbook, 3rd Ed., Richard C. Johnson (ed.), Chap. 12 (New York: McGraw-Hill, 1993). 64. R. Elliot, “Spherical Surface Wave Antennas,” IRE Trans. Antennas Propagat., vol. 4, no. 3 (July 1956): 422–428. 65. R. Hougardy and R. C. Hansen, “Scanning Surface Wave Antennas—Oblique Surface Waves over a Corrugated Conductor,” IRE Trans. Antennas Propagat., vol. 6, no. 4 (October 1958): 370–376. 66. L. B. Felson, “Radiation from a Tapered Surface Wave Antenna,” IRE Trans. Antennas Propagat., vol. 8 (November 1960): 577–586. 67. F. J. Zucker and J. A. Storm, “Experimental Resolution of Surface Wave Antenna Radiation into Feed and Terminal Patterns,” IEEE Trans. Antennas Propogat., vol. 18 (May 1970): 420–422. 68. Fan Yang, Yahya Rahmat-Samii, and Ahmed Kishk, “A Novel Surface Wave Antenna with a Monopole Type Pattern: A Thin Periodic Loaded Slab Excited by a Circular Disk,” 2005 IEEE APS International Symposium Digest, vol. 1A (July 3–8, 2005): 742–745. 69. D. R. Jackson and N. G. Alexópoulos, “Gain Enhancement Methods for Printed Circuit Antennas,” IEEE Trans. Antennas Propagat., vol. 33, no. 8 (September 1985): 976–987. 70. H. Mosallaei and K. Sarabandi, “Magneto-Dielectrics in Electromagnetics: Concept and Applications,” IEEE Trans. Antennas Propagat., vol. 52, no. 6 (June 2004): 1558–1567. 71. R. C. Hansen and M. Burke, “Antennas with Magneto-Dielectrics,” Microwave and Opt. Tech. Lett., vol. 26, no. 2 (July 2000): 75–78. 72. D. R. Smith, W. J. Padilla, D. C. Vier, S. C. A. Nemat-Nasser, and S. Schultz, “Composite Medium with Simultaneously Negative Permeability and Permittivity,” Phys. Rev. Lett., vol. 84, no. 18 (May 2000): 4184–4187. 73. H. Mosallaei, K. Sarabandi, and Y. Rahmat-Samii, “Novel Artificial Meta-Materials with Both e-m Parameters: A Composite Periodic Structure of Dielectric/Split Ring Resonators,” IEEE URSI International Symposium (June 16–21, 2002): 40. 74. H. Mosallaei and K. Sarabandi, “Embedded-Circuit Meta-Materials for Novel Design of Tunable Electro-Ferromagnetic Permeability, Band-Gap, and Bi-anisotropic Media,” IEEE AP-S International Symposium and USNC/CNC/URSI National Radio Science Meeting (June 22–27, 2003): 355–358. 75. ______, “Design and Modeling of Patch Antenna Printed on Magneto-Dielectric EmbeddedCircuit Metasubstrate,” IEEE Trans. Antennas Propagat., vol. 55, no. 1 (January 2007): 45–52. 76. K. Buell, H. Mosallaei, and K. Sarabandi, “A Substrate for Small Patch Antennas Providing Tunable Miniaturization Factors,” IEEE Trans. Microwave Theory Tech., vol. 54, no. 1 (January 2006): 135–146.


Low Profile Antenna Performance Enhancement Utilizing Engineered Electromagnetic Materials



Chapter 35

Reflectarray Antennas John Huang Jet Propulsion Laboratory California Institute of Technology


35-2

35.2 REVIEW OF DEVELOPMENT HISTORY . . . . . . . . . . . . . . . . . . . . .

35-3

35.3 ANALYSIS AND DESIGN PROCEDURES . . . . . . . . . . . . . . . . . . . .

35-9

35.4 BANDWIDTH ISSUES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35-12 35.5 APPLICATIONS AND RECENT DEVELOPMENTS . . . . . . . . . . . . . 35-14 35.6 SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35-20


Reflectarray Antennas 35-2

CHAPTER THIRTY-FIVE

35.1 INTRODUCTION This chapter gives an overview of the development history, key design methodologies, bandwidth issues, and applications for the reflectarray antenna, in particular the printed reflectarray. The reflectarray is an antenna consisting of either a flat or slightly curved reflecting surface and an illuminating feed, as shown in Figure 35-1. On the reflecting surface, there are many isolated elements (e.g. open-ended waveguides, printed microstrip patches, dipoles, or rings) without any power-division transmission lines. The feed antenna spatially illuminates these isolated elements, which are predesigned to re-radiate and scatter the incident field with electrical phases that are required to form a planar phase front in the far-field distance. This operation is similar in concept to the use of a parabolic reflector that utilizes its unique curvature to reflect and form a planar phase front when a feed is placed at its focal point. Thus the term “flat reflector” is sometimes used to describe the reflectarray, which utilizes both technologies of parabolic reflector and array. As shown in Figure 35-2, there are several methods for reflectarray elements to achieve a planar phase front. For example, one is to use identical microstrip patches with different-length phase delay lines attached so that they can compensate for the phase delays over the different paths from the illuminating feed. Another is to use variable-size patches, dipoles, or rings so that elements can have different scattering impedances and, thus, different phases to compensate for the different feedpath delays. With the third method, for circular polarization only, the reflectarray has all identical circularly polarized elements but with different angular rotations to compensate for the feedpath length differences. Reflectarrays using printed microstrip elements have been developed to achieve low reflecting surface profile, small antenna mass, and low manufacturing cost. These reflectarrays combine some of the salient features of the traditional parabolic reflector antenna and the microstrip array technology. Similar to a parabolic reflector, the reflectarray can achieve very good efficiency (> 50 percent) for a very large aperture since no power divider is needed and thus very little resistive insertion loss is encountered here. On the other hand, very similar to an array antenna, the reflectarray can have its main beam designed to tilt at a large angle (> 50°) from its broadside direction. Low-loss electronic phase shifters can be implanted into the elements for wide-angle electronic beam scanning. With this beam scanning capability

FIGURE 35-1

Configuration of a reflectarray antenna


Reflectarray Antennas REFLECTARRAY ANTENNAS

35-3

FIGURE 35-2 Various reflectarray elements: (a) identical patches with variable-length phase delay lines, (b) variable-size dipoles or loops, (c) variable-size patches, and (d) variable angular rotations

of the reflectarray, the complicated high-loss beamforming network and high-cost transmit/ receive (T/R) amplifier modules of a conventional phased array are no longer needed. One significant advantage of the printed reflectarray is that, when a large aperture (e.g. 10-m size) spacecraft antenna requires a deployment mechanism, the flat structure of the reflectarray allows a much simpler and more reliable folding or inflation mechanism to be accomplished than is possible with a specifically curved surface of a parabolic reflector. The flat reflecting surface of the reflectarray also lends itself to flush mounting onto an existing flat structure without adding a significant amount of mass and volume to the overall system structure. A reflectarray with hundreds or thousands of elements, in the form of a printed microstrip antenna, can be fabricated with a simple and low-cost etching process, especially when produced in large quantities. Another major feature of this antenna is that, with a large number of elements in a reflectarray having elemental phase adjustment capability, it can achieve a very accurate contour beam shape with a phase synthesis technique. With all the above capabilities, there is one distinct disadvantage associated with the reflectarray antenna: its inherent narrow bandwidth, which generally cannot exceed much beyond 10 percent depending on its element design, aperture size, focal length, etc. This narrow bandwidth behavior will be discussed further in a later section. Although the reflectarray has narrow bandwidth, due to its multitude of capabilities, the development, research, and application of the printed reflectarray antenna will continue throughout this century and beyond.

35.2 REVIEW OF DEVELOPMENT HISTORY Although the reflectarray antenna has not been widely exposed to the antenna community until recently, it was invented more than 40 years ago. Since then, many interesting reflectarray technologies have been developed. This section briefly reviews the development of



CHAPTER THIRTY-FIVE

this antenna technology throughout the past several decades. This review will not only inform a reader about this technology’s evolution, but also may lead him or her, by combining old thoughts with new technologies, to other new discoveries. Waveguide Reflectarray in the 1960s The reflectarray antenna concept was first conceived in the early 1960s.1 Either openor short-ended waveguide elements with variable-length waveguides, as illustrated in Figure 35-3, were used to demonstrate the capability of achieving co-phasal reradiated far-field beams. Because most wireless operations during this early time were done at relatively low microwave frequencies, the large-waveguide reflectarrays resulted in very bulky and heavy antennas, and thus this antenna concept was not pursued until more than ten years later. In addition, the efficiencies of these reflectarrays were not studied and optimized. Spiralphase Reflectarray in the 1970s In the mid 1970s, a very clever concept of “spiraphase” reflectarray was developed,2 where switching diodes, as illustrated in Figure 35-4, were used in a four-arm spiral or dipole element of a circularly polarized reflectarray to electronically scan its main beam to large angles from the broadside direction. This is possible because, by angularly rotating a circularly polarized radiating element, its propagating electrical phase will also change by an appropriate amount proportional to the amount of rotation. However, due to the thick spiral cavity (quarter-wavelength depth) and large electronic components, the spiraphase reflectarray was still relatively bulky and heavy. Its aperture efficiency was still relatively poor. Thus, no continued development effort was followed. It should be noted here that, in order to have good efficiency for the reflectarray, the intricate relations between the element beamwidth, element spacing, and focal-length/diameter ( f/D) ratio must be well designed; otherwise, a large backscattered component field or a mismatched surface impedance would result.

FIGURE 35-3 Conceptual drawing of the very early reflectarray using open-ended waveguides as elements



35-5

FIGURE 35-4 Reflectarray using a four-arm spiral element with switching diodes at center to achieve a 2-bit phase-shift system for circular polarization

Microstrip Reflectarray in the 1980s With the introduction of printable microstrip antennas, the technologies of reflectarray and microstrip radiators were combined, a typical configuration of which is illustrated in Figure 35-5. The first mention of using microstrip elements for reflectarray3 was in 1978. During the same year, the first attempt to analyze the microstrip reflectarray element using the infinite array approach was carried out.4 Since then, various printed microstrip reflectarray antennas were developed in the late 1980s and early 1990s for the purpose of achieving reduced antenna size and mass. These printed reflectarrays came in various forms, as shown in Figure 35-2, and all had flat, low-profile, low-mass reflecting surfaces. The ones that used identical patch elements with different-length phase delay lines5–10 had their elements arranged similar to those shown in Figure 35-2a. The phase delay lines, having lengths on the order of half-wavelength long or less, were used to compensate for the phase

FIGURE 35-5 Microstrip reflectarray with identical patches but different-length phase delay lines



CHAPTER THIRTY-FIVE

differences of different path lengths from the illuminating feed. The second approach, shown in Figure 35-2b, used elements that are made of printed dipoles with variable dipole lengths.11 Different dipole lengths will yield different scattering impedances, which then provide the different phases needed to compensate for the different path-length delays. Similarly, microstrip patches with variable patch sizes,12 shown in Figure 35-2c, were also developed. The concept of using circularly polarized microstrip patches with identical size but variable angular rotations,13,14 shown in Figure 35-2d, to form a co-phasal far-field reflectarray beam was conceived and developed. Recent Developments In addition to the various reflectarray elements shown in Figure 35-2, several other reflectarray or equivalent developments during the 1990s are worth mentioning here. Printed variable-length dipole elements, shown in Figure 35-6, were used to form a frequencyscanned grating-reflector antenna with an offset feed.15 Printed annular rings of variable diameters arranged in Fresnel Zone configuration, as shown in Figure 35-7, were also used to focus the beam.16 In the 1996 Phased Array Conference, a 94-GHz monolithic reflectarray17 fabricated in a single waffle, using 1-bit PIN diode phase shifters, was reported to achieve wideangle (± 45°) electronic beam scanning. In the same conference, a 35-GHz reflectarray, using waveguide/dielectric elements with 3-bit ferrite phase shifters,18 was also reported to achieve ± 25º beam scanning. One proposed technique,13 although not yet developed, is worth mentioning here. By using the angular rotation technique with circularly polarized elements, as depicted in Figure 35-8, miniature or micro-machined motors could be placed under each element to achieve wide-angle beam scanning without the need of T/R modules and phase shifters. For application in the spacecraft area, a deployable and low-mass

FIGURE 35-6 Printed dipole frequency-scanning grating-reflector antenna



35-7

FIGURE 35-7 Printed Fresnel Zone antenna with variablediameter annular rings

1-m-diameter inflatable reflectarray antenna19 at the X-band frequency was developed. Another unique spacecraft application of the reflectarray was conceived20 and developed21 by using its many elements, with a numerical phase synthesis technique, to form a uniquely shaped contour beam. From all the above developments, it can be seen that, at the end of the 20th century, the reflectarray antenna technology was becoming mature enough for possible applications throughout the microwave and millimeter-wave spectra.

FIGURE 35-8 Miniature micro-machined motors used to achieve beam scan by employing the angular rotation technique with CP elements



CHAPTER THIRTY-FIVE

FIGURE 35-9 bandwidth

Multilayer reflectarray to achieve wider

In the early 2000s, the development of reflectarray has mushroomed, and several performance improvement techniques are worth mentioning here. One uses multilayer stacked patches, as shown in Figure 35-9, to improve the reflectarray bandwidth from a few percent to more than 10 percent.22 As an extension to the 1-m X-band inflatable reflectarray mentioned above, a 3-m Ka-band inflatable reflectarray consisting of 200,000 elements was also developed,23 which currently is the electrically largest reflectarray. An X-band amplifying reflectarray was developed24 for each element of the reflectarray, as shown in Figure 35-10, to amplify the transmitted signal and, thus, achieve very high overall radiated power. To achieve

FIGURE 35-10 reflectarray

Unit-cell element of an X-band amplifying



FIGURE 35-11

35-9

Folded reflectarray to achieve a more compact antenna profile

good antenna efficiency, the most critical segment of the reflectarray design is its elements. The element performance was optimized by using the technique of genetic algorithm.25 The reflectarray using subreflector and array feed configuration to achieve fine beam scanning was also studied.26 To combat the shortcoming of narrow bandwidth, dual-band multilayer reflectarrays using annular rings27 and crossed dipoles28 are also being developed. Another development that is worth mentioning here is a folded reflectarray configuration,29 where two reflecting surfaces, as depicted in Figure 35-11, are used to reduce the overall antenna profile due to feed height of a conventional reflectarray.

35.3 ANALYSIS AND DESIGN PROCEDURES The design and analysis of a reflectarray can be separated into four essential steps, which are separately discussed below. Element Characterization The most important and critical segment of the reflectarray design is its element characterization. If the element design is not optimized, the reflectarray will not scatter the signal from the feed effectively to form an efficient far-field beam. Its beamwidth must correlate correctly with the reflectarray’s f/D ratio to accommodate all incident angles from the feed. Its phase change versus element change (patch size, delay line length, etc.) must be calibrated correctly. One of the most popular techniques for calibrating the phase is to use the infinite array approach12,30 to include local mutual coupling effect due to surrounding elements. It is not feasible for the current computer technology to have a complete rigorous solution to include all the mutual coupling effect of all different elements since the reflectarray generally consists of too many elements. The infinite array approach, which assumes all elements are identical, can be done by using the method of moments (MOM) technique12,30 or equivalently by using a finite difference time domain (FDTD) analysis on a unit cell of a single element.31 A mathematical waveguide simulator, which simulates the infinite array approach, can also be adapted by using the commercial software—HFSS (a finite element technique)—to achieve the element phase information. All of these techniques are used to derive the phase-versus-element-change curve, which is generally an S-shaped curve with a nonlinear relationship, as illustrated in Figure 35-12.



CHAPTER THIRTY-FIVE

FIGURE 35-12 A typical S curve of a reflectarray element phase change versus element change

The antenna designer should minimize the slope at the center of the curve so that the phase change will not be overly sensitive to the element change. If the curve is too steep, the element change or fabrication tolerance may become an issue, in particular at high microwave frequencies. Phase Delay Calculation The path lengths from the feed to all elements are all different, which leads to different phase delays. To compensate for these phase delays, the elements must have corresponding phase advancements designed according to a unique S curve similar to that shown in Figure 35-12. The following gives an example of how the compensating phase is calculated for each element of a reflectarray with a broadside-directed beam. The differential path length for each element is given as ∆ Lm,n = Lm,n − Lo,o

(35-1)

where Lm,n is the distance between the feed and the mnth element, which can be obtained by a simple geometry calculation. Lo,o is the distance between the feed and a reference point on the reflectarray surface, e.g. the center point. ∆ Lm,n is thus the differential feedpath length for the mnth element. To achieve a collimated radiation, the phase advancement ∆Φmn needed for the mnth element is given by ∆Φmn in degrees = [∆ Lm,n / lo − integer of (∆ Lm,n / lo )] × 360.

(35-2)

The above indicates that the compensating phase can be repeated every 360° and that the portion that is an integer multiple of a wavelength, or 360°, can be deleted. Pattern Calculation With all elements’ compensating phases known, the far-field radiation patterns can be calculated by the conventional array theory, where the radiation of all elements is summed



REFLECTARRAY ANTENNAS

together as follows. Consider a planar array consisting ofM × N elements that are nonuniformly illuminated by a low-gain feed at position vector rf . Let the desired beam direction be specified by unit vector uô. Then the far-field pattern of the reflectarray in the uˆ direction will be of the form M N       E (uˆ ) = ∑ ∑ F (rmn • rf ) ⋅ A(rmn • uô ) ⋅ A(uˆ • uô ) ⋅ exp[ jk ( | rmn − rf | + rmn • uˆ ) + jα mn] (35-3) m =1 n =1

where F is the feed pattern function, A is the reflectarray element pattern function, rmn is the position vector of the mnth element, and amn is the required compensating phase of the mnth element calculated by Eq. 35-2. Cosqq factor is used for both F and A functions with no azimuth (f) dependence. Geometry Design To determine the geometry of a reflectarray is basically to determine its f/D ratio, which is governed by its desired aperture efficiency. The aperture efficiency (ha) can be defined as the product of the illumination (hI) and spillover (hs) efficiencies: ha = hI × hs. By integrating the pattern function of Eq. 35-3, the illumination efficiency for a center-fed reflectarray can be obtained in a close form32 as given by

) (

(

)

 (1 + cos q +1 θ e) / (q + 1) + (1 − cos q θ e) / q   ηI =  2 tan 2 θ e (1 − cos 2 q +1 θ e) / (2q + 1)

2

(35-4)

and the spillover efficiency is given by

ηs = 1 − cos 2 q +1 θ e

(35-5)

where q is the exponent of the feed pattern function represented by cosqq, and qe is half of the subtend angle from the feed to the reflectarray aperture. The reflectarray element is approximated by a cosine function. Equations 35-4 and 35-5 are calculated by assuming a circular aperture only for the demonstration of the design procedures. Similar closedform equations can be easily obtained for square, rectangular, or elliptical apertures by performing proper integrations. To give an example of how Equations 35-4 and 35-5 can be utilized to optimize a reflectarray design, Figure 35-13 shows the calculated curve of spillover and illumination efficiencies versus the feed pattern factor q (feed beamwidth) for a 0.5-m 32-GHz reflectarray with a fixed f/D ratio of 1.0 (qe = 26.6°). It demonstrates that the maximum aperture efficiency is achieved at q = 10.5 or when the feed has a −3-dB beamwidth of 29°. Another curve, shown in Figure 35-14, gives aperture efficiency as a function of f/D ratio for the same 0.5-m 32-GHz reflectarray when the feed beamwidth is fixed at 33.4° with q = 8. In this case, the maximum aperture efficiency is achieved when the f/D ratio is 0.87. It can be seen that curves derived from Eqs. 35-4 and 35-5 are essential in obtaining an optimum efficiency design. The above discussion has been limited to center-fed reflectarray. Offset-fed reflectarray can also be optimally designed by using equations similar to Eqs. 35-4 and 35-5.



CHAPTER THIRTY-FIVE

FIGURE 35-13 shape

Spillover and illumination efficiencies versus feed pattern

35.4 BANDWIDTH ISSUES The bandwidth performance of a reflectarray13 is no match to that of a parabolic reflector, where, theoretically, infinite bandwidth exists. For a printed microstrip reflectarray, its bandwidth is primarily limited by two factors. One is the narrow bandwidth of the microstrip patch elements on the reflectarray surface, and the other is the differential spatial phase delay.

FIGURE 35-14 Aperture efficiency versus f/D ratio



35-13

Limit by Element Bandwidth The microstrip patch element generally has a bandwidth of about 3 to 5 percent. To achieve wider bandwidth for a conventional microstrip array, techniques such as using thick substrate for the patch, stacking multiple patches,22 and using sequentially rotated subarray elements have been employed. Bandwidths of more than 15 percent have been reported. Limit by Differential Spatial Delay The second reflectarray limiting factor, the differential spatial phase delay, can be best explained by referring to Figure 35-15, where the differential spatial phase delay, ∆S, is the phase difference between the two paths S1 and S2 from the feed to the reflectarray elements. This ∆S can be many multiples of the wavelength (l) at the center operating frequency. It can be expressed as ∆S = (N + d)l where N is an integer and d is a fractional number of a free-space wavelength l. At each element location, d is compensated by an appropriate phase delay achieved by the reflectarray element design (achieved by variable patch size, variable phase delay line length, etc.). As frequency changes, the factor (N + d)l becomes (N + d)(l + ∆l). Since the design and the compensating phase for each element are fixed for the center frequency, a frequency excursion error will occur in the reradiated phase front. The amount of phase change in each path when compared to a reference path, say S1, is (N + d)∆l, which can be a significant portion of a wavelength (360°). To reduce the amount of frequency excursion error, the integer number N must be reduced. There are several methods to reduce N. One is to design the reflectarray with a larger f/D ratio and hence to minimize the difference between paths S1 and S2. The second way is simply to avoid the use of a reflectarray with a large electrical diameter. The effects of f/D ratio and diameter on bandwidth performance were given previously in Figures 35-13 and 35-14. The third method to reduce frequency excursion error is to use time delay lines or partial time delay lines instead of the phase delays. In other words, when using the phase delay line technique (not the variable patch size technique), instead of using d∆l for the delay line length, (N + d)∆l could be used for the delay line. Certainly, additional line insertion loss and needed real estate for the lines are issues to be encountered. Another method to increase the bandwidth is to use, instead of a complete flat reflectarray surface, a concavely curved reflectarray with piecewise flat surfaces. This curved

FIGURE 35-15

Differential spatial phase delay of reflectarray



CHAPTER THIRTY-FIVE

reflectarray will continue to have advantages over a completely curved parabolic reflector; for example, its beam is able to be scanned to large angles with a phase shifter inserted into each element, and, for a space-deployable antenna, the piecewise flat surfaces in some cases are more easily folded into a smaller stowed volume. To mitigate the bandwidth problem, a recent technique of using multilayer stacked-patch elements22 not only has increased the element bandwidth but also has reduced the effect of differential spatial phase delay. As a net result, the bandwidth has increased from a few percent to more than 10 percent. Multiband techniques can also be applied to the reflectarray. Recently, two dual-band techniques have been developed for the X- and Ka-band frequencies. One uses double-layer membranes with two different-size rings and variable angular rotations,27 and the other also uses a double-layer approach with X-band crossed dipoles over Ka-band patches.28 To summarize, although the narrow bandwidth characteristic is the primary shortcoming of a reflectarray, there are several techniques that can be employed to alleviate the problem associated with the bandwidth issue.

35.5

APPLICATIONS AND RECENT DEVELOPMENTS

In addition to the possible reflectarray applications mentioned in the introduction and review sections, it is worthwhile here to present some details of several important applications and recent developments. Inflatable Reflectarray A Ka-band circularly polarized inflatable reflectarray23 with a 3-m-diameter aperture was developed by the Jet Propulsion Laboratory (JPL) and ILC Dover, Inc. for NASA’s future spacecraft communication antenna application. As shown in Figure 35-16, the antenna

FIGURE 35-16 A 3-m Ka-band inflatable reflectarray (the shining structure in front of the aperture is the surface flatness measurement device). The second photo shows the expanded view of reflectarray elements on thin membrane.



35-15

uses a torus-shaped inflatable tube to support and tension a 3-m-thin membrane reflectarray surface. This circularly polarized reflectarray, having approximately 200,000 elements using the variable angular rotation technique,13,14 is considered to be electrically the largest reflectarray ever built. It is much easier for the reflectarray, having a “natural” flat surface, to maintain its required surface tolerance (0.3-mm rms in this case) by the inflatable structure than a “nonnatural” parabolic surface; in particular, for long space flight. This inflatable antenna was later improved to equip it with rigidizable inflatable tubes23,33 in order to survive the hazardous space environment, such as bombardment by space debris and strenuous thermal effect. This reflectarray achieved an aperture efficiency of 30 percent with room for improvement and excellent far-field pattern shape with average sidelobe and cross-polarization levels below –40 dB, as shown in Figure 35-17.

Shaped Contour Beam A second important development of the reflectarray is the achievement of a shaped contour beam by using phase synthesis technique. This reflectarray, shown in Figure 35-18, was developed by the University of Massachusetts21 for a commercial application to provide Earth contour beam coverage. A typical calculated contour beam of this antenna, by using phase synthesis technique, is given in Figure 35-19. Since a reflectarray generally has more than thousands of elements, it thus has many degrees of freedom in design to provide an accurate and uniquely required contour beam. However, this capability of beam shaping is limited by its frequency bandwidth. A more recent reflectarray development34 using triplelayer stacked patches achieved a specified contour beam shape within a relatively wide bandwidth of 10 percent.

Dual-band Cassgrain Reflectarray A third important development is a dual-frequency reflectarray, where the two frequencies are widely separated, such as the X- and Ka-bands. The developed prototype antenna,

FIGURE 35-17

Measured radiation pattern of the 3-meter Ka-band inflatable reflectarray



CHAPTER THIRTY-FIVE

FIGURE 35-18 Ku-band reflectarray with shaped contour beam capability (Courtesy of Professor Dave Pozar, Univ. of Massachusetts)

shown in Figures 35-20 and 35-21, is circularly polarized and uses variable angularly rotated annular rings.14,27 It was developed by Texas A&M University for JPL/NASA’s future space communication application. This antenna, with a Cassegrain offset feed and an aperture diameter of 0.75 m, uses a multilayer technique where the X-band annular rings

FIGURE 35-19 A measured contour beam plot of the reflectarray shown in Figure 35-18



35-17

FIGURE 35-20 X/Ka dual-band Cassegrain reflectarray antenna (0.75-m diameter) using annular ring elements

are placed above the Ka-band rings and serve as a frequency-selective surface to let the Ka-band signal pass through. The measured results indicate that there is very little impact on the X-band performance due to the presence of the Ka-band elements. The measured radiation patterns at the X-band and Ka-band frequencies are shown in Figures 35-22 and 35-23, respectively. These patterns show excellent behavior with relatively low sidelobe and low cross-polarization levels. However, the measured Ka-band gain of the dualfrequency dual-layer antenna is about 1.0 dB lower than the Ka-band-alone antenna. The Ka-band-alone reflectarray has a measured aperture efficiency of 50 percent, while the dual-frequency dual-layer antenna has a Ka-band efficiency of about 40 percent. In other words, the X-band annular rings did impact somewhat the Ka-band performance. Efforts need to be carried out in the future to minimize this impact.

FIGURE 35-21 Sketch and photo of the X/Ka dual-band reflectarray antenna showing two membranes with annular ring elements



FIGURE 35-22

CHAPTER THIRTY-FIVE

Measured X-band radiation pattern of the X/Ka dual-band reflectarray

Foldable Piecewise-Flat Globally Curved Reflectarray A recent development that is worth mentioning is a reflectarray having a rectangular aperture that is intended for NASA/JPL’s Wide Swath Ocean Altimeter (WSOA) radar application. This reflectarray uses variable-size patches as elements. The required rectangular aperture,

FIGURE 35-23

Measured Ka-band radiation pattern of the X/Ka dual-band reflectarray



FIGURE 35-24

35-19

Piece-wise flat reflectarray for space application

as shown in Figure 35-24, consists of five flat subapertures that are connected together to form a curved reflectarray.35 The curving of the long dimension of the rectangular surface is to minimize the incident angles from the feed for the end elements and, thus, to optimize the radiation efficiency for all elements. The radiation efficiency here indicates the measure of the amount of energy of each element that is reradiated in the desired main beam direction. The advantage of using reflectarray with flat subapertures is to allow mechanical folding of the flat panels into a compact structure for spacecraft launch-vehicle stowage. Test data indicates that this reflectarray is functioning properly with good radiation patterns and an aperture efficiency above 50 percent. Beam Scanning Reflectarray One of the key advantages of a reflectarray is that its elements can be implanted with lowloss phase shifters and achieve wide beam scanning without the use of a complicated beamformer (power divider) and expensive T/R amplifier modules. As mentioned in Section 35.2, in the mid 1990s a 94-GHz monolithic reflectarray17 fabricated in a single waffle, using 1-bit PIN diode phase shifters, achieved wide-angle (±45°) electronic beam scanning. Also, a 35-GHz reflectarray, using waveguide/dielectric elements with 3-bit ferrite phase shifters,18 was reported to achieve ±25° beam scanning. More recent developments in this area include ferroelectric thin-film phase shifters36,37 for reflectarray. Interest in ferroelectric thin-film circuits developed because of their high power-handling capability, negligible DC power consumption, small size, and potential for low cost. Current research effort is to reduce the phase shifter’s relatively high insertion loss (from many decibels to 1 or 2 dB), in particular at the millimeter-wave frequency band. Another development worth mentioning here is the use of a miniature motor to rotate each element of a circularly polarized reflectarray. This concept was proposed in 199513 and has not been investigated in detail until recently, in which case two miniature motors with two annular-ring reflectarray elements were used in a waveguide simulator38 as well as a 5 × 1–element array39 to demonstrate the beam scan. A third recent important development is the use of the tunable varactor diode40,41,42 to control the phase change, as illustrated in Figure 35-25. The phase is analog controlled by its variable control voltage. It not only provides more precision phase but also minimizes the number of control lines when compared to the conventional quantized switched-line phase shifters. For instance, a tunable varactor diode only needs a single variable-voltage control to achieve a full 360° of phase variation, while a 3-bit switched line needs 12 control lines with 45° of quantization error. It is important to minimize the number of control/ bias lines in a reflectarray system where generally a huge number of elements is involved.



CHAPTER THIRTY-FIVE

FIGURE 35-25 Beam scanning reflectarray using varactor diodes as phase shifters, with photo showing a C-band 70-element reflectarray using varactor diodes (Courtesy of Sean Hum, University of Calgary, Canada)

35.6 SUMMARY The reflectarray antenna technology has come a long way. However, its development and application had not been widely adapted until about a decade ago when the printable microstrip reflectarray was introduced. Except for its narrow bandwidth characteristic, the reflectarray has many advantages over a parabolic reflector antenna type. The main beam of a reflectarray can be designed to tilt to a large angle from its broadside direction. Phase shifters can be implanted into the elements for wide-angle electronic beam scanning. For large-aperture spacecraft antenna applications, the reflectarray’s flat surface allows the antenna to be made into an inflatable structure with relative ease in maintaining its surface tolerance compared to a curved parabolic surface. Its flat surface also can be made of multiple flat panels for ease in folding into a more compact structure for launch vehicle stowage. Very accurate contour beam shape can be achieved with phase synthesis technique for Earth coverage application. Due to these multitudes of capabilities, the door has just opened for the development, research, and application of the printed reflectarray antennas. Major areas that need continuing improvement of the reflectarray performance are its bandwidth, radiation efficiency, and beam scanning capability. Acknowledgment Portions of the research activities described in this chapter were carried out by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration.

REFERENCES 1. D. G. Berry, R. G. Malech, and W. A. Kennedy, “The Reflectarray Antenna,” IEEE Trans. Antennas Propagat., vol. AP-11 (Nov. 1963): 645–651. 2. H. R. Phelan, “Spiralphase Reflectarray for Multitarget Radar,” Microwave Journal, vol. 20 (July 1977): 67–73.



35-21

3. C. S. Malagisi, “Microstrip Disc Element Reflect. Array,” Electronics and Aerospace Systems Convention (Sept. 1978): 186–192. 4. J. P. Montgomery, “A Microstrip Reflectarray Antenna Element,” Antenna Applications Symposium, University of Illinois, Sept. 1978. 5. R. E. Munson and H. Haddad, “Microstrip Reflectarray for Satellite Communication and RCS Enhancement and Reduction” (1987): U.S. Pat. 4,684,952. 6. J. Huang, “Microstrip Reflectarray,” IEEE AP-S/URSI Symposium Digest (June 1991): 612–615. 7. T. A. Metzler, “Design and Analysis of a Microstrip Reflectarray,” Ph.D. dissertation, University of Massachusetts, September 1992. 8. Y. Zhang, K. L. Wu, C. Wu, and J. Litva, “Microstrip Reflectarray: Full-wave Analysis and Design Scheme,” IEEE AP-S/URSI Symposium (June 1993): 1386–1389. 9. R. D. Javor, X. D. Wu, and K. Chang, “Beam Steering of a Microstrip Flat Reflectarray Antenna,” IEEE AP-S/URSI Symposium (June 1994): 956–959. 10. D. C. Chang and M. C. Huang, “Multiple Polarization Microstrip Reflectarray Antenna with High Efficiency and Low Cross-Polarization,” IEEE Trans. Antennas Propagat., vol. 43 (Aug. 1995): 829–834. 11. A. Kelkar, “FLAPS: Conformal Phased Reflecting Surfaces,” Proc. IEEE National Radar Conf. (March 1991): 58–62. 12. D. M. Pozar and T. A. Metzler, “Analysis of a Reflectarray Antenna Using Microstrip Patches of Variable Size,” Electronics Letters (April 1993): 657–658. 13. J. Huang, “Bandwidth Study of Microstrip Reflectarray and a Novel Phased Reflectarray Concept,” IEEE AP-S/URSI Symposium (June 1995): 582–585. 14. J. Huang and R. J. Pogorzelski, “A Ka-band Microstrip Reflectarray with Elements Having Variable Rotation Angles,” IEEE Trans. Antennas Propagat., vol. 46 (May 1998): 650–656. 15. F. S. Johansson, “A New Planar Grating-Reflector Antenna,” IEEE Trans. Antennas Propagat., vol. 38 (Sept. 1990): 1491–1495. 16. Y. T. Gao and S. K. Barton, “Phase Correcting Zonal Reflector Incorporating Rings,” IEEE Trans. Antennas Propagat., vol. 43 (April 1995): 350–355. 17. J. M. Colin, “Phased Array Radars in France: Present and Future,” IEEE Symposium on Phased Array System and Technology (Oct. 1996): 458–462. 18. A. A. Tolkachev, V. V. Denisenko, A. V. Shishlov, and A. G. Shubov, “High-Gain Antenna System for Millimeter-Wave Radars with Combined Electrical and Mechanical Beam Steering,” IEEE Symposium on Phased Array System and Technology (Oct. 1996): 266–271. 19. J. Huang and A. Feria, “A 1-m X-band Inflatable Reflectarray Antenna,” Microwave and Optical Technology Letters, vol. 20 (Jan. 1999): 97–99. 20. J. Huang, “Capabilities of Printed Reflectarray Antennas,” IEEE Symposium on Phased Array System and Technology (Oct. 1996): 131–134. 21. D. M. Pozar, S. D. Targonski, and R. Pokuls, “A Shaped-Beam Microstrip Patch Reflectarray,” IEEE Trans. Antennas Propagat., vol. 47 (July 1999): 1167–1173. 22. J. A. Encinar, “Design of Two-Layer Printed Reflectarray Using Patches of Variable Size,” IEEE Trans. Antennas Propagat., vol. 49 (Oct. 2001): 1403–1410. 23. J. Huang, V. A. Feria, and H. Fang, “Improvement of the Three-Meter Ka-band Inflatable Reflectarray Antenna,” IEEE AP-S/URSI Symposium (July 2001): 122–125. 24. M. Bialkowski, A. W. Robinson, and H. J. Song, “Design, Development, and Testing of X-band Amplifying Reflectarrays,” IEEE Trans. Antennas Propagat., vol. 50 (Aug. 2002): 1065–1076. 25. R. E. Zich, M. Mussetta, M. Tovaglieri, P. Pirinoli, and M. Orefice, “Genetic Optimization of Microstrip Reflectarrays,” IEEE AP-S/URSI Symposium (June 2002): III-128–131. 26. B. Khayatian and Y. Rahmat-Samii, “Characterizing Reflectarray Antenna Radiation Performance,” IEEE AP-S/URSI Symposium, Columbus, Ohio, June 2003. 27. J. Huang, C. Han, and K. Chang, “A Cassegrain Offset-Fed Dual-band Reflectarray,” IEEE AP-S/URSI Symposium, Albuquerque, NM, July 2006.



CHAPTER THIRTY-FIVE

28. M. Zawadzki and J. Huang, “A Dual-band Reflectarray for X- and Ka-bands,” PIERS Symposium, Honolulu, Hawaii, October 2003. 29. W. Menzel, D. Pilz, and M. Al-Tikriti, “Millimeter-Wave Folded Reflector Antennas with High Gain, Low-loss, and Low Profile,” IEEE Antennas & Propagation Magazine, vol. 44, no. 3 (June 2002): 24–29. 30. D. Pozar, S. D. Targonski, and H. D. Syrigos, “Design of Millimeter Wave Microstrip Reflectarrays,” IEEE Trans. Antennas Propagat., vol. 45 (Feb. 1997): 287–296. 31. E. Girard, R. Moulinet, R. Gillard, and H. Legay, “An FDTD Optimization of a Circularly Polarized Reflectarray Unit Cell,” IEEE AP-S/URSI Symposium (June 2002): III-136–139. 32. J. Huang, “Analysis of a Microstrip Reflectarray Antenna for Microspacecraft Applications,” JPL TDA Progress Report No. 42–120 (Feb. 15, 1995). 33. H. Fang, M. Lou, J. Huang, L. M. Hsia, and G. Kerdanyan, “An Inflatable/Self-Rigidizable Structure for the Reflectarray Antenna,” 10th European Electromagnetic Structure Conference, Munich, Germany, Oct. 2001. 34. R. Hodges and M. Zawadzki, “Design of a Large Dual Polarized Ku-band Reflectarray for Spaceborne Radar Altimeter,” IEEE AP-S Symposium (June 2005): 4356–4359. 35. J. A. Encinar and J. A. Zornoza, “Three-layer Printed Reflectarrays for Contoured Beam Space Applications,” IEEE Trans. Antennas Propagat., vol. 52 (May 2004): 1138–1148. 36. R. R. Romanofsky, J. T. Bernard, F. W. Van Keuls, F. A. Miranda, G. Washington, and C. Canedy, “K-band Phased Array Antennas Based on BST Thin-Film Phase Shifters,” IEEE Trans. Microwave Theory and Tech., vol. 48 (Dec. 2000): 2504–2510. 37. F. Xiong and R. R. Romanofsky, “Study of Behavior of Digital Modulations for Beam Steerable Reflectarray Antennas,” IEEE Trans. Antennas Propagat., vol. 53 (March 2005): 1083–1097. 38. A. E. Martynyuk, J. I. M. Lopez, J. R. Cuevas, and Y. K. Sydoruk, “Wideband Reflective Array Based on Loaded Metal Rings,” IEEE MTT-S Microwave Symposium, Long Beach, Calif., June 2005. 39. V. F. Fusco, “Mechanical Beam Scanning Reflectarray,” IEEE Trans. Antennas Propagat., vol. 53 (Nov. 2005): 3842–3844. 40. L. Boccia, G. Amendola, and G. Di Massa, “A Microstrip Patch Antenna Oscillator for Reflectarray Applications,” IEEE AP-S/URSI Symposium (June 2004): 3927–3930. 41. S. V. Hum, M. Okoniewski, and R. J. Davies, “Realizing an Electronically Tunable Reflectarray Using Varactor Diode-tuned Elements,” IEEE Microwave and Wireless Components Letters, vol. 15 (June 2005): 422–424. 42. M. Riel and J. J. Laurin, “Design of a C-band Reflectarray Element with Full Phase Tuning Range Using Varactor Diodes,” IEEE AP-S/URSI Symposium, Washington D. C., July 2005.



Chapter 36

Mobile Handset Antennas Yiannis C. Vardaxoglou Professor of Wireless Communications Loughborough University, UK

Jim R. James Emeritus Professor, Cranfield University, UK Consultant Engineer, Visiting Professor, Loughborough University, UK CONTENTS 36.1 IMPACT ON ANTENNA DESIGN . . . . . . . . . . . . . . . . . . . . . . . . . .

36-2

36.2 CELLULAR HANDSET ANTENNA DESIGN ISSUES . . . . . . . . . .

36-2

36.3 HELICAL WIRE ANTENNAS AND VARIANTS . . . . . . . . . . . . . . . .

36-5

36.4 EVOLUTION OF THE PIFA AND ITS VARIANTS . . . . . . . . . . . . . .

36-7

36.5 CERAMIC CHIP AND RESONATOR ANTENNAS . . . . . . . . . . . . . 36-11 36.6 SAR MEASUREMENT AND MINIMIZATION . . . . . . . . . . . . . . . . 36-14 36.7 PROVISION FOR GPS AND BLUETOOTH. . . . . . . . . . . . . . . . . . . 36-20 36.8 MEASUREMENT OF HANDSET ANTENNAS . . . . . . . . . . . . . . . 36-21 36.9 SATCOM HANDSET ANTENNAS . . . . . . . . . . . . . . . . . . . . . . . . . 36-22 36.10 FUTURE TRENDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36-24 36.11 SYMBOLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36-25 36-1 Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2007 The McGraw-Hill Companies. All rights reserved. Any use is subject to the Terms of Use as given at the website.

Mobile Handset Antennas 36-2

CHAPTER THIRTY-SIX

36.1 IMPACT ON ANTENNA DESIGN Antenna engineering design has many interesting and surprising features, not least of which is its ability to reinvent itself if the demand arises. Such has been the impact of mobile communications on antenna design, resulting in unbelievably compact and densely packaged dipole-like handset antennas housed entirely within the mobile phone case itself. The dipole and monopole were some of the first antennas developed and, together with fundamental electromagnetic theory, have been well established over the past century. Yet new versions of these generic radiators continue to be analyzed, simulated, developed, and even researched today as is evident from the research journals worldwide. Quite simply, product demand is the driving force for this seemingly endless creation of different physical realizations of what is essentially an antenna capable of giving a dipole-like radiation pattern. While the vectorial radiation field of an antenna is uniquely defined by the antenna itself, this is not so for typical engineering radiation patterns specified as the modulus of the radiation field whereby phase information is not retained. Furthermore, this specification is often made even less restrictive by only stating limiting values of sidelobes and other pattern features, between which the measured pattern must lie. Bandwidth specifications are similarly pinned down with only a simple limiting level on the S11 response, hence little information is demanded about the shape of the bandpass characteristic itself. The nonuniqueness of such specifications thus leaves plenty of scope to recast the antenna in a very different form. The dipole antenna family, and in fact most other types of antennas, can be reinvented in a different physical form to satisfy the demands of new equipment products. An outstanding example is the microstrip patch antenna, made possible with printed antenna technology. When compared with conventional monopoles and loop antennas, the printed version can be expected to have a somewhat lower radiation efficiency but it has opened the door to many innovative new products requiring a very low profile dipole-like radiator that is compatible with cost-cutting manufacturing processes. The previous example not only illustrates how new and very different physical realizations can be developed with similar radiation pattern characteristics but also highlights the need for a system’s design approach whereby the performance of individual components is not the main criterion but rather the satisfactory performance of the entire system. The deployment of a system approach has gradually become more evident in mobile handset antenna design over some two decades. The initial thinking was to mount a conventional antenna on the mobile phone case itself, and early examples were physically large and included balanced fed dipoles, monopoles, and loops. The astronomical demand for mobile phones together with the pull of user preference has resulted in the present-day integral handset antenna. Seldom does one ask of the performance of the antennas in isolation where it is understood that the matching and radiation efficiency data may be less impressive than that of conventional antennas. What matters is how it all functions together as a system.

36.2 CELLULAR HANDSET ANTENNA DESIGN ISSUES Without doubt, integral handset antenna design is very complicated and presents the designer with perhaps the most formidable antenna design criteria demanded by a product. There are numerous design issues for cellular operation that are summarized next, but the integration as a system further complicates matters, and many, if not most, of the design


Mobile Handset Antennas MOBILE HANDSET ANTENNAS

36-3

issues are interrelated, thus there is a high degree of iteration in the design work. The requirements for a global roaming satcom handset antenna have different constraints and are considered separately in Section 36.9. Antenna Electrical Size What constitutes an electrically small antenna is well established and seminal contributions include Wheeler,1 Chu,2 and Collin and Rothschild.3 The radian sphere4 of radius a is defined as a=

λ 2π

(36-1)

where l is the free-space wavelength. If an antenna’s physical structure lies within this sphere, then electrically small antenna characteristics will be evident. The effect of shape and volume of the antenna structure has also been considered by Wheeler.1 For a lossless antenna, the effects are (i) a high Q factor and hence small bandwidth, (ii) a small radiation resistance resulting in a low radiation efficiency, and (iii) likely greater antenna sensitivity to mechanical and electrical tolerances. When the antenna has intrinsic losses, the bandwidth is less narrow. It is widely acknowledged5 that for diminishing a Q ∼ 1 / (ka)3

(36-2)

where k = 2p /l. From an antenna engineering standpoint, the interest6,7 over many years has centered on realizing a small antenna in practice that lies on or exceeds the performance limits represented by the Q formula. This topic has been vigorously revisited recently7,8 with the upsurge of activity in small antenna design for both mobile and wireless applications. For handset antennas, it is the lower bands that enforce antenna electrical size constraints. For instance, at 800 MHz, a = 5.97 cm and the dimensions of a typical handset are 1 × 4 × 10 cm, so if the antenna has dimensions of 1 × 2 × 4 cm, it lies within the radiansphere. At 1800 MHz, a = 2.64 cm, so the situation is less critical. An unknown factor is, of course, the extent to which a given handset antenna couples to the handset circuit components and battery, etc., because this will in effect increase the electrical size of the antenna. Ground Plane Electrical Size Effects When a monopole is mounted on a horizontal ground plane (GP) of finite extent, the dipolelike radiation patterns can be corrupted with a variety of pattern effects depending on the electrical size of the GP. An excellent theoretical illustration with measurements has been given9 showing the increase in pattern distortion as a function of reduced GP electrical size. The mobile handset GP is enforced mainly by the metallized assembly consisting of the circuit board and its components, which as mentioned above is typically 4 × 10 cm. Any plastic outer case will have little effect. Neglecting hand effects, such a GP could act as an electromagnetic counterpoise to a 10-cm monopole mounted in the same plane as the GP at 750 MHz. Operation at a lower frequency is possible using a monopole version of the sleeve dipole10 to isolate the handset



CHAPTER THIRTY-SIX

case, but the resulting antenna has greater height and is not acceptable today where the antenna element is usually mounted on the GP itself. The GP extent in one direction is only 4 cm and some degree of radiation pattern perturbation is likely to be experienced at the lower mobile bands. Consequently, it is common practice in research papers and manufacturers’ data sheets to show measured and/or simulated handset antenna radiation patterns with the antenna embedded on the GP. There continues to be much interest in creating a stand-alone chip type antenna11 that is immune to the proximity of the GP and its components. Another way of addressing the latter requirements concerns a balanced feed arrangement,12 but since this concept requires two identical antennas in phase opposition, its application13 so far has been confined to the higher mobile bands due to size limitations. Time Varying Pattern Effects Cellular handset antennas have to function in a rich multipath propagation environment which allows the use of antennas with polarization properties that are not tightly specified. Hand and head movement, together with the angular movement of the handset, creates further time varying effects which, together with multipath effects, are mainly beyond the designer’s control. At the lower mobile bands below 1 GHz, the antenna’s constrained electrical size commits it to smoother dipole-like patterns with few nulls and thus time varying effects can be better tolerated than at progressively higher frequencies where additional pattern nulls appear. The situation is further assisted by the self-correcting adaptivity of speech communication. Increasing Band Coverage It is generally thought that phone users have an appetite for mobile phones having additional functions that can obtain wider news and sales information and provide more games and a host of other facilities like GPS and wireless links. Financial returns from such a multimedia operation are likely to be very significant but at the same time the cost of the mobile handset must be constrained. Not surprisingly, handset manufacturers aim to achieve several bands from one central antenna not only to maintain cost levels but also because there is little available space to do otherwise. A multifunction antenna is now commonly the stated requirement, and antenna designers have responded with a variety of configurations that demonstrate design concepts and feasibility. Just how far this concept can be extended remains to be seen, but antennas with three bands have been demonstrated.14 It is evident that the multiband operation may require some trading of other performance parameters such as radiation efficiency and antenna input match levels. This will limit the number of bands that can eventually be obtained from a single antenna. Radiation Efficiency Whatever type of antenna is chosen to embed in the handset, the designer will strive to minimize the power lost to dissipation in the antenna itself. A measure of the losses is the antenna radiation efficiency ha defined by ha = Prad / (Prad + Ploss )

(36-3)



MOBILE HANDSET ANTENNAS

where the radiated power is Prad and the power dissipated in the antenna is Ploss. In practice the antenna is likely to need a matching network, and losses in the latter are expressed as a matching network efficiency hm where hm = ha Qm / (Qm + Qa )

(36-4)

where Qm = matching network Q factor and Qa = antenna Q factor. The combined efficiency of the antenna and its matching network is then ha hm.15 Further losses are of course incurred when the antenna is embedded in the handset, due to coupling to handset components, and finally when operating with the hand and head present, which account for much of the additional power loss.

36.3 HELICAL WIRE ANTENNAS AND VARIANTS The remarkable property of wire antennas, or indeed any form of linear metal conductors, is that the resonant frequency is mainly determined by the electrical length of the current path. As such, a wire monopole (or dipole) can be compacted into an electrically smaller size. The antenna bandwidth, however, is more affected and increases for compacted structures occupying a greater volume. A common example where the wire monopole is partially wound into a coil at some location on the wire16 is shown in Figure 36-1a. For handset applications, a spiral coil of wire alone constitutes the antenna (see Figure 36-1b) and for small diameters is known as the normal mode helical antenna (NMHA). The device gives dipole radiation patterns when placed on an electrically large GP but has circularly polarized radiation characteristics when the spiral diameter exceeds about one wavelength. The bandwidth, radiation efficiency, and radiation resistance at resonance are significantly reduced but the latter can be restored by tapping into the spiral, as shown in Figure 36-1c, at the expense of a more complicated construction. There are a multitude of variations, and in Figure 36-1d a second spiral coil is wound in antiphase directly on top of the first spiral coil, creating a large bandwidth but very poor radiation efficiency. This is referred to as the double wound antenna.17 Figures 36-1e and f show zigzag-like versions and other designs that can be configured in fractal form or generated by genetic algorithms. Some performance details for spiral antenna devices are compared to those of a wire monopole in Table 36-1. A compact manufactured NMHA is illustrated in Figure 36-2 where the conducting path is printed into a groove in the cylindrical plastic former to avoid the increased production cost of winding on a wire.

TABLE 36-1

Performance of Spiral-type Antennas Compared with Monopole Antenna

Antenna Type

Resonance (MHz)

Bandwidth (MHz)

Loss (dB)

Efficiency (%)

Normalized Size

Resistance (Ω)

Monopole

150

∼20

∼0

∼100

1

25

NMHA

142

∼1.5

1.04

81

0.22

5

Double wound

150

∼9

7.8

16

0.18

21



CHAPTER THIRTY-SIX

(a)

(b)

(c)

(d)

(e)

(f)

FIGURE 36-1 (a) Coil loaded whip antenna; (b) NMHA; (c) NMHA with tapping connection for matching; (d) double wound spiral antenna; (e) and ( f ) zigzag antennas. (Figures 16-1a to 1d reproduced by permission of K. Fujimoto and J. R. James11 © Artech House, Inc. 2001.)

FIGURE 36-2 Example of typical NMHA mounted on a handset case showing construction11 (Courtesy of Nippon Antenna Co. Ltd.)



36-7

(a)

(b) FIGURE 36-3 (a) Printed zigzag antennas printed on dielectric films11 (Courtesy of Allgon); (b) The structure rolled into a cylindrical shape mounted on a plastic former with and without the cover in place

The zigzag antenna configurations have an outstanding manufacturing advantage over the spiral devices because they can be printed on thin dielectric films and then rolled into a cylindrical or other three-dimensional shape. Such assemblies have less size reduction than the spiral types but have attracted use both mounted on a handset case, as shown in Figure 36-3, and embedded within the handset interior as a planar structure. Two printed films designed for different resonant frequencies can also be connected in parallel to obtain a dual-band antenna, and this is more difficult to achieve with spiral conductors.

36.4 EVOLUTION OF THE PIFA AND ITS VARIANTS The planar inverted F antenna (PIFA) has come into prominence due to its common use in mobile phone handsets. In fact, the PIFA is often regarded as a generic antenna in its own right, but from physical fundamentals it is seen to be a derivative of both wire antennas and printed patch antennas. For instance, a low profile inverted L antenna (ILA) has a low radiation resistance, but by tapping the feed connection along the wire, it can be increased, which is the inverted F antenna (IFA), as shown in Figure 36-4a. This antenna is useful for low profile applications where only a thin wire can be used. Where space is available, the L-shaped wire can be replaced by a planar conductor of width L1, as shown in Figure 36-4b. The planar conductor increases the radiation resistance and offers a two-dimensional choice of feed position to facilitate matching. The end shorting plate of height H need not occupy the entire plate width, as shown in Figure 36-4b.



CHAPTER THIRTY-SIX

(a)

(b)

FIGURE 36-4 (a) Inverted F wire antenna and (b) typical PIFA

Examination of Figure 36-4 also reveals that the PIFA can be regarded as a quarter-wave patch antenna18 with an air-spaced substrate. The PIFA has both vertical and horizontal radiation surfaces, and the radiation patterns can be expected to contain significant crosspolarization, which is not a problem for cellular mobile phones since they have to operate in a high multipath environment. Being purely a simple bent metal antenna with no lossy substrate, the PIFA has good radiation efficiency. However, it is sensitive to its environment, and when embedded in a handset with the presence of the hand and head, the radiation efficiency reduces drastically.19 The radiation patterns are also significantly perturbed by the hand and head movement and appreciable power is dissipated in the head. However, the simplicity of the PIFA and the low manufacturing cost remain outstanding advantages that are attractive to an industry where cost-cutting is a paramount consideration. The demand for multiband operation at no additional antenna cost has inspired the creation of the multiband PIFA; a dual-band example is shown in Figure 36-5. The simplicity of this antenna is remarkable and there would appear to be an unlimited number of ways of configuring the current paths on the top plate of the PIFA or indeed to other metal-plate antennas. This concept has been extended to the tri-band version shown in Figure 36-6.

(a)

(b)

FIGURE 36-5 (a) Dual-band PIFA (after Y.-X. Guo et al © IEEE 2004) and (b) bandwidth characteristics of dual-band antenna 20




(a)

(b)

FIGURE 36-6 (a) Tri-band monopole (after G.-Y. Lee and K.-L. Wong © Micro. Opt. Technol. Lett. 2002); (b) Side view of monopole antenna mounted on the handset GP; (c) Measured and simulated bandwidth characteristics of tri-band antenna 21

As already mentioned, radiation patterns having a high cross-polarization content can be expected, and examples of radiation patterns for the tri-band antenna of Figure 36-6 are shown in Figure 36-7. Similar performance can be obtained with the meander-line concept, which allows a printed zigzag conductor to be printed on a thin planar dielectric film prior to folding into the rectangular 6 × 6.5 × 25 mm3 box shape illustrated in Figure 36-8. This tri-band antenna is very compact and covers the GSM (900 MHz), DCS (1800 MHz), and PCS (1900 MHz) bands. The large degree of design freedom offered by these simple metal antenna structures has enabled manufacturers to generate their own in-house low-cost designs. Further compacting of these antennas can be achieved by loading the low-frequency resonant zones of the structure with a high-permittivity dielectric slab, albeit at the expense of a somewhat lower radiation efficiency and bandwidth for the low band and the extra cost of the ceramic material.

(a)

(b)

FIGURE 36-7 (a) Radiation patterns of the tri-band antenna at 900 MHz (after G.-Y. Lee and K.-L. Wong21 © Micro. Opt. Technol. Lett. 2002); (b) Radiation patterns of the tri-band antenna at 1800 MHz



CHAPTER THIRTY-SIX

(a)

(b)

(c) FIGURE 36-8 (a) The assembled and mounted tri-band meander-line antenna (after P.-L. Teng and K.-L. Wong22 © Micro. Opt. Technol. Lett. 2002); (b) The meander-line antenna before assembly; (c) Measured and simulated bandwidth characteristics of the tri-band meander-line antenna



36-11

36.5 CERAMIC CHIP AND RESONATOR ANTENNAS A major manufacturing problem with PIFAs and other air-spaced antennas embedded in the handset is that they are not stand-alone electronic components. These antennas couple to the handset components, and a previously proven antenna cannot be simply embedded into a new handset configuration without further design adjustments. The ceramic chip antenna concept goes some way toward alleviating this coupling problem and also reduces the antenna size at the expense of additional material costs. The fields are intensified within the ceramic material, and unwanted characteristics are the narrowing of the bandwidth together with some lowering of the radiation efficiency even though the intrinsic material loss (tande ) is low. Numerous chip antenna designs have been described whereby a highpermittivity ceramic material is associated with a meander-line23 or helical configurations.24 The conducting lines can be readily printed onto the outer surface of the ceramic material. An example of a helical construction is illustrated in Figure 36-9a and its deployment as an antenna embedded in a handset is shown in Figure 36-9b together with radiation patterns. The typical ceramic chip antenna is likely to be too small to permit resonant modes within its interior. In contrast, the dielectric resonator antenna (DRA)25 relies on a mode being excited, which can be achieved by a small probe or coupled slot. Although the concept of a DRA was introduced by Richtmyer26 in 1939, the practical versions were not investigated until the 1980s, no doubt due to the new availability of high-permittivity, low-loss, temperature-stable ceramic materials at that time. The feasibility of embedding a DRA within a handset has been investigated27 but single mode operation has the characteristic narrow bandwidth which

(a)

(b) FIGURE 36-9 (a) Constructional view of helical ceramic chip antenna11 (Courtesy of Murata Manufacturing Co.) and (b) the chip antenna embedded in a handset and the radiation patterns



CHAPTER THIRTY-SIX

limits its application. In another development28 a quadrifilar-type antenna structure is plated onto a high-permittivity dielectric cylinder. A balun structure at the base isolates the antenna from the GP, and there is a radiation null at broadside, thus reducing radiation exposure to the head. The bandwidth as anticipated is very narrow. The antenna is illustrated in Figure 36-10. The demand for multiband handset antennas has prompted designers to develop DRAs that have more than one mode excited, thus providing bandwidth windows that correspond to the required cellular bands. This is not difficult if the resonator is not restricted in size, but for handset installation the DRA thickness must be < 10 mm, while the handset width and limited space allocation place severe constraints on the other dimensions. Such a multimode DRA has been described29 and the mode generation is governed by the excitation mechanism, which is commonly a coupling strip plated on the dielectric surface. Additional conductors can be plated on the dielectric surfaces to provide additional bands, and the nature and location of the conductors are mainly determined by experiment and/or simulation. The electrical compactness of DRAs, at least for the higher mobile bands, allows provision of balanced pairs12 to reduce currents in the handset GP and hence reduce hand and other environmental effects. An example is given in Figure 36-11, which

FIGURE 36-10 Dielectrically loaded, balanced quadrifilar-type antenna (Courtesy of Sarantel Ltd.)

(a)

(b)

(c)

FIGURE 36-11 (a) Balanced ceramic antenna on GP30 (Courtesy of Antenova Ltd.); (b) The 900-MHz bandwidth characteristic; (c) The 1800- and 1900-MHz bandwidth characteristics



36-13

covers the 900-, 1800-, and 1900-MHz bands.30 The 900-MHz band remains unbalanced while the two higher bands are balanced. It is also claimed that there is less radiation exposure to the head. The use of ferrite material to reduce the electrical size of wire monopoles and other antennas has received some attention31 over the past 60 or so years but has remained a specialized technique for a few applications. More recently32 its application to DRAs has emphasized the enhancements in antenna performance that can be achieved, these being a significant increase in bandwidth and the radiation efficiency. Best performance is obtained when there is approximate equality between the real parts of the material’s complex permittivity and permeability. Theory and simulations33 have established the performance benefits of applying this technique to handset antennas. An example34 is given in Figure 36-12a, which shows a ferrite-loaded monopole antenna covering the four bands 1800 MHz (GSM), 1900 MHz (GSM), 2100 MHz (UMTS), and 2450 MHz (Bluetooth). The conducting rings facilitate the adjustment of the band shape. The quadband bandpass characteristics are shown in Figure 36-12b together with a previous triband characteristic which did not cover the UMTS band.

(a)

(b) FIGURE 36-12 (a) View of ferrite antenna on GP (after M. I. Kitra et al34 © IEEE 2007) and (b) the bandwidth characteristics


Mobile Handset Antennas 36-14 TABLE 36-2

CHAPTER THIRTY-SIX

Performance of Ferrite Handset Antenna (after M. I. Kitra et al34 © IEEE 2007) tande = 0.03 tandm = 0.03

tande = 0.0001 tandm = 0.06

e′

6

6.38

m′

6.6

6.18

4.5

4.5

Achieved

Achieved

35.05, 32.6, 15.15, and 22.34

38.5, 34.9, 19.2, and 22

Monopole length (mm) Coverage at 1800, 1900, 2100, and 2450 MHz at −6 dB level Efficiency (%) with the head present at 1800, 1900, 2100, and 2450 MHz, respectively

A summary of the simulated performances for various material properties is given in Table 36-2. The complex permeability and permittivity of the material are defined as mr = m ′ + jm ″ ;

tandm = m ″/m ′

er = e ′ + je ″ ;

tande = e ″/e ′

(36-5)

Simulation using a commercially available ferrite material designed for screening purposes with m ′ = 2, tandm = 0.06, e ′ = 24.1, and tan de = 0.0001 established that a quad-band performance was obtainable. The low radiation efficiency may be improved with the development of a ferrite material specifically for antenna rather than screening applications.

36.6 SAR MEASUREMENT AND MINIMIZATION This section examines the effects on the user’s health when using a mobile communications device. Methods for reducing the radiation absorbed in the head are also presented. Radiation Exposure The public has been concerned for many years about the electromagnetic radiation effects from microwave ovens, overhead power lines, and household electrical installations even though no established evidence of harm has emerged. Non-ionizing electromagnetic waves in the microwave and lower frequency bands create thermal effects in tissues when the water molecules and dissolved ions are made to vibrate. The water content is therefore an indication of the level of absorption. Nonthermal effects are due to the interaction of the applied field with the molecules where it is considered that the latter align themselves along the electric field to minimize potential energy.35 Elegant theories and mathematical models have been published, for instance, “perturbation of DNA due to soliton waves,” “effects on body organs and glands,” and “acceleration and retardation of seed growth at specific millimeter wave frequencies.” There remains a large amount of literature globally on this topic, but the results are often ambiguous and no clear evidence appears to have emerged so far.36



36-15

It is not surprising, therefore, that the astronomical growth of mobile phone usage has been accompanied by much public concern, particularly as the handset is held close to the head and earlier analog handsets could dissipate a watt of power in the brain. Despite extensive worldwide research and strong support from governments for over a decade, no evidence of a heath risk has been uncovered so far. Protests about the sitting of base station masts and power lines are continuing, but at the present time the situation with mobile phones remains calm and stable, with most users aware that minimizing handset use is desirable, particularly for children. Text messaging has escalated for children and eases the exposure problem. Manufacturers for their part have reduced exposure levels with the introduction of digital systems and continue to be responsibly mindful of the situation. On the research front, the electromagnetic exposure to the head and other body regions continues to be of major interest, such as the development of low SAR handsets (as described later in this section), the effect of metal spectacles,37 body-worn jewelry,38 metal spectacles on the mucous membranes,39 the exposure to passengers in automobiles when using mobile phones,40 to mention but a few. Below we give background details of some current work in this field. Radio Frequency Dosimetry With the concern about mobile phone exposure to the human body and particularly the head, it was evident that a regulatory process was urgently needed. The radiation exposure to any part of the human body is now assessed by the Specific Absorption Rate (SAR), which is defined by SAR = (s / 2r ) | E |2 W/kg

(36-6)

where s = tissue conductivity, r = tissue density, and | E |2 = the square of the electric field intensity in the tissue. Background details are given in Fujimoto and James11 (Chapter 6, Section 6.5, and Chapter 7). Radio frequency dosimetry in the human body is very complex due to many factors that affect the absorption rate in tissues. Safety guidelines recommend that the 1g averaged peak SAR should not exceed 1.6 W/kg and the whole body averaged peak SAR should be less than 0.08 W/kg. The 1g averaged peak SAR is preferred because it represents local variations more accurately. Each mobile phone user has a usage profile ranging from those who use their mobile phone occasionally to those in a busy occupation necessitating many hours of continuous phone use each day. An idea of how SAR needs to be interpreted is given in Fujimoto and James11 (Chapter 6, Section 6.5). To determine the likely SAR for a mobile phone, it must be measured in as near a natural scenario as possible together with a model of the human head composed of representative artificial tissue. Table 36-3 summarizes the electrical properties of some common tissues. Purpose-built equipment is available to assist in standardization, but computer modeling has played a major role in rapidly evaluating different human body and handset situations; these aspects are detailed below. Measurement of SAR Although computer modeling has proliferated, the direct practical measurement of SAR in a phantom head and body is regarded as an important complementary tool. Handset manufactures have now built up their bespoke practical measurement equipment over several years with phantom models comprised of tissue simulating fluid with the appropriate electrical parameters as listed in Table 36-3. Elaborate three-dimensional electrically small



CHAPTER THIRTY-SIX

TABLE 36-3

Electrical Parameters of Human Body Tissues Relative Permittivity e ′

Conductivity s (S/m)

300

45.3

0.87

450

43,5

0.87

835

41.5

0.90

900

41.5

0.97

1450

40.5

1.20

1800

40.0

1.40

1900

40.0

1.40

2000

40.0

1.40

2450

39.2

1.80

3000

38.5

2.40

Frequency (MHz)

probes are positioned within the phantom fluid by sophisticated control systems, enabling the SAR to be rapidly assessed within a few minutes. The homogeneous tissue fluid enables a bulk SAR assessment to be performed but cannot represent the layering of different tissues and the presence of organs and glands. It does, however, allow the rapid comparison of different handset antennas and the effects of different handset orientations and separation distance from the body region. Most of the measurements are concerned with the absorption in the head, with or without a phantom hand model, and the variation in SAR for the same handset can be up to 100 percent depending on the measurement conditions.41 Consumer groups have published league tables of SAR head absorption for commercially available mobile phones but its influence on customer purchasing preference is not very apparent, maybe because of the strong fashion element associated with phones, particularly with younger people. A photograph of a measurement system is given in Figure 36-13, showing the phantom in the forefront undergoing probe measurements in its right ear region. Details of the probes and control apparatus are given in Fujimoto and James11 (pp. 338–340). There is also some demand from universities and research establishments for ready-made commercially available SAR measurement equipment. One such example is the SPEAG Dosimetric Assessment System (DASY 4), which provides a sixaxes robotic arm scanning the inside of FIGURE 36-13 View of phantom measurement a phantom head with facial features that equipment (after K. Fujimoto and J. R. James11 42 © Artech House, Inc. 2001) are filled with tissue simulating liquid.



FIGURE 36-14

36-17

DASY 4 SAR measurement system by SPEAG

The system is computer controlled, enabling measurement of SAR to be performed in a short time. A photograph of the equipment is shown in Figure 36-14. Techniques for Minimizing SAR A variety of earlier attempts sought to reduce the handset antenna radiation to the hand and head regions. A pull-up metal mesh shield around the handset case was manufactured but, as might have been expected, it interfered and reduced the handset transmitted and received power. This idea has very recently been revisited with success using a metamaterial shield43 and a view of the latter in position on the handset case is shown in Figure 36-15. A way of reducing SAR in the head has been described44 that is based on the presence of a ferrite material screen. The research was simulated using FDTD with a detailed head model. It is claimed that provided the handset GP currents are not perturbed, then the SAR in the head is reduced without a reduction in radiation efficiency. An early attempt to reduce the head SAR by creating a radiation pattern null in the direction of the head involved a slot/patch combination.45 A layered spherical head constituted the simulation model. The antenna was bulky and unlikely to be applicable to present-day handset requirements, particularly with the demand for single multiband antenna operation. Computer modeling is now a foremost tool associated with the minimization of SAR, particularly for the head region. The models have ranged from simple homogeneous cube blocks, spheres, and layered spheres to complex, more-realistic head models exhibiting the heterogeneous regions of the head and brain. These anatomical models are very detailed and thus escalate computer processing requirements. Some models can be freely downloaded on the Web. Two different anatomical models46,47 are illustrated in Figures 36-16a and b.



CHAPTER THIRTY-SIX

FIGURE 36-15 Metamaterial shield on a handset case (after G.Goussetis et al43 © URSI 2005)

Much research has been carried out on the comparison of different head models and it has been deduced that even though a heterogeneous head model will affect the SAR distribution within the head, it has little effect on the maximum integrated SAR, which is determined mainly by the antenna itself and its distance from the phantom.48–50 It is concluded that the maximum averaged SAR is lower in most cases in a heterogeneous head model compared to a homogeneous head model and that a spherical head can be used to give a reasonable, slightly pessimistic assessment.50–52

(a)

(b)

FIGURE 36-16 (a) Visible human head (after U. Tiede et al © IEEE 1996); (b) 1.1-mm resolution image of human head with 26 tissue types (after W. Whittow47 © Sheffield University 2004) 46




(a)

(b)

FIGURE 36-17 (a) Far-field gain patterns of the half-wavelength antenna with and without the sixlayered spherical head (after K. Fujimoto and J. R. James11 © Artech House, Inc. 2001); (b) 1g averaged and unaveraged SAR distributions along z axis at 900 MHz (BR = brain, C = cerebrospinal fluid [CSF], D = dural, BN = bone, F = fat, and S = skin)

Currently, the IEEE Standards Coordinating Committee 34, Subcommittee 2, Work Group 2 is proposing new FDTD computational techniques for determining the SAR in the human body.53 It is expected that this standard will provide better guidelines for simulations involving heterogeneous heads.54 A layered spherical head has been investigated using the eigenvalue expansion method, and an interesting example of the results is shown in Figures 36-17a and b (Fujimoto and James11 pp. 373–374). The result demonstrates that the absorption in the skin exceeds that in the CSF at higher frequencies. The majority of head model simulation research appears to mainly involve the use of PIFAs and other handset antennas that couple strongly into their environment, as described earlier in this chapter. The head SAR levels can therefore be near the recommended SAR limits.19 The ferrite-loaded handset antenna described in Section 36.5 represents a significant advance in techniques for reducing the head SAR, and the reductions in SAR are very significant. Table 36-4 gives details of the SAR in the head for this ferrite antenna. A spherical homogeneous head model was used in this research. TABLE 36-4

SAR in the Head for Ferrite Handset Antennas35

Handset Antenna

Frequency Band Coverage (MHz)

SAR in the Head (10g [W/kg])

Ferrite antenna (Figure 36-12)

1800, 1900, 2100, and 2450, respectively

0.076, 0.11, 0.06, and 0.0006, respectively



CHAPTER THIRTY-SIX

The excellent low SAR in the head is brought about by the use of an antenna monopole excitation probe, the orientation of the probe, and the use of ferrite material, and, for the balanced version, the deployment as well of two coupled antennas in phase opposition. As far as we are aware, these ferrite handset antennas are probably the only way of effectively reducing the SAR in the head while maintaining good prospects for realistic manufacturing potential.

36.7 PROVISION FOR GPS AND BLUETOOTH The provision for GPS and Bluetooth services in recent years was the commencement of adding value to mobile phones. Handset manufacturers are clearly influenced by the relative ease of embedding additional software processing chips within the handset. Additional bulky sensors are invariably required to input, output, and display the information, and this has not been an insurmountable obstacle for TV reception and camera phones where the display panel and camera can protrude somewhat outside the case. The clamshell handset also eases the space problem. The power of the chips continues to increase significantly while chip prices continue to fall. In keeping with this trend, handset manufacturers are deterred from increasing the number of antennas within the handset, both for cost reasons and because of the acute shortage of space. Therefore, we can expect to have only one multiband handset antenna that will encompass the new bands in addition to the established original bands. The isolation between adjacent bands needs to be small to avoid additional software filtering. We have already shown in this chapter that some of the multiband antennas are in fact able to include the Bluetooth transceiver band at 2450 MHz. One thing in the designer’s favor is that at this higher frequency, the Bluetooth antenna is less electrically small with a quarter-wave length of 3.06 cm. For GPS receiver provision in the L1 1575.42-MHz band, the antenna size limitation is more critical with a quarter-wave length of 4.76 cm. Both GPS and Bluetooth are low-level signal devices and consequently they involve no SAR issue. A way of reducing the GPS antenna length has already been demonstrated19 and involves an IFA plated onto a dielectrically loaded PIFA. The dielectric loading condenses the IFA fields and hence the antenna length. This example is illustrated in Figures 36-18a and b, showing a view of the dielectric loaded PIFA and the isolation between the PIFA and the IFA.

(a)

(b)

FIGURE 36-18 (a) View of dielectric loaded PIFA with mounted IFA (after Z. Li and Y. Rahmat-Samii19 © IEEE 2005); (b) Isolation between the PIFA and IFA antennas



36-21

How handset manufacturers continue to provide for GPS and Bluetooth depends very much on their own preferred handset and antenna designs, but the examples shown in this section illustrate the feasibility of adding these two services to mobile phones.

36.8 MEASUREMENT OF HANDSET ANTENNAS The mobile phone handset is a remarkable electronic product: it is packed with communication circuitry that must fit in the handset case, and it must be affordable by the public at large. The complexity of handsets has demanded new thinking about not only how to measure the circuit functions during the manufacturing process but also its overall performance when assembled. All of this must be accomplished in a cost-effective way. The situation is further complicated by the continually increasing processing speeds and the associated high-frequency EMC difficulties within the handset. Instrument manufacturers have responded in earnest and numerous suites of sophisticated instruments are now available to handset manufacturers to automate their component and subsystem production lines. Measuring the performance of the finished product is a somewhat less-well-defined task because it has to embrace the relationship of the handset with its environment. This includes not only radio propagation effects in the everyday urban scenario but, as already discussed in this chapter, the absorption of waves by the operator’s body and, in particular, head. To what extent these issues are addressed will depend on the handset manufacturers but the cost of testing will be a major consideration and it has to be assumed that some sampling of the manufacturing output will take place. The performance of an antenna in cellular operation can be defined and related to a particular route. Multipath reflections from moving vehicles, people, airplanes, doors in buildings, etc., create random fluctuations in the propagating radio wave polarization and signal strength. For measurement purposes, the wave polarization can be resolved into vertical polarization (VP) and horizontal polarization (HP), so upon moving the mobile handset along a particular urban route, the mean powers Pv and Ph, respectively, associated with these polarized components can be measured. The antenna will not be able to capture all the polarized power, and Pr < (Pv + Ph), where Pr is the mean antenna power that would have been received over the same route, in the same time period. The antenna performance is defined55 by its mean effective gain (MEG), where MEG = Pr / (Pv + Ph) MEG =

(36-7)

2π π

∫ ∫ [ Pv / (Pv + Ph ) Gθ (θ , φ ) Pθ (θ , φ ) + Ph / (Pv + Ph ) Gφ (θ , φ )Pφ (θ , φ )] .sin θ dθ d φ 0 0

(36-8)

where (q, f ) are spherical angular coordinates, Gq (q, f) and Gf (q, f) are the q and f components of the antenna power gain pattern, respectively, and Pq (q, f) and Pf (q, f) are the q and f components of the angular density functions of the incoming plane waves, respectively. The recent trend toward extreme handset compactness and internal integrated antennas does not allow the handset antenna to be evaluated in isolation from the handset itself. In principle the MEG assessment could be applied to the handset itself, but it is clearly not affordable as a production test. Measurements on handset reception and transmission are also carried out56 in a controlled environment, such as an anechoic chamber, but the process can again be very time-consuming and indeed costly.



CHAPTER THIRTY-SIX

The use of a reverberation chamber to test antennas in Rayleigh fading propagation conditions is well established, but its recent application57 to mobile handset testing appears to offer manufacturers an efficient, low-cost method of assessing handset performance. An anechoic chamber requires highly absorbent walls, but a reverberation chamber is a large, metal-walled cavity capable of supporting many modes, which are then greatly perturbed by rotating reflectors within the excited chamber. This mode-stirring action is random enough to create a Rayleigh distributed transfer function between the receiving and transmitting antennas inside the chamber: in this application, these are the radiating handset and the receiving wall probe. The Rayleigh distribution model is a useful approximation of realistic mobile propagation conditions, and a demonstration chamber has been described57 that has many attractive features. For instance, a phantom tissue head model can be placed in close proximity to the handset under measurement to enable the actual operating radiation efficiency to be obtained. The chamber can also test the addition of diversity functions and the newly emerging MIMO systems. The installation costs and operating times are low and compatible with handset manufacturing requirements.

36.9 SATCOM HANDSET ANTENNAS Unlike cellular phones, satcom phone systems have had a very uncertain start due to the inadequate public takeup, which has led to financial difficulties for manufacturers. The systems have, however, survived in a leaner, more responsive form and a stable market has been developed. Around 1998 the systems under development were Iridium, Globalstar, ICO, and Thuraya. A summary of their salient technical details are listed in Mobile Antenna Systems Handbook11 (p. 582). Iridium, Globalstar, and ICO are based on LEO, LEO, and MEO constellations, respectively, while Thuraya has a GEO satellite. Phased-array antennas directing spot beams to earth are used on all systems. Global roaming is provided by satcom phones, enabling the user to communicate from any place on earth. The user community includes military, business people, politicians, emergency services, those working in remote parts of the world or at sea, and so on, and this service differs greatly from that of the cellular phone system. A point in favor of the handset manufacturers is that the handset cases are less restricted in size than cellular phones, cost is not such a limitation, and the handset antenna can protrude well above the handset case. It was evident58 in 1998 that the perceived market may not be adequate although the global roaming capability would attract support from the defense sector. These satcom systems introduced new problems for handset antenna designers because of the low signal margins and the need for circular polarization.58 The quadrifilar helix antenna (QHA) is extensively used, and reduced-diameter versions have been developed, giving some loss of bandwidth capability while ensuring compact user convenience. QHAs are fitted with a balun to feed the spiral arms from an unbalanced feeder, and various designs exist. A more conventional QHA is shown in Figure 36-19. The current Iridium 9505A handset has a size of 158 × 62 × 59 mm and has many added-value features including fax and data; an auxiliary handset antenna FIGURE 36-19 Quadrifilar helix antenna (QHA) 59 is also supplied. The Globalstar SAT600 (after C. Kilgus © IEEE 1964)



36-23

FIGURE 36-20 Thuraya 7101 handset11 (Courtesy of Thuraya)

handset has a size of 177 × 58 × 39 mm and offers a dual-mode facility combining cellular GSM900 and the satcom channel. Add-on facilities for the Internet are also available. The Thuraya 7101 handset, shown in Figure 36-20, offers data, fax, and GPS facilities; its size is 146 × 53 × 27 mm. As already mentioned, the QHA is constructed so that it can be pulled up out of the handset’s case for use, thus giving it an elevated position clear of the user’s head. Construction details for the Thuraya handset antenna are given in Figure 36-21, which shows the narrowdiameter QHA with the GSM NMHA mounted on its tip.

FIGURE 36-21 Constructional details of the Thuraya handset antenna11 (Courtesy of Allgon)



CHAPTER THIRTY-SIX

(a)

(b)

FIGURE 36-22 (a) Thuraya handset antenna passband characteristics11 (Courtesy of Allgon); (b) Thuraya handset antenna’s LHCP and RHCP elevation radiation pattern

The Thuraya earth-to-space link is 1626.5–1660.5 MHz with a space-to-earth link of 1525.0–1559.0 MHz. The Thuraya passband characteristics are shown in Figure 36-22a, and the antenna elevation radiation pattern is shown in Figure 36-22b.

36.10 FUTURE TRENDS The mobile communications market will continue to grow fueled by user demand for increased functionality, aesthetics, smaller handsets, longer battery life, and ubiquitous access. Some future trends are presented next. ●

●

●

●

●

In the immediate future, handset manufacturers will continue to add value to handsets by way of low-cost processing software. This of course will be dependent on the consumer takeup, which from past experience is dictated by individual requirements, fashion, and quite likely other serendipity factors. When the new services require extended band coverage, handset antenna innovation will continue to be much in demand. The system design approach will continue to be much in evidence in handset antenna design. This takes a holistic view of how all the components function together even if their individual performance may be degraded. When compared with conventional antennas, present-day handset antennas, with their low efficiency and proximity to other components, would have been seen as very poor radiators a few decades ago. Increasing the data rates of mobile phones will continue to receive research attention as more services are sought. New processing algorithms and techniques are being addressed, including the use of multiple antennas and quantum computing.60 The clamshell type of mobile handset has provided more space on handsets for facilities and displays while allowing the handset to be compacted when not in use, but additional antenna design problems are invoked.61 Other types of handset case designs can be expected as services increase. The quest for a single multiband handset antenna that will cover all the band requirements for new services is inhibited by well-established fundamental antenna action whereby antenna resonances are necessary to achieve good matching, bandwidth, and radiation efficiency. Resonance is determined by the length of current paths and sets constraints



36-25

on antenna size even if material loading and other compacting techniques are used, as described in this chapter. A future breakthrough in antenna design cannot be ruled out. Much attention has been focused recently on how materials can be synthesized using metamaterials62 but to date the cell sizes of the latter remain too large for the synthesis of an electrically small antenna.

36.11 SYMBOLS A

radius of minimum sphere

DRA

dielectric resonator antenna

GP

ground plane

GPS

global positioning system

IFA

inverted F antenna

K

free-space wave number

LEO

low earth orbit

MEG

mean effective gain

MEO

medium earth orbit

NMHA

normal mode helical antenna

ha

antenna radiation efficiency

hm

matching network radiation efficiency

PIFA

planar inverted F antenna

Q

Q factor

Qa

antenna Q factor

Qm

matching network Q

QHA

quadrifilar helix antenna

SAR

specific absorption rate

UMTS

universal mobile telephone system

tandm

permeability loss tangent

tande

permittivity loss tangent

mr = m ′+ jm ″

complex permeability

er = e ′ + je ″

complex permittivity

REFERENCES 1. H. A. Wheeler, “Fundamental Limitations of Small Antennas,” IRE Proc., vol. 35 (December 1947): 1479–1484. 2. L. J. Chu, “Physical Limitations of Omni-directional Antennas,” J. App. Phys., vol. 19 (December 1948): 1163–1175.



CHAPTER THIRTY-SIX

3. R. E Collin and S. Rothschild, “Evaluation of Antenna Q,” IEEE Trans. Antennas and Propagat., vol. AP 12 (January 1964): 23–27. 4. H. A. Wheeler, “The Radian Sphere Around a Small Antenna,” IEE Proc., vol. 47 (August 1959): 1325–1331. 5. J. S. McLean, “A Re-examination of the Fundamental Limits on the Radiation Q of Electrically Small Antennas,” IEEE Trans. Antennas and Propagat., vol. 44, no. 5 (May 1996): 672–676. 6. G. Goubau, “Multielement Monopole Antennas,” Proc. ECOM_ARO Workshop on Electrically Small Antennas (May 6–7, 1976): 63–67. 7. R. C. Hansen, “Fundamental Limitations in Antennas,” Proc. IEEE, vol. 69 (February 1981): 170–182. 8. A. D. Yaghjian and S. R. Best, “Impedance, Bandwidth and Q of Antennas,” IEEE Trans. Antennas and Propagat., vol. 53, no. 4 (April 2005): 1298–1324. 9. D. M. Bolle and M.D. Morganstern, “Monopole and Conic Antennas on Spherical Vehicles,” IEEE Trans. Antenna. and Propagat., vol. 17 (1969): 477–484. 10. A. W. Rudge, K. Milne, A. D. Olver, and P. Knight (eds.), The Handbook of Antenna Design, vol. 2 (London: Peter Peregrinus, IEE, 1983): 727. 11. K. Fujimoto and J. R. James, Mobile Antenna Systems Handbook, 2nd Ed. (Boston: Artech House 2001). 12. H. Morishita and K. Fujimoto, “A Balanced-Fed Loop Antenna System for Handset,” IEICE Trans. Commun., vol. E82-A, no. 7 (1999): 1138–1143. 13. S. Kingsley, “Advances in Handset Antenna Design,” http://rfdesign.com/mag/radio_advances_ handset_antenna/ (May 1, 2005): 16–22. 14. K. Wong, Planar Antennas for Wireless Communications (Hoboken, NJ: Wiley-InterScience, 2003): 53. 15. K. Fujimoto, A. Henderson, K. Hirasawa, and J. R. James, Small Antennas (Baldock, UK: Research Studies Press, 1987): 7–9. 16. R. C. Hansen, “Optimum Inductive Loading of Short Whip Antennas,” IEEE Trans., vol. VT-24 (1975): 21–29. 17. J. R. James and A. Henderson, “Investigation of Electrically Small VHF and HF Cavity-type Antennas,” Proc. Int. Conf. Ant. Propagat. (1978): 322–326. 18. J. R. James and P. S. Hall, Handbook of Microstrip Antennas, vols. 1 and 2 (London: Peter Peregrinus, IEE, 1989): 25, 1104. 19. Z. Li and Y. Rahmat-Samii, “Optimization of PIFA-IFA Combination in Handset Antenna Design,” IEEE Trans. Antennas and Propagag., vol. 53, no. 5 (May 2005): 1770–1778. 20. Y.-X. Guo, M. Y. W. Chia, and Z. N. Chen, “Miniature Built-In Multiband Antennas for Mobile Handsets,” IEEE Trans. Antenna. and Propagat., vol. 52, no. 8 (August 2004): 1936–1944. 21. G.-Y. Lee and K.-L. Wong, “Very-Low-Profile Bent Planar Monopole Antenna for GSM/DCS Dual-band Mobile Phone,” Microwave and Optical Technol. Lett., vol. 34, issue 6 (September 2002): 406–409. 22. P.-L. Teng and K.-L. Wong, “Planar Monopole Folded into a Compact Structure for Very-LowProfile Multiband Mobile-Phone Antenna,” Microwave and Optical Technol. Lett., vol. 33, issue 1 (April 2002): 22–25. 23. J. Lee, C. Jeon, and B. Lee, “Design of Ceramic Chip Antenna for Bluetooth Applications Using Meander-Lines,” IEEE Antennas and Propagat. Symposium (2002): 68–71. 24. C. Lin, Y. Cheng, and H. Chuang, “Design of a 900/1800 MHz Dual-Band LTCC Chip Antenna for Mobile Communication Applications,” Microwave Journal (January 2004): 78, 80, 82, 84, 86. 25. K. M. Luk and K. W. Leung, Dielectric Resonator Antennas (Baldock, UK: Research Studies Press Ltd, 2003). 26. R. D. Richtmyer, “Dielectric Resonators,” Jour. Appl. Phys., vol. 10 (June 1939): 391–398. 27. M. T. K. Tam and R. D. Murch, “Compact Sector and Annular Dielectric Resonator Antennas,” Trans. Antennas and Propagat., vol. 47, no. 5 (May 1999): 837–842.



36-27

28. O. Leisten et al., “Miniaturised Dielectrically Loaded Quadrifilar Antenna for Global Positioning System (GPS),” Electronics Letters, vol. 37, issue 22 (October 2001): 1321–1322. 29. G. Ba-Babik, C. Di Nallo, and A. Farone, “Multi-mode Dielectric Resonator Antenna of Very High Permittivity,” IEEE Int. Conf. Antennas and Propagat. (2004): 1383–1386. 30. Antenova Ltd., GSM RADIONOVA Radio Antenna Module Data Sheet, February 4, 2005. 31. J. R. James and A. Henderson, “Electrically Short Monopole Antennas with Dielectric or Ferrite Coatings,” Proc. IEE, vol. 125, no. 2 (1978): 793–803. 32. J. R. James, R. Chair, K. M. Luk, K. M. Chow, K. W. Leung, and J. C. Vardaxoglou, “Influence of Magnetic Material on Dielectric Resonator Antenna Excitation,” IEE Proc. Microw. Antennas and Propagat., vol. 151, no. 4 (August 2004): 293–298. 33. M. I. Kitra, P. McEvoy, J. C. Vardaxoglou, and J. R. James, “A Theoretical and Simulation Study of Dielectrically Loaded Antennas and Their Contribution Towards Low SAR,” Inter. ITG Conf. on Antennas (INICA) (September 2003): 245–248. 34. M. I. Kitra, C. J. Panagamuwa, P. McEvoy, J. C. Vardaxoglou, and J. R. James, “Low SAR Ferrite Handset Antenna Design,” IEEE Trans. Antenna and Propagation, vol. 55, no. 4 (April 2007): 1155–1164. 35. A. de Salles, “Biological Effects of Microwave and RF,” Proc. SBMO/IEEE MMT-S IMOC ’99, (April 1999): 51–56. 36. R. Goldberg, “Literature Resources for Understanding Biological Effects of Electromagnetic Fields,” EMF-link Multitimedia Resource, http://infoventures.com/emf/top/lit-rev.html (January 1996). 37. W. G. Whittow and R. M. Edwards, “A Study of Changes to Specific Absorption Rates in the Human Eye Close to Perfectly Conducting Spectacles Within the Radio Frequency Range 1.5 to 3.0 GHz,” IEEE Trans. Antennas and Propagat., vol. 52, no. 12 (2004): 3207–3212. 38. N. A. Samsuri and J. A. Flint, “On the Effect of Jewellery Rings on Specific Absorption Rate (SAR),” Proc. Loughborough Antennas and Propagat. Conf. (2006): 421–423. 39. W. G. Whittow and R. M. Edwards, “A Study of Changes to Specific Absorption Rates in the Mucous Membrane Close to Perfectly Conducting Spectacles Within the Radio Frequency Range 0.8 to 2.6 GHz” Proc. Loughborough Antennas and Propagat. Conf. (2006): 417–420. 40. D. O. Coy, D. M. Zakharia, and Q. Balzano, “Field Strengths and Specific Absorption Rates in Automotive Environments,” IEEE Trans. Vehicular Tech., vol. 48, no. 4 (1999): 1287–1303. 41. J. C. Lin, “Specific Absorption Rates (SARs) Induced in Head Tissues by Microwave Radiation from Cell Phones,” IEEE Antennas and Propagat. Mag., vol. 42, no. 5 (October 2000): 138–139. 42. DASY 4 Dosimetric Assessment System Manual, Schmidt and Partner Engineering AG, 2003. 43. G. Goussetis, A. P. Feresidis, G. K. Palikaras, M. Kittra, and J. C. Vardaxoglou, “Miniaturisation of Electromagnetic Band Gap Structures for Mobile Applications,” Radio Science, vol. 40, no. 6, RS6S04 (Nov. 2005). 44. J. Wang, O. Fujjwar, and T. Takagi, “Reduction of Electromagnetic Absorption in the Human Head for Portable Telephones by a Ferrite Sheet Attachment,” IEICE Trans. Commun., vol. E80-B, no. 12 (1997): 1810–1815. 45. H. Ruoss and F. M. Landstorfer, “Slot Antenna for Handheld Mobile Phones Showing Significantly Reduced Interaction with the Human Body,” Electron. Lett., vol. 32, no. 6 (1996): 513–514. 46. U. Tiede, T. Schiemann, and K. H. Höhne, “Visualizing the Visible Human,” IEEE Comput. Graphics Appl., vol. 16 (1996): 7–9. 47. W. G. Whittow, “Specific Absorption Rate Perturbations in the Eyes and Head by Metallic Spectacles at Personal Radio Communication Frequencies,” Ph.D. thesis, EEE Dept., University of Sheffield, UK, 2004. 48. S. Watanabe, M. Taki, T. Nojiama, and O. Fujiwara, “Characteristics of the SAR Distributions in a Head Exposed to EM Fields Radiated by a Hand-held Portable Radio,” IEEE Trans. on Microw. Theory and Tech., vol. 44, no. 10 (October 1996): 1874–1883. 49. A. Lee, H. Choi, B. Kim, H. Lee, and J. Pack, “Effect of Head Size for Mobile Phone Exposure on EM Absorption,” Proc. APMC2001 (2001): 384–387.



CHAPTER THIRTY-SIX

50. K. Kim and Yahya Rahmat-Samii, “Antennas and Human in Personal Communications: Applications of Modern EM Computational Techniques,” Proc. 12th Inter. Conf. on Microw. and Radar, vol. 4 (May 20–22, 1998): 36–55. 51. S. Khalatbanri, D. Sardaari, A. A. Mirzaee, and H. A. Sadafi, “Calculating SAR in Two Models of the Human Head Exposed to Mobile Phone Radiation at 900 and 1800 MHz,” Prog. in Electromag. Res. Symp., Cambridge, Mass., March 26–29, 2006. 52. K. Kim and Yahya Rahmat-Samii, “EM Interactions Between Handheld Antennas and Humans: Anatomical vs Multi-layered Spherical Head,” IEEE Trans. APS Conf. on Antennas and Propagation for Wireless Communications (November 1998): 69–72. 53. “Recommended Practice for Determining the Peak Spatial-Average Specific Absorption Rate (SAR) in the Human Body from Wireless Communications Devices, 30 MHz–6 GHz: General Requirements for Using the Finite Difference Time Domain (FDTD) Method for SAR Calculations,” IEEE 1528.1 draft standard, April 2006. 54. B. B. Beard et al, “Comparisons of Computed Mobile Phone Induced SAR in the SAM Phantom to That in Anatomically Correct Models of the Human Head,” IEEE Trans. Electro. Comp., vol. 48, no. 2 (May 2006): 397–407. 55. T. Taga, “Analysis for Mean Effective Gain of Mobile Antennas in Land Mobile Antenna Environments,” IEEE Trans. Veh. Tech., vol. VT-39, no. 2 (May 1990): 117–131. 56. H. Arai, Measurement of Mobile Antenna Systems (Boston: Artech House, 2001). 57. P. S. Kildal, “Characterisation of Small Antennas and Active Mobile Terminals in Rayleigh Fading by Using Reverberation Chamber,” Proc. Loughborough Antennas and Propagat. Conf. (April 2005): 234–239. 58. J. R. James. “Realising Personal Satcom Antennas,” IEE Electronics and Communications Eng. Jour. (April 1998): 73–82. 59. C. C. Kilgus, “Spacecraft and Ground Station Applications of the Resonant Quadrifilar Helix,” IEEE Ap-S Int. Symp. Digest (June 1964): 75–77. 60. R. Calderback, “Quantum Computing and Cellular Phones,” Lecture at The Royal Academy of Engineering, London, June 20, 2006. 61. B. S. Collins, “Improving the Performance of Clamshell Handsets,” Proc. Loughborough Antennas and Popagat. Conf. (2006): 7–12. 62. R. W. Ziolkowski and A. D. Kipple, “Application of Double Negative Materials to Increase the Power Radiated by Electrically Small Antennas,” IEEE Trans. Antenna. and Propagat., vol. 51, no. 10 (October 2003): 2626–2640.



Chapter 37

Broadband Planar Antennas for High-Speed Wireless Communications Zhi Ning Chen Institute for Infocomm Research, Singapore

CONTENTS 37.1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

37-2

37.2 SUSPENDED PLATE ANTENNAS . . . . . . . . . . . . . . . . . . . . . . . . . .

37-2

37.3 PLANAR MONOPOLE ANTENNAS . . . . . . . . . . . . . . . . . . . . . . . . . 37-11


Broadband Planar Antennas for High-Speed Wireless Communications 37-2

CHAPTER THIRTY-SEVEN

37.1 INTRODUCTION High data speeds and multiband operations of modern wireless communication systems have significantly increased the demand for broadband antennas that are capable of supporting such requirements. In general, the systems need low-cost solutions with desired performance in terms of impedance bandwidth, polarization, and gain. Owing to unique merits such as small volume or low profile, low manufacturing cost, and easy integration into planar circuits, planar antennas are playing important roles in various wireless communication applications. The planar antennas can usually be categorized in terms of radiation performance into microstrip patch antenna, suspended plate antenna (SPA), planar inverted-L/F antenna, and planar monopole/dipole antenna.1 Usually, they are attractive to antenna engineers due to their low profile and/or small volume. The changes in such antenna design are, in general, from the specific requirements of applications. For example, the microstrip patch antenna in its basic forms has a low profile, which is conducive to conformal design, but suffers narrow impedance bandwidth on order of 1 percent. In contrast, the planar monopoles usually have a high profile above a ground plane but enjoy broad bandwidth. Making a trade-off between antenna profile and impedance, as well as radiation performance, the SPAs are good options for fixed base stations in wireless communication systems, and planar monopoles/dipoles for mobile wireless terminals. Accordingly, a variety of techniques have long been developed to further enhance the broadband performance of the SPAs and planar monopoles. Due to the merits of acceptable performance, low profile, and, in particular, low manufacturing cost, the SPAs and planar monopoles have widely been applied in high-speed wireless communication systems. This chapter is divided into two parts, suspended plate antennas, and planar monopoles for high-speed wireless communications. The basic characteristics of the antennas will be introduced. The state-of-the art designs will be covered, and the important issues for the antenna applications will be addressed.

37.2 SUSPENDED PLATE ANTENNAS In this section, the SPAs will be introduced as base station antennas for modern broadband wireless connections such as high-speed wireless local area networks (WLANs) based on IEEE 802.11. The impedance and radiation characteristics of the SPAs are described first. Then, the techniques used to broaden the impedance bandwidth of the SPAs are addressed. After that, the techniques for enhancing the radiation performance of the SPAs are introduced. Finally, the applications of the SPAs to WLANs will be exemplified as a case study. Definition An SPA with a thick, low-permittivity dielectric substrate can be considered as a variation of microstrip patch antenna, as shown in Figure 37-1.1 The planar radiator may be of any shape and placed above a ground plane. The planar radiator may be supported by a dielectric substrate with very low permittivity. Besides air, the dielectric substrate may be low-loss dielectric foam with a relative dielectric constant, er of ~ 1. The dielectric substrate may be multilayered partially with air. The substrate thickness h is usually around 0.06 times the operating wavelength. The antenna may be fed in various ways. In Figure 37-1



BROADBAND PLANAR ANTENNAS FOR HIGH-SPEED WIRELESS COMMUNICATIONS

z Radiator

Radiator

w

Probe

l

h

d h

l

x

Ground plane

Ground plane

(a)

RF feed

(b)

FIGURE 37-1 Coordinate system and geometry of a probe-fed SPA above air: (a) three-dimensional view and (b) side view

the SPA is fed by a coaxial probe at a distance d away from the edge of the radiator. The operating frequency fo of the SPA can be estimated by fo =

c c = λ o 2 ε r (l + h )

(37-1)

where lo denotes the operating wavelength at the frequency fo, c is the speed of light (299,792,458 m/s), and er designates the relative dielectric constant. The length l, width w, distance d, and thickness h are in meters. The SPAs can be fed in many ways like microstrip patch antennas. Typically, the SPAs can be excited by a microstrip line, a probe, and an aperture-coupled feeding structure, as shown in Figure 37-2. Figure 37-2a describes the very popular feeding structure of the planar antennas, where a stripline is directly connected to the edge of the SPA.2 Different from the microstrip patch antenna with a thin dielectric substrate supporting the radiator, the stripline for the SPAs is usually suspended above a ground plane with a large thickness. The large spacing between the feeding stripline and ground plane causes a wide stripline of characteristic impedance of 50 Ω. For example, within a frequency range of 1–10 GHz, the width of a stripline is 5.1 mm for h = 1 mm and 10 mm for h = 2 mm. Therefore, a coaxial probe, namely the inner conductor of a coaxial connector, is used to vertically (not horizontally) excite the end of the stripline. As a result, the frequency-dependent input impedance of the probe is observed.

Radiator

Microstrip line

Radiator

Feed point

Microstrip line

Probe

Ground plane

Probe Ground plane

(a) FIGURE 37-2

Microstrip line

RF connector

Radiator

Ground plane

Ground plane

Ground plane

Aperture

Aperture

Dielectric substrate

Ground plane

(b)

(c)

Three types of feeding structures for SPAs




In Figure 37-2b, the SPA is fed by a coaxial probe attached to the radiator directly while its outer conductor is connected to the ground plane.3 This configuration greatly simplifies the feeding structure. Due to the large spacing between the radiator and ground plane, this is a proper solution for the SPAs. The feeding probe can also be used to support the radiator when no dielectric substrate is used below the radiator. However, both the stripline and coaxial probe feeds may cause narrow impedance bandwidth due to the narrowband impedance transition between the stripline or probe and radiators. Moreover, such asymmetric feeding structures degrade the radiation performance, especially at higher frequencies with high cross-polarization radiation levels and/or squinted radiation from the boresight axis (beam-squint effect). Figure 37-2c shows an electromagnetically coupled feeding structure,4 where a microstrip line centrally feeds the radiator through a nonresonant slot right below the radiator, which is cut from the ground plane. The symmetric configuration of the antenna to some degree alleviates the degraded radiation performance problems by suppressing the high cross-polarization radiation and the severe beam-squint effect. The nonattaching transition structure also improves the impedance bandwidth to some extent.

Important Features Consider the SPA and coordinate system shown in Figure 37-1 again. The radiator is a square with a length l = 58 mm. A coaxial probe feeds the radiator with a distance d away from the radiator edge. Figure 37-3 shows the input impedance of the probe when the thickness h is 5, 8, and 11 mm. The start and stop frequencies are 2 and 3 GHz, respectively. From the Smith chart in Figure 37-3, the variation of the input impedance for varying thickness h is clearly observed. When the thickness becomes larger, the input impedance becomes larger, in particular, with larger inductance. This is an important feature of the SPAs. Therefore, to achieve a broad impedance bandwidth, it is necessary to employ the techniques to compensate for the large inductance.

FIGURE 37-3 Input impedance of the probe-fed SPA shown in Figure 37-1 with d = 0 mm, l = 58 mm, and h = 5, 8, and 11 mm


Broadband Planar Antennas for High-Speed Wireless Communications BROADBAND PLANAR ANTENNAS FOR HIGH-SPEED WIRELESS COMMUNICATIONS

37-5

FIGURE 37-4 Return losses of the SPA shown in Figure 37-1 with l = 58 mm and h = 5, 8, and 11 mm.

Figure 37-4 shows the return losses of the SPA shown in Figure 37-1. The location of the feed point is optimized for good impedance matching as h = 5 mm. The achieved impedance bandwidths for –10-dB return loss are, respectively, 4.7 percent, 7.1 percent, and 6.2 percent at h = 5, 8, and 11 mm. The larger the thickness, the broader the impedance bandwidth. However, for the larger thickness, such as 11 mm, the good matching cannot be achieved by changing the position of the feedpoint, so that the achieved impedance bandwidth is not larger as expected. Therefore, the impedance-matching techniques should be applied for broad impedance bandwidth. Owing to the large thickness, the radiation performance of the SPAs is slightly different from that of the conventional microstrip patch antennas. Figure 37-5 shows the radiation patterns in both E and H planes at the lower-edge, center, and higher-edge frequencies of the impedance bandwidth. In the E planes, the cross-polarization radiation levels are very low and thus can be ignored. With the increase in operating frequency, the sidelobe levels greatly increase, and the beam-squint effect can be observed at the higher-edge frequency due to the asymmetric configuration of the antenna. In the H planes, the copolarized radiation patterns are hardly changed. However, the cross-polarization radiation levels increase by 5.4 dB when increasing the frequency. In short, the longer probe results in higher impedance, in particular, inductance. The larger antenna thickness offers potential for a broad impedance bandwidth but degrades the radiation performance with higher cross-polarization radiation levels in the H planes and higher sidelobe levels as well as the beam squinting in the E planes. As a result, it is necessary to develop the techniques to broaden impedance bandwidth and enhance the radiation performance in the applications of the SPAs in wireless communication systems. Techniques for Broad Impedance Bandwidth In the SPA design, the large thickness and low permittivity of the substrate provides the potential for a broad impedance bandwidth due to low antenna Q values. For example, the probe-fed SPAs in their basic forms with air as their substrate usually can achieve around 7 percent impedance bandwidth for –10-dB return loss.1 The achieved bandwidth is limited by the high impedance caused by the longer feeding probe. To further broaden the impedance bandwidth, the techniques for impedance matching have been developed and applied.1




(a)

(b)

(c) FIGURE 37-5 Radiation patterns for the probe-fed SPA shown in Figure 37-1 with l = 58 mm and h = 8 mm at (a) lower-edge frequency, 2.28 GHz; (b) center frequency, 2.36 GHz; (c) higher-edge frequency, 2.44 GHz

Figure 37-6 shows the techniques for broadening impedance bandwidths of the probefed SPAs, which are extensively used in wireless communication systems. Figure 37-6a shows the method to offset the large inductance caused by a longer probe.5 An annular slot is cut around the feedpoint so that the radiator is fed through capacitive coupling. The capacitive loading can also be introduced between the radiator and feeding probes in other ways.6–8 Figure 37-6b illustrates the technique to counterbalance the large inductance and achieve a broad bandwidth by slotting the radiator.9 The slot cut from the radiator acts as an impedance-matching network across a broad bandwidth. The slot can be of other shapes.10 Such a technique can be applied without any increase in the size of antenna and can easily be fabricated at a low cost; two factors that make the applications of planar antenna arrays appealing. The introduction of strong electromagnetic coupling between the feeding probe and radiator is an important and effective way to achieve broadband impedance matching.11 The idea has been applied to the SPAs with great success, as shown in Figure 37-6c. Due to the application of the L-shaped feeding probe, the impedance bandwidth can reach up to more than 36 percent.12



Slot

37-7

Slot

Probe

Probe

(b)

(a) Vertical Sheet

Probe

Probe

(c)

(d)

FIGURE 37-6 Techniques for broadening impedance bandwidth of the SPAs

The key to achieving the broadband impedance response of the SPA is the broadband transition between the feeding probe and the radiator. Figure 37-6d shows the other way to modify the transition, where a vertical conducting sheet is used between the feeding probe and the radiator.13 The vertical sheet basically acts as a broadband impedance-matching network. As an example, an SPA fed by an L-shaped probe is designed that operates at 2.4 GHz. The geometry of the SPA is shown in Figure 37-7a, where a rectangular stub (25×4 mm) forms the horizontal portion of the L-shaped probe and is vertically excited by a probe with a radius of 0.65 mm. By adjusting the length and height of the vertical portion of the L-shaped probe as well as the height of the radiator, the impedance locus forms a tight loop around the center of the Smith chart, as shown in Figure 37-7b. The strong electromagnetically coupled structure functions as a broadband impedance-matching network. Therefore, the well-matched bandwidth for –10-dB return loss reaches up to 25 percent, ranging from 2.04 GHz to 2.62 GHz.

50

4

50 6 12.5

7.5

mm Probe

Ground plane

(a) FIGURE 37-7

25

(b)

(a) Geometry of an L-shaped probe-fed SPA and (b) impedance in Smith chart




FIGURE 37-8 Radiation patterns for gain of the L-shaped probe-fed SPA.

Figure 37-8 shows the radiation patterns in both E and H planes at frequencies 2.3 and 2.6 GHz. With the increase in frequency, the copolarization sidelobe levels in the E planes increase by 5.2 dB. The beam-squint effect is observed at the higher frequency of 2.6 GHz. In the H planes, the cross-polarization radiation levels become high also by 5.2 dB at 2.6 GHz. The ratios of co- to cross-polarization radiation levels are reduced from 15.8 dB at 2.3 GHz to 10.4 dB at 2.6 GHz. Therefore, the L-shaped probe-fed SPA features a broad impedance bandwidth but suffers the degraded radiation performance. Techniques for Enhanced Radiation Performance For base station antennas in wireless communication systems, the pure radiation polarization and stable maximum radiation direction are often required to avoid possible interference between different wireless systems and to save radiation power. The techniques to enhance the radiation performance of the broadband SPAs have been developed. As mentioned in Broadband Planar Antennas: Design and Applications,1 the structurally asymmetric and electrically unbalanced antenna design usually suffers high cross-polarization radiation levels and beam-squint effects. Consequently, the solution to alleviate the degraded radiation performance problem is to modify the feeding structure of the SPAs to keep the antenna geometry symmetric and feeding structure balanced. Figure 37-9 demonstrates some of the solutions that can be easily applied to the SPA design with the enhanced radiation performance. First, a dual-probe feeding configuration is illustrated in Figure 37-9a.14 Two feeding probes are used to excite the symmetric radiator simultaneously. The probes are symmetrically located in the E plane with respect to the H plane so that the antenna configuration is kept geometrically symmetric. The out-ofphase RF signals from the two probes form an electrically balanced feeding structure. As a result, the currents that distribute symmetrically and in balance at the radiator effectively suppress the cross-polarization radiation levels in the H planes and beam-squint effect in the E planes. However, the dual-probe feeding scheme needs a broadband feeding network to form a feeding system with two anti-phase probes, although this can be eased in array design.15 To simplify the dual-probe feeding structure, a single-probe scheme is presented in Figure 37-9b.16 The radiator is fed by a probe through a half-wavelength strip. The probe



37-9

FIGURE 37-9 Feeding schemes applied to SPAs for enhancement of radiation performance: (a) dual-probe; (b) probe-fed half-wavelength strip; (c) dual-probe-fed center slot (d ) simplified dual-probe-fed center slot scheme.

vertically excites one of the ends of the strip. The horizontal portion of feeding strip is half-wavelength long at the center operating frequency. Another end of the strip is vertically attached to the planar radiator. The RF signals in the two ends are expected to be out of phase. There is strong electromagnetic coupling between the horizontal portion of the strip and radiator, which achieves a good impedance matching within a broad bandwidth of 20 percent. Due to the balance-like feeding configuration, the cross-polarization radiation levels are effectively reduced so that the ratio of co- to cross-polarization levels is higher than 20 dB across the whole bandwidth. The beam-squint effect in the E planes is, however, observed due to the asymmetric geometry caused by the feeding structure. A modified dual-probe feeding structure is used to enhance the radiation performance, as shown in Figure 37-9c, where the two out-of-phase probes are positioned around the center of the radiator, and a narrow slot is used to separate two feeding points.17,18 The separation between the two probes is small. Varying the length of the center slot can shift the resonant frequency and improve impedance matching. As discussed in Broadband Planar Antennas: Design and Applications,1 the current distribution at the radiator will be kept symmetric so that the cross-polarization radiation is effectively suppressed in all directions. Such a structure also needs a feeding network in the single element design similar to the conventional dual-probe scheme. A modified version was proposed and used as shown in Figure 37-9d, where a grounded pin is used to replace the second probe. The radiation performance of the SPAs with such a simplified feeding scheme is comparable with that shown in Figure 37-9c.19 Applications in High-Speed Wireless Communications As mentioned in previous sections, the SPAs feature broad impedance bandwidth and acceptable radiation performance. One more important merit is that the SPAs have a cost-effective design, because they can be readily fabricated with conducting sheets partially or completely without use of any expensive dielectric substrate. Such an advantage enhances the competence of such




an antenna technology in the massive wireless market. The application of the SPA in wireless communication systems, for example, the access point (AP) of WLANs, is exemplified with an actual product design. As is well known by now, wireless networks can offer many advantages over wired networks due to mobility. With wireless networks, users are free from the tether of an Ethernet cable at a desk. Also, the wireless networks are flexible, fast, and easy to use, deploy, and maintain. The WLAN is a mature option for wireless networks, which are based on the IEEE 802.11 protocol and widely used. One of the WLAN’s operating frequency bands (IEEE 802.11b) is the 2.4-GHz band (2.4–2.4835 GHz for North America and Europe; 2.471–2.497 GHz for Japan; 2.4465–2.4835 GHz for France; and 2.445–2.475 GHz for Spain). Thus, the antennas for the WLANs should cover the band of 2.4–2.5 GHz as well. Figure 37-10a shows a photo of one design that has been successfully applied in the APs of WLANs as a product. The SPA was formed by a piece of copper sheet and supported by two stands above a 100×100-mm ground plane at a distance of 7.5 mm, as shown in Figure 37-10b. The stands were symmetrically cut and bent from the radiator and separated by a center slot. One stand was grounded as a shorting pin and the other was fed by a 50-Ω microstrip line through the ground plane. Adjusting the length of the slot can change the operating frequency and improve impedance matching. The detailed optimized parameters are shown in Figure 37-10b. The antenna is embedded in a casing of the APs for indoor WLAN wireless access applications. Figure 37-11 shows the measured return loss and radiation patterns for the SPA. The bandwidth for –15 dB can cover the operating band of 2.4–2.5 GHz very well, as shown in Figure 37-11a. The measured radiation patterns in both the E and H planes at 2.4 and 2.5 GHz in Figure 37-11b and c show the gain in the boresight direction is higher than 8.5 dBi. The wider coverage of 76o in the H planes and slight tilt beam direction of 5–10o in the E planes are conducive to application of the SPA design in the indoor WLAN AP applications, where the APs are operating at the mode of point-to-multiple point. It is concluded that the suspended plate antenna is one type of broadband microstrip antenna but with many unique features such as larger antenna thickness, broader impedance bandwidth, more easily degraded radiation performance, and lower fabrication cost. A variety of the techniques for the enhancement of impedance and radiation performance have long been developed. The more cost-effective designs based on suspended plate antennas have been accepted by the massive wireless communication markets.

Ground plane Slot

1 2

Radiator 4

8 10

21

Slot

10 2 21 Slot

46

10 4

8

46

7.5

(a)

mm

(b)

FIGURE 37-10 Geometry of an SPA embedded in a WLAN access point operating at 2.4 GHz



(a)

37-11

(b)

(c) FIGURE 37-11 Measured results for the SPA applied in the WLAN AP applications (a) return loss; copolarized radiation patterns in E and H planes (b) at 2.4 GHz and (c) at 2.5 GHz

37.3 PLANAR MONOPOLE ANTENNAS In this part, the state-of-the-art planar monopole antennas will be introduced as embeddable terminal antennas for mobile ultrawideband (UWB) devices, for instance, wireless universal serial bus (USB) dongles. First, the state-of-the-art planar monopoles are briefly summarized. Unique design criteria for UWB antenna design are described. After that, the planar monopoles applied to wireless USB are introduced. The State of the Art In general, UWB antennas indicate frequency-independent designs, for instance, TEM horn antennas, as shown in Figure 37-12a.20–22 The antenna radiates linearly polarized TEM waves. Figure 37-12b shows a photo of a planar two-arm spiral antenna. The operating frequency range is determined by the inner and outer radii of the spiral. The radiated wave is a circularly polarized wave at the boresight. The spiral antenna is usually backed by a lossy or conducting cavity to improve impedance matching at low frequency and axial ratio by reducing the reflections from the end of the spiral arm. The phase center will change with operating frequency when a conical spiral antenna is used.23 Bi-conical antennas may be the earliest antennas used in




(a)

(c)

(b)

(d)

FIGURE 37-12 Frequency-independent antennas: (a) a TEM horn antenna; (b) a planar two-arm spiral antenna backed by a conducting cavity; (c) a bi-conical antenna; (d) two discone antennas

wireless communications because of their relatively stable phase centers and broad impedance bandwidths.24 Figure 37-12c displays an asymmetrical finite bi-conical antenna with a broad impedance bandwidths.25–27 Figure 37-12d shows a typical discone antenna. The existing frequency-independent antennas having a constant performance at all frequencies also include self-complementary log-periodic structures, such as planar log-periodic slot antennas, bidirectional log-periodic antennas, log-periodic dipole arrays, two- and four-arm log spiral antennas, and conical log-spiral antennas.28 Furthermore, the cylindrical antennas can achieve broadband impedance characteristics by using resistive loading to form traveling wave along the dipole arms.29–31 However, the antennas mentioned above are seldom used in portable/mobile devices due to the constraints of bulky size and high manufacturing cost, although they have been widely used in radar systems, electromagnetic compatibility (EMC), and channel measurements, as well as base stations for wireless communications. Therefore, planar monopoles or disc antennas have been alternatively used in wireless communications because they have shown excellent performance in impedance and radiation and because they have the significant advantage of small size.32–35 The earliest planar dipole may be the bow-tie antenna invented by Brown-Woodward.36 Figure 37-13 reviews the planar trapezoidal designs.37–40 A general radiator with dimensions of Wu, Wb, H, and h that is installed above a large ground plane forms a planar monopole. The monopole antenna can be excited by a coaxial cable through a surfacemounted adapter (SMA) connector. The triangular monopole antenna is the case with Wb = 0 or Wu = 0, which can be considered as the planar versions of bi-conical antennas.



Wu>Wb H

Wu

Wb

Ground plane

h

Wb=0

Wu=Wb

SMA

FIGURE 37-13

37-13

Wu=0

Review of the planar trapezoidal monopoles

The rectangular monopole antenna can be formed as Wb = Wu ≠ 0, which is the planar version of cylindrical monopole antenna. The frequency corresponding to the lower edge of impedance bandwidth of the planar trapezoidal monopole can be estimated by Equations 37-2 and 37-3. F (GHz) =

F (GHz) =

0.25 × 300 W + Wb H +h+ u 4π 0.25 × 300

(37-2)

2

Wu  2 max(Wu , Wb ) − Wu − Wb  H2 +   +h+ 2 2π  

(37-3)

where W max(Wu , Wb ) =  u Wb

Wu ≥ Wb Wb ≥ Wu

and all dimensions are in millimeters. Equation 37-3 is the modified version of Eq. 37-2 to estimate the frequency corresponding to the lower edge of the impedance bandwidth and has been used to estimate the frequency corresponding to the lower edge of the impedance bandwidth of a rectangular radiator as a special case of polygonal radiators. The rectangular planar monopole antenna can be derived from the trapezoidal planar antenna, as shown in Figure 37-13.41–45 The monopole performance is mainly determined by the shape and size of the planar radiator as well as the feeding section. The overall size of the monopole and shape of the radiator dominate the frequency corresponding to the lower edge of the impedance bandwidth. The feed gap, the location of the feedpoint, and the shape of the radiator greatly affect the impedance matching.46–49 Figure 37-14 shows a rectangular planar monopole antenna and its modified versions. By modifying the shapes of the bottom sides of the radiators, the impedance matching of the planar rectangular monopole can be further improved, as shown in Figure 37-14b, c, and d, where the impedance transition




(b)

(c)

(d)

(e)

(f)

(g)

(a)

FIGURE 37-14 versions

A rectangular planar monopole antenna and its modified

between the coaxial probe and the radiator has been improved by using step transition, two branch feeding strips, and a beveled radiator.46,47,49–51 A broadband impedance transition will ensure an impedance matching across a broad bandwidth. A shorting pin can also be used to make the planar antenna more compact.52 The impedance and radiation performance of four types of square planar monopole antennas were studied.48 By slotting the radiators of the planar monopoles, the impedance bandwidth of the monopole can be improved where the current distributions on the radiators are changed, as shown in Figure 37-14e and f.53–55 By adding a horizontal strip, the height of the planar monopole can be reduced because of the longer effective current path, as shown in Figure 37-14g.56 Theoretically, the radiators of the planar monopole antennas can be of any shape for broad operating bandwidth. The elliptical planar monopole antenna is one of the important types of planar monopoles in planar antenna design due to its broadband-even high-pass impedance performance. Figure 37-15 illustrates the elliptical monopoles in their general forms (Figure 37-15a), circular monopole (Figure 37-15b), semicircular monopole (Figure 37-15c), rectangular monopole with a semicircular bottom (Figure 37-15d), and annular monopole (Figure 37-15e).57–62

(b)

(c)

(d)

(e)

(a)

FIGURE 37-15

Elliptical planar antennas and their variations



37-15

FIGURE 37-16 The prototypes of printed PCB monopoles: (a) a microstrip-fed triangular monopole; (b) a microstrip-fed circular PCB monopole; (c) a microstrip-fed rectangle+semicircle PCB monopole; (d) a microstrip-fed rectangular PCB monopole; (e) a CPW-fed triangle+semicircle PCB monopole; (f ) a directly cable-fed rectangular PCB monopole with a strip; (g) a directly cable-fed circular PCB monopole with a slot; (h) a CPW-fed circular slot PCB antenna (from left to right and from the first to the second row).

However, in mobile wireless applications, the antennas are expected to be embeddable or easy to be integrated into wireless devices in system design. Therefore, the antennas directly printed onto a printed circuit board (PCB) are the most promising designs. Such antennas are usually constructed by etching the radiators onto the dielectric substrate of PCB stabs and a ground plane around the radiators. Figure 37-16 demonstrates some prototypes of PCB antennas. It can be seen that the PCB antennas can be fed by a microstrip transmission line (Figure 37-16a–d ), a coplanar waveguide (CPW) structure (Figure 37-16e and h), or directly by an RF cable where the inner conductor is connected to the upper radiator and the outer conductor is grounded into a system ground (Figure 37-16f and g), as shown in Figure 37-16. For the microstrip-fed cases, the ground planes are usually etched on opposite surfaces, which are connected into the system ground. The ground planes are etched close to the feeding strip on the same surface as the feeding strip for the CPW-fed PCB monopoles. The directly cable-fed monopoles are often used in scenarios where the PCB monopoles are not close to the main board and an RF cable is used to connect the antenna into the main board—for example, the antennas embedded into the cover of a laptop computer. Antennas Applied in High-Speed UWB Wireless Communications UWB is a technology for the transmission of information by using an extremely large operating bandwidth. Since the first Report and Order by the Federal Communications Commission (FCC) released the unlicensed use of UWB on February 14, 2002, the UWB technology has been expected to be the next-generation technology for short-range but high-speed wireless communications.63 The frequency range of 3.1–10.6 GHz has been allocated for such applications. The emission limit masks are regulated by regulators such




FIGURE 37-17 Emission limit masks for indoor and outdoor UWB applications

as the FCC for co-existence with other electronic systems, as shown in Figure 37-17. The emission power limits are lower than the noise floor in order to avoid possible interference between UWB devices and existing electronic systems. The masks may vary against regions or countries but the maximum emission levels are always kept lower than –41.3 dBm / MHz. The extremely large spectrum provides room to use extremely short pulses on the order of picoseconds. Thus, the pulse repetition or data rates can be low or very high, typically up to giga pulse/second. The pulse rates are dependent on the applications. Due to the unique requirements of UWB systems, the antenna design faces many challenges. Besides the requirements for conventional broadband antennas in terms of impedance bandwidth, gain, radiation patterns, and polarization, the special requirements for the UWB antenna in general include the following two points. First, both the frequency and time domain responses of the UWB antenna should be taken into account due to the extremely wide bandwidth and direct pulsed signals. The frequency domain response includes all conventional parameters such as impedance, radiation, and transmission. The impedance bandwidth is measured in terms of return loss or voltage standing wave ratio (VSWR). Usually, the return loss should be less than –10 dB or VSWR < 2:1. An antenna with an impedance bandwidth narrower than the operating bandwidth tailors the spectrum of transmitted and received signals, acting as a bandpass filter in the frequency domain, and reshapes the waveforms of radiated or received pulses in the time domain. The radiation performance includes radiation efficiency, radiation patterns, polarization, and gain. The radiation efficiency is an important parameter especially for small antenna design, where it is difficult to achieve impedance matching due to small radiation resistance and large reactance. For the small antenna with weak radiation directivity, the radiation efficiency is of greater practical interest than the gain. The radiation patterns show the directions where the signals will be transmitted. Second, the effect of the modulation schemes of the systems on the antenna design should be taken into consideration. The wireless communication systems based on UWB technology can be OFDM based or pulsed based. In an OFDM-based system, the whole UWB band is occupied by many sub-bands, where each sub-band having a few hundred megahertz (larger than 500 MHz) can be considered as broadband. Within the sub-bands, the effect of nonlinearity of the phase shift on the receiving performance can be ignored because the phase varies very slowly with frequency. Therefore, the design of the antenna is more focused on achieving constant frequency response in terms of the radiation efficiency,



37-17

gain, return loss, radiation patterns, and polarization over the operating band, which may fully or partially cover the UWB bandwidth of 7.5 GHz. The pulse-based UWB systems usually require linear phase, impedance, and gain responses that can entirely cover the operating bandwidth or partially cover the bandwidth where the majority of the pulse energy is distributed. To prevent the distortion of the received pulses, a UWB antenna that can produce radiation fields of constant magnitude and a phase shift that varies linearly with frequency is desired. In brief, an OFDM-based UWB system has almost the same requirements as those in a broadband system but with an extremely broad bandwidth, which usually varies from 50 percent (for the lower UWB band of 3.1−5 GHz) to 100 percent (for the entire UWB band of 3.1−10.6 GHz). However, paying additional attention to pulse-based UWB systems is needed, where a UWB antenna usually acts as a bandpass filter that reshapes the spectra of the radiated /received pulses. To design a good UWB antenna, we should keep the following important design factors in mind: ●

●

●

●

●

●

●

●

Enough impedance bandwidth to cover the operating bandwidth Steady directional or omnidirectional gain radiation patterns High and constant gain along desired direction(s) Linear phase, especially for pulse-based systems Consistent polarization Mobility with small size/low profile Low design complicity Low material /manufacturing cost

Based on these requirements, a circular planar dipole antenna is exemplified. Figure 37-18 shows a pair of the circular planar dipoles, which are face-to-face positioned for transfer function (S21) measurement. The planar circle has a radius of 10.5 mm and feed gap of 2.2 mm. The identical antennas are used as transmit and receive antennas in the test setup shown in Figure 37-18, a pair of the planar dipole antennas with a separation of 100 mm and positioned in parallel and face-to-face. y

y’ 100 mm

10.5

x’

x 2.2

~

~ z

FIGURE 37-18

z’

A pair of circular planar dipole antennas




(a) FIGURE 37-19 dipole antenna

(b)

(a) Return loss |S11| and magnitude of |S21|; (b) the phase response of |S21| of the circular

Figure 37-19 demonstrates the return losses |S11| and the magnitude of the transfer function |S21| for the dipole shown in Figure 37-18. The circular planar dipole antenna demonstrates the broad impedance bandwidth covering the entire UWB band and broad transmission coverage 2−8.5 GHz within a 10-dB variation. The results shown in Figure 37-19 are simulated. Antennas in UWB-based Wireless USB The most promising applications of UWB technology in the short-range and high-speed wireless interfaces may be wireless USB (WUSB), which replaces wired USB to match the data rate of USB 2.0 at 480 Mbps. The technology is a hub-and-spoke connection that supports dual-role devices in which a product such as a camera can act either as a device to a host laptop/desktop or as a host to a device such as a printer. In this section, a printed PCB antenna will be designed for a WUSB dongle, which is used in a laptop environment. Figure 37-20a shows a printed PCB antenna on a dielectric substrate with a popular USB flash drive. The radiator and ground plane are etched on opposite sides of the PCB, as shown in Figure 37-20b. The radiator consists of a rectangular section and a horizontal

y Printed antenna

Feedpoint

x Dielectric substrate

Top View

(a)

Ground plane

Bottom View

(b)

FIGURE 37-20 A small PCB antenna for a WUSB dongle



37-19

250 Antenna

100° 250

WUSB dongle mm 300

(a) FIGURE 32-21

(b)

(a) Placement of the antennas; (b) the antennas under test

strip, which is used to reduce the length of the rectangular radiator.56 A rectangular notch was cut close to the horizontal strip.64 The effects of the ground plane on the performance of the printed PCB antenna have greatly been reduced because of the introduction of the notch. The radiator is fed by a microstrip line from the left side of the radiator. The excitation is located at the bottom end of the microstrip line. The gap between the radiator and the ground plane is typically around 1 mm. The ground plane has a size of a typical wireless USB dongle, where the system will be installed. A WUSB dongle is placed beside a laptop. The laptop usually has a metallic and lossy cover and main body. To reduce the effects of the cover and body, the antenna is installed as far as possible from the cover and main body. As an example, the effect of the laptop on the radiation performance of the antenna is investigated. Figure 37-21 shows a dongle with a small printed PCB antenna installed at an IBM ThinkPad laptop under test. The keyboard panel of the laptop measures 250 × 300 × 25 mm. The screen measures 250 × 300 mm and is opened as if it is in use such that the angle between the screen and the keyboard panel is 100°. The antenna can be placed at different positions. Here, two different positions on the laptop for analysis are shown in Figure 37-21. The antenna can be placed at the right side of the back panel P1 and at the edge on the right-side panel P2. The size of the ground plane of the antenna does not affect the impedance and radiation performance significantly because the effect of the ground plane on the antenna performance has been significantly reduced.64 Therefore, no matter if the dongle is electrically connected to the metal casing or not, the performance of the antenna is unchanged. Figure 37-22 plots the radiation patterns for the total fields in the horizontal (x-y) plane when the PCB antenna is placed at two positions on the laptop and in free space. The coordinate system is shown in Figure 37-20. By comparing the gain for the antenna in free space and at positions P1 and P2, it can be seen that the antenna at P1 has higher gain than that at P2 due to the blockage effect of the cover on the radiation from the antenna in




(a)

(b)

(c) FIGURE 37-22 Measured gain patterns for total fields in the x-y plane (a) in free space and (b) for the antenna positions P1 and (c) P2 on the laptop

proximity of the laptop. The reflection from the cover leads to the stronger radiation from the antenna at P1 in the span of 160° –90° –0° –270° than that within the rest range. The radiation for the antenna at P2 is nearly omnidirectional but with several nulls and lower gain than that for the antenna at P1 in most of the directions, although the blockage of the cover on the antenna at P2 should be less than that on the antenna at P1. In short, UWB is a promising technology for short-range high-speed wireless communications. To accommodate the special requirements for the antennas, the traits of UWB systems should be studied carefully and effects of the installation environments on the antenna performance should be taken into account in the system’s point of view, in particular for small and embedded antennas.

REFERENCES 1. Z. N. Chen and M. Y. W. Chia, Broadband Planar Antennas: Design and Applications (New York: John Wiley & Sons, Inc. 2006). 2. J. Howell, “Microstrip Antennas,” Antennas and Propagation Society International Symposium, vol. 10 (December 1972): 177–180. 3. R. Munson, “Conformal Microstrip Antennas and Microstrip Phased Arrays,” IEEE Trans. on Antennas and Propagat., vol. 22, no. 1 (January 1974): 74–78.



37-21

4. D. M. Pozar, “A Microstrip Antenna Aperture Coupled to a Microstrip Line,” Electronics Letters, vol. 21 (January 17, 1985): 49–50. 5. J. M. Griffin and J. R. Forrest, “Broadband Circular Disc Microstrip Antenna,” Electronics Letters, vol. 18, no. 5 (March 1982): 266–269. 6. K. F. Lee, K. Ho, and J. Dahele, “Circular-Disk Microstrip Antenna with an Air Gap,” IEEE Trans. on Antennas and Propagat., vol. 32, no. 8 (August 1984): 880–884. 7. K. S. Fong, H. F. Guse, and M. J. Withers, “Wideband Multiplayer Coaxial-Fed Microstrip Antenna Element,” Electronics Letters, vol. 21, no. 11 (May 1985): 497–499. 8. G. Mayhew-Ridgers, J. W. Odendaal, and J. Joubert, “Single-Layer Capacitive Feed for Wideband Probe-Fed Microstrip Antenna Elements,” IEEE Trans. on Antennas and Propagat., vol. 51, no. 6 (June 2003): 1405–1407. 9. T. Huynh and K. F. Lee, “Single-Layer Single-Patch Wideband Microstrip Antenna,” Electronic Letters, vol. 31, no. 16 (August 1995): 1310–1312. 10. Z. N. Chen, “Experimental Investigation on Rectangular Plate Antenna with Ω-shaped Slot,” Radio Science, vol. 36, no. 5 (August-September 2001): 833–840. 11. H. Nakano, M. Yamazaki, and J. Yamauchi, “Electromagnetically Coupled Curl Antenna,” Electronic Letters, vol. 33, no. 12 (June 1997): 1003–1004. 12. K. M. Luk, C. L. Mak, Y. L. Chow, and K. F. Lee, “Broadband Microstrip Antenna,” Electronic Letters, vol. 34, no. 15 (July 1998): 1442–1443. 13. Z. N. Chen and M. Y. W. Chia, “A Feeding Scheme for Enhancing Impedance Bandwidth of a Suspended Plate Antenna,” Microwave and Optical Technology Letters, vol. 38, no. 1 (2003): 21–25. 14. P. S. Hall, “Probe Compensation in Thick Microstrip Patches,” Electronic Letters, vol. 23 (1987): 606–607. 15. L. Levis, A. Ittipiboon, and A. Petosa, “Probe Radiation Cancellation in Wideband Probe-Fed Microstrip Arrays,” Electronic Letters, vol. 36 (2000): 606–607. 16. Z. N. Chen and M. Y. W Chia, “Broadband Suspended Probe-Fed Plate Antenna with Low CrossPolarization Levels,” IEEE Trans. on Antennas and Propagat., vol. 51, no. 2 (February 2003): 345–346. 17. Z. N. Chen and M. Y. W. Chia, “A Novel Center-Fed Suspended Plate Antenna,” IEEE Trans. on Antennas and Propagat., vol. 51, no 6 (June 2003): 1407–1410. 18. Z. N. Chen and M. Y. W. Chia, “Center-Fed Microstrip Patch Antenna,” IEEE Trans. Antennas and Propagat., vol. 51, no 3 (March 2003): 483–487. 19. Z. N. Chen, and M. Y. W. Chia, “Broadband Suspended Plate Antennas Fed by Double L-shaped Strips,” IEEE Trans. on Antennas and Propagat., vol. 52, no. 9 (September 2004): 2496–2500. 20. M. Kanda, “Transients in a Resistively Loaded Linear Antennas Compared with Those in a Conical Antenna and a TEM Horn,” IEEE Trans. on Antennas and Propagat., vol. 28, no. 1 (January 1980): 132–136. 21. L. T. Chang and W. D. Burnside, “An Ultrawide-Bandwidth Tapered Resistive TEM Horn Antenna,” IEEE Trans. on Antennas and Propagat., vol. 48, no. 12 (December 2000): 1848–1857. 22. R. T. Lee and G. S. Smith, “On the Characteristic Impedance of the TEM Horn Antenna,” IEEE Trans. on Antennas and Propagat., vol. 52, no. 1 (January 2004): 315–318. 23. T. W. Hertel and G. S. Smith, “On the Dispersive Properties of the Conical Spiral Antenna and Its Use for Pulsed Radiation,” IEEE Trans. on Antennas and Propagat., vol. 51, no. 7 (July 2003): 1426–1433. 24. J. D. Kraus, Antennas, 2nd Ed. (New York: McGraw-Hill, 1988): 340–358. 25. C. W. Harrison, Jr. and C. S. Williams, Jr., “Transients in Wide-Angle Conical Antennas,” IEEE Trans. on Antennas and Propagat., vol. 13, no. 2 (March 1965): 236–246. 26. S. S. Sandler and R. W. P. King, “Compact Conical Antennas for Wide-band Coverage,” IEEE Trans. on Antennas and Propagat., vol. 42, no. 3 (March 1994): 436–439. 27. S. N. Samaddar and E. L. Mokole, “Biconical Antennas with Unequal Cone Angles,” IEEE Trans. on Antennas and Propagat., vol. 46, no. 2 (February 1998): 181–193. 28. P. E. Mayes, “Frequency-Independent Antennas and Broad-band Derivatives Thereof,” Proceedings of the IEEE, vol. 80, no. 1 (January 1992): 103–112.




29. T. T. Wu and R. W. P. King, “The Cylindrical Antenna with Nonreflecting Resistive Loading,” IEEE Trans. on Antennas and Propagat., vol. 13, no. 3 (May 1965): 369–373. 30. D. L. Senguta and Y. P. Liu, “Analytical Investigation of Waveforms Radiated by a Resistively Loaded Linear Antenna Excited by a Gaussian Pulse,” Radio Science, vol. 9, no. 3 (June 1974): 621–630. 31. J. G. Maloney and G. S. Smith, “A Study of Transient Radiation from the Wu-King Resistive MonopoleFDTD Analysis and Experimental Measurements,” IEEE Trans. on Antennas and Propagat., vol. 41, no. 5 (May 1993): 668–676. 32. H. Meinke and F. W. Gundlach, Taschenbuch der Hochfrequenztechnik (Berlin: Springer-Verlag, 1968): 531–535. 33. G. Dubost and Zisler, Antennas a Large Bande (Paris: Masson, 1976): 128–129. 34. S. Honda, M. Ito, H. Seki, and Y. Jinbo, “A Disk Monopole Antenna with 1:8 Impedance Bandwidth and Omnidirectional Radiation Pattern,” Int. Symp. Antennas Propagat. (1992): 1145–1148. 35. M. Hammoud, P. Poey, and F. Colombel, “Matching the Input Impedance of a Broadband Disc Monopole,” Electronics Lett., vol. 29 (1993): 406–407. 36. G. H. Brown and O. M. Woodward, “Experimentally Determined Radiation Characteristics of Conical and Triangular Antennas,” RCA Review, vol. 13 (December 1952): 425–452. 37. Z. N. Chen and M. Y. W. Chia, “Impedance Characteristics of Trapezoidal Planar Monopole Antenna,” Microw. Opt. Techno. Lett., vol. 27, no. 2 (October 2000): 120–122. 38. J. A. Evans and M. J. Ammann, “Planar Trapezoidal and Pentagonal Monopoles with Impedance Bandwidths in Excess of 10:1,” IEEE Antennas and Propagation Society International Symposium, vol. 3 (July 11–16, 1999): 1558–1561. 39. X. H. Wu, Z. N. Chen, and N. Yang, “Optimization of Planar Diamond Antenna for Single/ Multiband UWB Wireless Communications,” Microw. Opt. Techno. Lett., vol. 42, no. 6 (2004): 451–455. 40. Z. N. Chen, “Impedance Characteristics of Planar Bow-tie-like Monopole Antennas,” Electronics Letters, vol. 36, no. 13 (2000): 1100–1101. 41. M. J. Ammann, “Square Planar Monopole Antenna,” IEE National Conference on Antennas and Propagation (March 31–April 1, 1999): 37–40. 42. M. J. Ammann, “Impedance Bandwidth of the Square Planar Monopole,” Microw. Opt. Techno. Lett., vol. 24 (2000): 185–187. 43. M. J. Ammann and Z. N. Chen, “An Asymmetrical Feed Arrangement for Improved Impedance Bandwidth of Planar Monopole Antennas,” Microw. Opt. Techn. Lett., vol. 40, no. 2 (2004): 156–158. 44. Z. N. Chen, M. J. Ammann, and M. Y. W. Chia, “Broadband Square Annular Planar Monopoles,” Microw. Opt. Techno. Lett., vol. 36, no. 6 (March 2003): 449–454. 45. Z. N. Chen, “Experimental on Input Impedance of Tilted Planar Monopole Antennas,” Microw. Opt. Techno. Lett., vol. 26, no. 3 (2000): 202–204. 46. K. G. Thomas, N. Lenin, and R. Sivaramakrishnan, “Ultrawideband Planar Disc Monopole,” IEEE Trans. on Antennas and Propagat., vol. 54, no. 4 (April 2006): 1339–1341. 47. S. Su, K. Wong, and C. Tang, “Ultrawideband Square Planar Antenna for IEEE 802.16a Operating in the 2-11 GHz Band,” Microw. Opt. Techn. Lett., vol. 42, no. 6 (September 2004): 463–466. 48. X. H. Wu and Z. N. Chen, “Comparison of Planar Dipoles in UWB Applications,” IEEE Trans. on Antennas and Propagat., vol. 53, no. 6 (2005): 1973–1983. 49. M. J. Ammann and Z. N. Chen, “A Wideband Shorted Planar Monopole with Bevel,” IEEE Trans. on Antennas and Propagat., vol. 51, no. 4 (2003): 901–903. 50. X. H. Wu, A. A. Kishk, and Z. N. Chen, “A Linear Antenna Array for UWB Applications” IEEE Int. Symp. Antennas and Propagat., vol. 1A (July 3–8, 2005): 594–597. 51. E. Antonino-Daviu, M. Cabedo-Fabres, M. Ferrando-Bataller, and A. Valero-Nogueira, “Wideband Double-Fed Planar Monopole Antennas,” Electronics Letters, vol. 39, no. 23 (November 2003): 1635–1636.



37-23

52. E. Lee, P. S. Hall, and P. Gardner, “Compact Wideband Planar Monopole Antenna,” Electronics Letters, vol. 35, no. 25 (December 1999): 2157–2158. 53. H. S. Choi, J. K. Park, S. K. Kim, and J. Y. Park, “A New Ultra-Wideband Antenna for UWB Applications,” Microw. Opt. Techno. Lett., vol. 40, no. 11 (May 2004): 399–401. 54. S. Y. Suh, W. L. Stutzman, and W. A. Davis, “A New Ultrawideband Printed Monopole Antenna: the Planar Inverted Cone Antenna (PICA),” IEEE Trans. on Antennas and Propagat., vol. 52, no. 5 (May 2004): 1361–1364. 55. D. Valderas, J. Meléndez, and I. Sancho, “Some Design Criteria for UWB Planar Monopole Antennas: Application to a Slotted Rectangular Monopole,” Microw. Opt. Techno. Lett., vol. 46, no. 1 (July 2005): 6–11. 56. A. Cai, T. S. P. See, and Z. N. Chen, “Study of Human Head Effects on UWB Antenna,” IEEE Intl. Workshop on Antenna Technology (March 7–9, 2005): 310–313. 57. Z. N. Chen, M. J. Ammann, M. Y. W. Chia, and T. S. P. See, “Circular Annular Planar Monopoles with EM Coupling,” IEE Proc. Microw. Antennas, Propagat., vol. 150, no. 4 (August 2003): 269–273. 58. N. P. Agrawall, G. Kumar, and K. P. Ray, “Wide-band Planar Monopole Antenna,” IEEE Trans. on Antennas and Propagat., vol. 46, no. 2 (February 1998): 294–295. 59. T. Yang and W. A. Davis, “Planar Half-Disk Antenna Structures for Ultra-Wideband Communications,” IEEE Int. Symp. Antennas and Propagat., vol. 3 (June 2004): 2508–2511. 60. C. Y. Huang and W. C. Hsia, “Planar Elliptical Antenna for Ultra-Wideband Communications,” Electronics Letters, vol. 41, no. 6 (March 2005): 296–297. 61. P. V. Anob, K. P. Ray, and G. Kumar, “Wideband Orthogonal Square Monopole Antennas with Semi-Circular Base,” IEEE Int. Symp. Antennas and Propagat., vol. 3 (July 2001): 294–297. 62. J. W. Lee, C. S. Cho, and J. Kim, “A New Vertical Half Disc-Loaded Ultra-Wideband Monopole Antenna (VHDMA) with a Horizontally Top-Loaded Small Disc,” Antennas and Wireless Propagat. Letters, vol. 4 (2005): 198–201. 63. First Report and Order, Federal Communications Commission (FCC), February 14, 2002. 64. Z. N. Chen, T. S. P. See, and X. M. Qing, “Small Ground-Independent Planar UWB Antenna,” IEEE Int. Symp. Antennas and Propagat., Albuquerque, New Mexico, July 2006.


Broadband Planar Antennas for High-Speed Wireless Communications



Chapter 38

Antennas for Medical Applications Cynthia M. Furse University of Utah

CONTENTS 38.1 OVERVIEW OF ANTENNAS FOR MEDICAL APPLICATIONS . . . .

38-2

38.2 THE ENVIRONMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

38-2

38.3 ANTENNAS FOR MEDICAL IMAGING . . . . . . . . . . . . . . . . . . . . . .

38-6

38.4 HEATING. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38-11 38.5 COMMUNICATION (BIOTELEMETRY) . . . . . . . . . . . . . . . . . . . . . . 38-13 38.6 PULSED ELECTROMAGNETIC FIELDS . . . . . . . . . . . . . . . . . . . . . 38-16 38.7 SENSING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38-17 38.8 FUTURE DIRECTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38-19


Antennas for Medical Applications 38-2

CHAPTER THIRTY-EIGHT

38.1 OVERVIEW OF ANTENNAS FOR MEDICAL APPLICATIONS Antennas used for medical applications span applications in imaging, communication with implantable medical devices, heating for treatment of cancer, cardiac abnormalities, and hypothermia, measurement of fields for assessment of RF safety, augmentation of healing, and reduction of pain. Some of these applications have gained worldwide acceptance and are currently used with human subjects, and others are still in the research and development stage. This chapter describes the nature of the human body environment in which these antennas are commonly used, relevant regulations and guidelines, and the antennas and their applications.

38.2 THE ENVIRONMENT The Regulatory Environment There are two types of regulations of particular interest to designers of antennas for medical applications. The first is the allowable frequency. Applications that are used external to the body or for short periods of time (hyperthermia treatment, pain control, cardiac ablation, etc.) utilize the Industrial, Scientific, and Medical (ISM) bands (433, 915, 2450 MHz) in both the United States and Europe. Higher frequencies have the advantage of smaller antenna sizes, but the disadvantage of lower depths of penetration within the body. Implantable medical devices that are meant to stay in the body for a long period of time have been allocated a band of their own in the United States, the Medical Implant Communication Service (MICS) band from 402–405 MHz.1 The maximum bandwidth that can be used by a single device is 300 kHz in this band. The maximum power limit is 25 m W Equivalent Radiated Power (ERP).2,3 MICS shares its frequency allocation with the Meteorological Aids Service (METAIDS), which is used primarily by weather balloons, and is therefore specified for indoor use. The second type of regulation of interest in the design of antennas for medical applications is the limit on allowable absorbed power in the body. The limits for whole-body exposure are generally not the limiting factor. Instead, limits on localized power are more critical. Localized power is defined by specific absorption rate (SAR), which is calculated as SAR (W / kg ) =

σ | E p |2 2δ

(38-1)

where s is the electrical conductivity of the tissue (S/m), E p is the peak value of the electric field, and d is the density of the tissue (kg/m3). Localized SAR for nontherapeutic applications is limited to 2 W/kg in any 10-gram region of the body with an approximately cubic volume.4,5,6 An important exception is made in the new IEEE Safety Standard5 for the pinna (the outer ear) for an increased limit of 4.0 W/kg for the general public and 20.0 W/kg for occupational exposures. For therapeutic applications such as cardiac ablation and hyperthermia, the absorption limit does not apply, and care must then be taken not to damage surrounding tissues by overheating them. The question of whether electromagnetic radiation causes harm to the body at nonthermal levels has been hotly contested and remains a topic of ongoing research. A focused


Antennas for Medical Applications ANTENNAS FOR MEDICAL APPLICATIONS

38-3

review of this research was completed by the NIEHS in 1999.7 Significant research and ongoing professional oversight have lead to today’s RF exposure standards, which are meant to ensure safe levels of RF fields.8 Still, designers of antennas for medical systems should be aware of public fears that may accompany their use and should anticipate being asked to address this question. The Physical Environment Antennas used for medical applications are strongly impacted by the lossy dielectric materials that make up the human body. This presents challenges when antennas are used for a communication system, because the tissues absorb the power and detune the antenna. This depends strongly on the frequency and the location and depth of the antenna in or near the body, and significant variation can be seen from patient to patient. Figure 38-1 shows the electrical properties of muscle and fat as a function of frequency.9 At low frequencies, the conductivity of the tissue dominates the impact on the field, and at high frequencies, the dielectric values tend to dominate. Table 38-19 shows the electrical properties of several different tissues in the body at 433 MHz, which is a commonly used ISM frequency. Muscle is highly conductive and therefore very lossy, whereas fat has lower conductivity and therefore lower loss. These two tissues are near extremes in the body. A common rough approximation is that the body can be modeled using average properties of 2/3 muscle. This is suitable for addressing global questions such as total power absorbed in the body, but is generally not suitable for evaluating near field effects such as peak SAR. The lossy tissues in the body have several effects on antennas used for medical applications. When antennas are used for deliberately depositing power in the body for hyperthermia or cardiac ablation, for instance, the power tends to stay more localized around the antenna, where it is absorbed and converted to heat. For cardiac ablation or in vitro hyperthermia applicators, this is good, because it means that the heat will not penetrate to nearby structures that are not meant to be heated. For external, whole-body applicators for hyperthermia or for medical imaging applications, this loss means that it can be difficult to get the power to penetrate deep within the body. Lower frequencies are used when possible, and regions near the surface of the body (such as the breast) are easier to work with than areas deep within the torso. Multiple antennas must be used outside the body and focused in some way in order to get the power deep into the body. For biotelemetry (communication) applications, the same types of problems are seen. Communication with subcutaneous implants loses less power in the body than communication with deep body implants, for instance. Power lost in the body has two effects: it is wasted and cannot be used for communication, and the RF exposure limits typically limit

FIGURE 38-1 Electrical properties of muscle and fat (from Gabriel9)


Antennas for Medical Applications 38-4 TABLE 38-1


Electrical Properties of Tissues at 433 MHz (from Gabriel9)

Tissue

er

s (S/m)

Aorta

49.15

0.7395

Bladder

17.67

0.3128

Blood

57.3

1.72

Bone (Cancellous)

21.08

0.02275

Bone (Cortical)

13.77

0.1032

Bone (Marrow)

5.137

0.03575

Breast Fat

5.62

0.04953

Cartilage

43.64

0.65

Cerebellum

52.9

0.91

Cerebro Spinal Fluid

68.97

2.32

Cervix

44.17

1.020

Colon

60.88

0.96

Cornea

54.4

1.070

Dura

51.03

0.8

Eye Tissues

57.69

1.010

Fat

5.028

0.04502

Gall Bladder

60.06

1.035

Gall Bladder Bile

76.55

1.613

Grey Matter

54.27

0.8775

Heart

60.74

0.9866

Kidney

57.3

1.152

Lens Cortex

52.75

0.6742

Lens Nucleus

38.76

0.38

Liver

50.34

0.68

Lung Deflated

52.83

0.7147

Lung Inflated

21.58

0.3561

Muscle

64.21

0.9695

Nerve

35.7

0.500

Ovary

51.55

1.033

Skin (dry)

42.48

0.5495

Skin (wet)

51.31

0.72

Small Intestine

74.1

2.053

Spleen

60.62

1.041

Stomach

74.55

1.120

Tendon

50.53

0.7554



ANTENNAS FOR MEDICAL APPLICATIONS

TABLE 38-1

Continued

Tissue

er

s (S/m)

Testes

65.2

1.137

Thyroid

60.02

0.8183

Tongue

58.79

0.8993

Tracea

42.93

0.673

Uterus

64.73

1.117

Vitreous Humour

66.16

0.3931

White Matter

39.84

0.5339

the power that can be used for communication and hence the range or bandwidth of the system. Also, the lower the frequency used (to enable penetration and minimize power loss), the larger the antenna must be. Determining the exact impact of the body on the antenna generally requires precise calculation of the antenna in the presence of the human body. This typically requires numerical methods that are described in the next section. Numerical Simulation of Antennas in or near the Human Body Several methods for analyzing antenna arrays for medical applications exist.10 For simple cases where the biological structure can be approximated as uniform or by very simple models such as layers or cylinders, classical methods such as analytical analysis11,12 or method of moments13,14 can be used. If the structure of the body varies so much that anatomically precise modeling is rendered imprecise by variation between individuals, these simple analyses can be used to determine an optimal array design for the range of expected variation between individuals. An example of this was done in Hadley15 for design of coils for vascular MRI. Another example in which the body can be modeled as near-uniform is in the case of arrays for hyperthermia of the brain. In Furse and Iskander16 stepped-impedance dipoles were modeled using method of moments in an homogenous brain with a localized (nonhomogenous) tumor. Method of moments with a simple pulsed basis function (which is the most numerically efficient form) has limitations for heterogeneous models, however, due to artificial charge buildup on the dielectric interfaces.13 Higher-order basis functions can overcome this limitation, although the computational complexity is significantly increased.17 In addition, the method of moments is very computationally expensive when heterogeneous models are evaluated. It requires N logN computations, where N is the number of cells in the model, including those making up the heterogeneous object. A more efficient method for calculation of heterogeneous objects is the finite difference time domain (FDTD) method, which has led to its tremendous popularity for numerical bioelectromagnetic calculations. For example, a interstitial array of hyperthermia applicators simulated using method of moments16 was simulated with a fraction of the computational resources using FDTD.18 Several individual hyperthermia applicators have been simulated using FDTD.19 FDTD requires N2 computations, where N is a cell in the (normally cubical) FDTD grid. Unlike method of moments, every cell in space (including at least a minimal amount of air surrounding the model) must be included in the discrete model, so the total number of cells, N, is likely to be larger. However, the significant improvement in computational efficiency generally makes this trade-off favor FDTD for bioelectromagnetic simulations. Complete detailed analysis of breast cancer imaging modalities was also done




with FDTD,20–43 as well as hyperthermia systems,44 and evaluation of cell phones (including those with dual antennas) near the human head.45,46 Antennas for implantation in the body (mostly microstrip or Planar Inverted F (PIFA) types) have been simulated with FDTD and in some cases optimized with genetic algorithms.47 Deep hyperthermia applicators (annular phased arrays) have been simulated extensively with FDTD.48,49 Several FDTD developments have been important for bioelectromagnetic simulations, including the development of frequency-dependent methods (FD)2TD,50 low-frequency FDTD methods,51 efficient FDTD computation,52 and evaluation of temperature using the bioheat equation.49 Model development is one of the significant challenges of numerical bioelectromagnetics. Models have progressed from the prolate spheroidal models of the human used during the 1970s53 to roughly 1-cm models based on anatomical cross sections used during the 1980s54 to a new class of millimeter-resolution MRI-based models of the body that have been the hallmarks of research since the 1990s.55–58 Today, probably the most widely used models are derived from the Visible Man Project.59 Once a tissue-segmented model has been chosen, the electrical properties of the tissues are defined. The properties of human tissue vary significantly with frequency, so it is essential to use data accurately measured at the frequency of interest. There is a wide range of published data on measured tissue properties,53,60–71 and work is still underway to measure and verify these properties. These and other references are electronically searchable at the University of Utah Dielectric Database OnLine.72

38.3 ANTENNAS FOR MEDICAL IMAGING One of the most promising uses of antennas in medical applications is for imaging the location of leukemia,73 breast tumors,20–43 and cardiac anomalies.74,75 Microwave imaging methods rely on the fact that the electrical properties of normal and malignant tissue are significantly different60–67 and that there is significant variation from tissue to tissue. Location of breast cancer shows particular promise, because its relatively low loss allows electromagnetic fields to propagate to the tumor and back, and the proximity of the tumor to the outer surface of the body means that the signal does not have more than a few inches to propagate. Two major microwave imaging methods utilize antenna arrays. Tomography20–27 attempts to map a complete electrical profile of the breast, and confocal imaging28–43 maps only the location of significant scatterers. Both of these methods have used antenna arrays made up of wideband elements to send and receive the test signals. Microwave thermography picks up the passive electromagnetic fields from the body.76–92 Magnetic Resonance Imaging (MRI) uses a strong magnetic field to cause the magnetic dipoles in the body to precess and then uses an array of loops to pick up the fields when they relax back to their normal state.11,93–113 Tomography for Breast Cancer Detection Microwave tomography is used to provide a complete spatial mapping of the electrical properties in the region of interest. During the acquisition phase, an array of antennas surrounds the region of interest. One of the antennas in the array is used to transmit a signal, normally a sine wave,20 set of sine waves,21–22 or a broadband signal,23 and all of the other antennas are used to receive the reflected signal. The array is scanned so that each antenna transmits each frequency, and these signals are received by each of the other antennas. After all of the data has been acquired, it is processed by comparing the received data with what



38-7

would be expected from a simulated model of the region. A numerical “forward model” is used to predict how much power is transmitted from the transmit antenna, passes into and reflects from the breast/tumor model, and is received by the receive antenna. Originally, the simulated model is just a good guess for what might be present, generally a generic breast model with no tumor. The differences between the measured and expected received data are used to modify the original guess to obtain an ideal model that best matches the measured data. This “inversion” is used to predict what model could have produced the measured data. Microwave tomography for breast cancer has been demonstrated by several groups.20–27 In the Dartmouth system,20 for instance, sine waves from 300–1000 MHz (being expanded to 3 GHz) are transmitted from a circular array of 16 transmit/receive monopole antennas, shown in Figure 38-2, to produce 2D reconstructed images of the breast. Quarter-wave monopole antennas (in the fluid) that were built by extending the inner conductor of semirigid coax were used for this application. Monopoles were chosen because they are easy to model as a line source in a 2D reconstruction algorithm with high accuracy.20 Water-filled waveguide apertures have also been used for tomography, but the monopole antennas were found to be as accurate and easier to build.24 Microwave tomography has been validated experimentally.27 The presence of 1.1- or 2.5-cm saline tubes (representing tumors) in excised breast tissue are seen to be clearly visible.26 Objects as small as 4 mm in diameter have been imaged at 900 MHz.27 Confocal Imaging for Breast Cancer Detection Confocal imaging for breast cancer detection is another exciting application of antenna arrays in medical imaging. Confocal imaging is similar to ground-penetrating radar. Unlike microwave tomographic imaging, this method does not provide a complete electrical mapping of the region of interest. Instead it identifies locations of significant scattering. This method typically uses a single antenna scanned in a flat array pattern above the breast or a cylindrical array of very small broadband antennas.28 For planar imaging, the patient lies face up, and the antenna is physically scanned in a plane above the breast.29–31 For cylindrical imaging, the patient lies face down, with the breast extending into the cylindrical array

FIGURE 38-2 2D monopole array used for tomographic imaging of the breast (after P. M. Meaney et al20 © IEEE 2000)




through a hole in the table.32,33 Matching fluid surrounding the breast, similar to that used for microwave tomography, is suggested in this case. Both methods provide similar results.33 One antenna in the array transmits an ultrawideband (UWB) pulse, which propagates into the breast, where it is reflected off significant electrical discontinuities, and is received in parallel by the other antennas in the array. Knowing the physical spacing between the array elements, the different delays between the transmit antenna, scattering point, and receiving antenna can be calculated geometrically. The received pulses representing a specific point in space can then be time delayed appropriately for each antenna, added up, and integrated to indicate the magnitude of the scattered energy from that point in space. This is effectively correlating the signals received from that point at all antennas. The antennas used for confocal imaging must be UWB and small enough to fit within the relatively small array area. Resolution of less than 1 cm requires a bandwidth of at least 5 GHz. The lossy nature of tissue attenuates high-frequency signals, limiting the upper frequency to about 10 GHz. Initially, resistively loaded bowties were suggested for the planar configuration,29–31,35,37 while dipole antennas were suggested for the cylindrical system.32,33 Resistively loaded Vee dipoles have also been proposed.36 In the cylindrical configuration, multiple antennas are present in the array, although they are not simultaneously active. In the planar system, a single antenna is scanned over the surface, creating a synthetic antenna aperture. To overcome the inherent inefficiency of resistively loaded antennas, a modified ridged horn antenna operating from 1 to 11 GHz has been introduced.38 Most of the antennas are designed to observe copolarized reflections from the breast; however, two resistively loaded bowtie antennas in the shape of a Maltese cross, as shown in Figure 38-3, have also been proposed to pick up the cross-polarized reflections.30 Cross-polarized reflections from simple tumor models were also examined.34,39 The antenna shown in Figure 38-334 consists of two cross-polarized bowtie antenna elements, an octagonal cavity behind the bowtie elements, and a metal flange attached to the cavity. The broadband bowties have flare angles of 45°. They are 1.67 cm long, which is a half-wavelength at 3 GHz in fat (similar to breast). The octagonal cavity blocks waves radiated away from the breast. The cavity is approximated as a circular waveguide filled with fat material for matching and size reduction. The first cutoff frequency is set to be Y Z

31.25mm 10mm X Inner flange

13.85mm 8.33mm X

11mm

62.5mm

Outer flange Cavity filled with epoxy

Epoxy

Bowtie elements 62.5mm (a)

Near field observation plane (z=1cm)

(b)

FIGURE 38-3 Cross-polarized antenna for confocal imaging. The properties of the substance inside the cavity and the medium outside the antenna are similar to fat (er = 9; s = 0:2 S/m) (after X. Yun et al34 © IEEE 2005).



38-9

2 GHz for 2 to 4 GHz operation. The cavity length is a quarter wavelength, which is 11 mm at 3 GHz. The flange consists of an inner and outer component, and is designed to block unwanted waves such as surface waves. The antenna performance does not change significantly when the flange size is varied between 10–6.25 mm, therefore, the width of the outer flange is set to be 6.25 mm. The inner flange is designed to prevent possible electric field overshoot at the inner corners of the opening of the octagonal cavity or at the ends of the bowtie elements. A slotline bowtie antenna has also been proposed.40 A resistively loaded monopole antenna, shown in Figure 38-4, suitable for use in a cylindrical array was proposed by Sill and Fear.41–43 Based on the Wu–King design114–115 this antenna was designed to be usable from 1 to 10 GHz immersed in canola oil (er = 3.0) for matching to breast tissue. The antenna is fabricated using high-frequency chip resistors (Vishay 0603HF, by Vishay Intertechnology, Malvern, PA) soldered to a high-frequency substrate (Rogers RO3203 series by Rogers Corporation, Chandler, AZ). The substrate (er = 3.02 and s = 0.001 S/m) has electrical properties similar to those of the canola oil. The antenna is soldered to a subminiature A (SMA) connector and attached to a metal plug for connection into the oil-filled test canister. The cylindrical confocal imaging system has been experimentally tested.41,43,116 Simulated tumors as small as 4 mm have been detected using a 3D system. Microwave Radiometry Microwave radiometry is a passive method where the natural electromagnetic radiation or emission from the body is measured to allow detection or diagnosis of pathogenic conditions.77 This method has been proposed for detection of breast cancer80,90,91 and brain cancer,81 in which the metabolism of cancer cells increases the localized temperature 1 to 3°C. This method has also been used for fluid and blood warming,77 for detection of rheumatology,79 and for monitoring temperature rise during hyperthermia treatment.78 Typical antennas include open-ended rectangular waveguides,82–85 small-loop antennas,87 and a horn antenna with a dielectric lens.88 Working around 3 GHz, all of these antennas have radiation patterns that have minimal penetration into the body, thus strongly weighting them to monitoring of surface temperatures.89,92 Increased focus, and therefore better spatial accuracy, was obtained with an array of six rectangular aperture antennas filled with low-loss

FIGURE 38-4 Fabricated resistively loaded monopole antenna soldered to an SMA connector and attached to a metal plug (after J. M. Sill and E. C. Fear43 © IEEE 2005)




dielectric (er = 25), which were scanned over the object of interest in an overlapping pattern. Preliminary results indicated promise for location of breast tumors.90–91 Magnetic Resonance Imaging Magnetic Resonance Imaging (MRI) uses a very strong magnetic field (0.5T, 1.5T, 3T, 4T, 7T, perhaps in the future 8T) to make the magnetic dipoles in the body precess (line up). When they are released, a set of receiver coils picks up the magnetic field created when these dipoles return to their normal orientations (position may change a lot as in blood imaging, diffusion, etc.). The relaxation properties of the different tissues affect the relative received signal intensities, and a 3D map of the body can be produced. There are two basic types of receiver coils used for MRI. Volume coils, such as the quadrature birdcage head coil shown in Figure 38-5,11 are used for imaging large and deep anatomic structures of the body and provide homogeneous field profiles. For highresolution applications that are more localized, such as angiographic imaging, hypocampus imaging, and functional imaging, in which the object features are very small, volume coils pick up less signal and more noise, thus having a lower signal-to-noise ratio (SNR) and poor-quality images. Modifications such as the use of an RF reflector or “endcap,”11 and modified shapes such as the elliptical94 or “dome”95,96 coils, have been developed. Smaller-volume surface coils97 have been shown to improve image quality, particularly when combined into phased arrays98–106 such as the one shown in Figure 38-6. Phasedarray coils are closer to the area of interest, so they pick up larger signal strength, and are smaller, so they pick up less noise, thus having a higher SNR. They are designed to overlap so that the mutual inductance between coils is zero, and so that the impedance at the preamplifier is very low, for optimal SNR.8 Part of the price for this improved image quality is the complexity of the receiver and data acquisition system, as each antenna is received on an individual channel. The image processing is also more computationally expensive, as the signal from each antenna is weighted depending on its proximity to the target region (and hence expected relative SNR), phase shifted, and combined with the other similarly processed signals.

FIGURE 38-5 Quadrature birdcage coil with endcap used for whole-volume head imaging (reprinted with permission from J. R. Hadley et al11 © Journal of Magnetic Resonance Imaging 2000)



38-11

FIGURE 38-6 (a) Two-element phased-array coil design. Dashed lines indicate the breaks in the underside of the double-sided copper section of the coil. (b) Image of finished phased-array coils (enclosed in foam) with triax balun cables and phased-array port connector box (reprinted with permission from J. R. Hadley et al11 © Journal of Magnetic Resonance Imaging 2000).

Among the practical considerations that are challenging with phased-array coils are the data acquisition time and the limited field of view, particularly for applications where the region of interest (an arterial occlusion, for instance) may not be precisely known and is therefore easy to miss. Phased-array coils have been used for numerous magnetic resonance angiography (MRA) applications including peripheral107,108 abdominal, intracranial, and carotid imaging.109–111 Recent coil designs have started to integrate phased-array elements into volume-like coils with the ability to control how the image is constructed to achieve maximum image quality.112,113 For these applications, the coil array functions much like the phased array in a synthetic aperture radar application. The image quality for the different coil types and configurations depends strongly on the application. The optimal image construction algorithm depends strongly on the application and region of interest, making the flexibility of being able to synthetically develop large or small subarrays very attractive.

38.4 HEATING Hyperthermia Hyperthermia (HT)117,118 is a method of treating cancer by heating the body. The tissue is typically heated to 41 to 45°C for 30 to 60 minutes. Often, this involves focusing the energy on the tumor region, relying on the tumor to be more sensitive to heat than the surrounding healthy tissues. This may be due to poor internal vasculature of the tumor or its higher conductivity and permittivity caused by increased water and ionic content. HT has also been shown to increase the effectiveness of radiation or chemotheraphy.119–120 The most commonly used frequencies for hyperthermia are 433, 915, and 2450 MHz. The type of antenna or antenna array used for HT depends on whether it is to be administered superficially, interstitially, or deep-body. Superficial HT applicators include microstrip,121 waveguide,122 current sheets,123 inductive,124 and the dual concentric conductor antenna, or DCC, shown in Figure 38-7.125–128 The DCC is particularly attractive, because it can be easily fabricated on flexible, printed,




FIGURE 38-7 DCC antenna geometry, E fields in the ring slot and edge currents (after P. F. Maccarini et al126 © IEEE 2004)

circuit-board material, which makes it easy to conform to virtually any part of the human body. The DCC aperture is a ring-slot configuration fed simultaneously on all four sides. Prediction and optimization of the heating is normally done by analyzing the near fields of the antenna (the heating region) numerically.126 Interstitial applicators for HT are typically monopole antennas made from coaxial cables with the center conductor extending beyond the outer ground shield of the cable.129 These antennas have a tear-drop-shaped radiation pattern, so the majority of the heating is near the feedpoint of the antenna (where the ground shield stops), leaving the tip of the antenna extended beyond the useable heating range. The heating distribution can be made more uniform by varying the width of the conductor16 or adding a choke to the antenna.130 Numerous other designs of interstitial applicators also exist. The heating pattern can be adjusted within the array by phasing the antenna elements16 or by using nonuniform insulation.131 Deep-body HT applicators are generally based on annular phased arrays (APA) of waveguides,118 coaxial TEM apertures,132 printed antennas,129,133 and induction systems.134 Originally, APA systems contained only one ring of 2D applicators surrounding the patient. The ring could be scanned vertically. Significant improvement with a true 3D HT system with the applicators vertically offset has been observed.129 The first clinically used 3D-type applicator is the SIGMA-Eye applicator (BSD Medical Corp., Salt Lake City, UT135). A detailed description of this applicator and different numerical antenna feed models can be found in Nadobny et al.136 Among the ongoing antenna design challenges in this area is the design of antennas that can be used to also monitor temperature and administer radiation therapy.128,133,137 One prototype combination device is shown in Figure 38-8 and another in Figure 38-9. Another research area is the use of optimization approaches to predict and control the heating pattern.138,139

FIGURE 38-8 A prototype of the new Berlin MR-compatible Water-Coated Antenna Applicator (WACOA) (after J. Nadobny et al137 © IEEE 2005)



MW PCB Array

Brachytherapy Catheters

Thermal Monitoring Catheters

38-13

RF Connectors

5mm water Bolue

FIGURE 38-9 Schematic of combination applicator showing component parts: parallel catheter arrays for brachytherapy sources and thermal mapping sensors, PCB antenna array, and water coupling bolus (after P. Stauffer et al133 © IEEE 2004)

Treatment of Hypothermia Patients who have experienced hypothermia are at great risk during the rewarming process. Conventional methods such as warm water baths use external sources of heat, which warm the peripheral regions while the heart is still cold. The demand for increased circulation to the extremities can overload the heart. EM applicators can produce deeper heating than methods that simply heat the body surface and rely on thermal conduction to carry the heat to the deeper tissues. EM heating provides heat deeper into the body, which increases cardiac output and circulates warmed blood to the peripheral tissues without overloading the heart.140 Cardiac Ablation Microwave catheter ablation uses monopole antennas inserted in catheters to the heart to treat cardia arrhythmias by creating deep lesions that destroy the source of the arrhythmias. Frequencies of 915 or 2450 MHz are typically used.141,142 Ablation requires localized temperatures of 50–90° C for a short time, typically a few minutes. Microwave ablation has also been used in conjunction with traditional balloon angioplasty to soften arterial plaque.143

38.5 COMMUNICATION (BIOTELEMETRY) Wireless communication systems and their associated implantable antennas are needed for communication with implantable medical devices such as cardiac pacemakers and defibrillators,144 neural recording and stimulation devices,145 and cochlear146 and retinal147 implants. Designing antennas for embedded applications is extremely challenging because of reduced antenna efficiency, impact of the environment on the antenna, the need to reduce antenna size, and the very strong effect of multipath losses. In addition to the present needs for embedded antennas, the expansion of MEMS, which are expected to play a dominant role in next-generation technologies, will add dramatically to the applications for imbedded antennas. Ultra-small devices (small enough to be injected in a human vein, for instance) and the desire to communicate with them will inevitably lead to the need for miniaturized antennas embedded in lossy environments. Emerging medical telemetry devices have led




to recent advances in the design of small, biocompatible antennas that can be implanted in the human body.148–175 Major challenges exist for implantable antennas. Not only do they have to be long-term biocompatible, but they must also be small, passive or highly efficient (to reduce battery requirements inside the body), and able to transmit power through the highly lossy body structure. In addition, they must meet the maximum SAR guidelines, which can be quite challenging. The very near field of the antenna is the inductive zone where power is not radiated. If lossy material is in the inductive zone of the antenna, which it normally is for implanted antennas, this near field power is absorbed and shows up in the SAR measurements or calculations. Thus, SAR is often the limiting factor for power transmitted from an implanted device. In the majority of wireless telemetry cases for implantable devices, inductively coupled coils often wound around a dielectric or ferrite core are used.152–167,176 In these cases, the timevarying magnetic field generated by the primary coil is received by the secondary coil, which results in an induced current in the implanted coil. Frequencies are often lower than 50 MHz to ensure that the presence of the human body does not significantly obstruct the coupling between the coils. In this case, the most important parameters for the design of the telemetry system are the self and mutual inductances of the coils. Several methods can be used to determine these parameters depending upon the frequency of interest and the geometrical shape of the coils. For the simplest traditional coil geometries (circle, square), analytical approximations of self and mutual inductances are often used, whereas more sophisticated methods (such as the partial inductance method and similar methods) can be used in the case of geometrically complex coils. A measure of the quality of coupling between two coils is given by the coupling coefficient K between two coils (0 ≤ K ≤ 1), defined by K=

M 12 L1 L2

where M12 is the mutual coupling between coils 1 and 2, and Li is the self inductance of coil i. To maximize the power delivered to the load in these applications, usually a capacitor is inserted in parallel with the inductance of the coil and the resistance of the load to form a parallel resonant LC circuit. Many other parameters may affect the design of inductively coupled coils for biomedical telemetry systems, such as implant size, maximum power, temperature increase in the implanted device, and specific absorption rate of power (SAR given in W/kg) induced in the tissue.177 Most inductive telemetry links are used for subcutaneous applications due to power restrictions for passive devices. Data rates are generally low, and size/weight and biocompatibility issues plague these devices. However, recent advances continue to reduce the power requirements and provide more biocompatible designs. The Utah Electrode Array (UEA), for example, uses a pickup coil printed on a ceramic substrate and integrated with the implanted neural electrode array, as shown in Figure 38-10.145 The implanted coil is energized by an external inductive programmer/reader that powers the implanted circuitry while transferring telemetry data. Radiofrequency links are also being developed for communication with medical implants. For cardiac telemetry, a dipole171 and spiral or serpentine microstrip,173,174 and a waffle-type patch178 have been designed for implantation in the shoulder. An insulated wire antenna has also been used, and this wire may be used as the lead between the heart and the battery pack/controls of the pacemaker.168 The antenna can be treated as a waveguide, where the lossy body acts as the outer conductor of the waveguide. The insulated antenna in tissue may be matched with a load resistor connected to the conducting tissue in order to reduce or eliminate the reflection.169



38-15

Telemetry Coil Receiver ASIC

Electrode Array

FIGURE 38-10 Utah Electrode Array packaged with a custom ASIC and printed receiver coil (after K. Guillory and R. A. Normann145 © J. Neurosci. Methods 1999)

Another type of antenna used for communication with cardiac devices is the circumference antenna, which is a monopole antenna that is mounted around the edge of the pacemaker case, as shown in Figure 38-11.169,170 The 94-mm-long circumference antenna shown in Figure 38-11 was centered in a plastic insulator with a thickness of 10 mm. The bandwidth where the SWR of this antenna is less than 2 is 42 MHz, which is larger than the required MICS allocation of 3 MHz. For smaller implants, a microstrip patch antenna has been successfully used for a retinal prosthesis,175 and a small dipole has been designed for communication with a brain implant.172 Deep-torso devices will experience more loss than subcutaneous devices. Furthermore, the location in the body controls the radiation pattern shape as well as magnitude. For example, the calculated radiation patterns for a small, multiturn loop antenna implanted in

FIGURE 38-11 Circumference antenna on a pacemaker model (after A. Johansson,169 Figure 5-3)




FIGURE 38-12 Calculated azimuthal radiation patterns for a 418-MHz vertical source: vertically polarized pattern (solid line) and horizontally polarized pattern (dotted line) (after W. G. Scanlon et al156 © IEEE 2000)

the vagina are shown in Figure 38-12. Measured net body losses (power absorbed in the body) for this 418-MHz antenna are 19.2 dB. The bodyworn radiation efficiency is

ηb =

Pbody = 1.2% Pair

where Pair is the total power radiated by the antenna in air, and Pbody is the total power radiated (external to the body) by the antenna implanted in the body. At 916.5 MHz, the measured net body loss is 24.3 dB, and the bodyworn radiation efficiency is 0.37 percent.156 The substantial losses in the body have so far limited deep-torso implants to communication with receivers held on or very near the body.

38.6 PULSED ELECTROMAGNETIC FIELDS Pulsed electromagnetic fields (PEMFs) have been developed for a number of medical applications. These fields are generally delivered by electrodes connected directly to the body, and as such are not truly an application of antennas. However, since this method is showing significant promise for many different medical applications, and since UWB antennas are being used in many other areas, it is not unlikely that antenna design concepts could be applied to PEMF applicators in future applications. Bone and Tissue Healing Pulsed electromagnetic fields have been found to be highly effective for healing fractures and soft-tissue injuries, particularly those that do not respond to ordinary healing methods. As early as 1812, passing “electric fluids” through needles inserted in the fracture gap was found to stimulate bone healing, and by the mid-1800s, this DC current stimulation was



38-17

considered the method of choice for slow-healing fractures. Today’s bone-healing PEMF systems typically use a 72-Hz single-pulse pattern. While the full biological mechanism is not minutely understood, it appears PEMFs of this type are able to retard the osteoclasts that destroy bone while increasing the rate of new bone formation.179 Pain Control Pulsed electromagnetic fields have also been used for controlling pain. Electrodes are placed strategically around the knee, shoulder, back, etc., and PEMF is applied. This has been found to provide both short- and long-term pain relief, although the exact reason is not fully understood. This method is currently used in both human and veterinary medicine.179 Drug Delivery and Electrochemotherapy Pulsed electromagnetic fields can be used for a wide variety of needle-less drug delivery applications. Iontophoresis is a method to electrically force drugs across a transdermal interface using a relatively small voltage (0.1–10 V) across the skin boundary. This method appears not to create structural changes in the cells or the skin, but rather just creates ion pathways that a conductive fluid (drug) will follow through pre-existing aqueous pathways. At present, a limited number of drugs can be delivered using this method.180 Pulsed electromagnetic fields can also be used to treat cancer, using a new therapy called electrochemotherapy, which has been used for a variety of cutaneous tumors, including head and neck tumors, melanomas, superficial breast cancer lesions, etc. In this therapy, the resistance of malignant cells to penetration by certain chemotherapeutic agents is temporarily lowered by electroporation, which creates temporary pores (pathways) in the membranes of the malignant cells by the application of short DC pulses that generate electric fields of several kV/cm. Once the cells are porated, the chemotherapeutic agents can enter the malignant cells and destroy them. Electrochemotherapy not only can increase the efficacy of certain chemotherapeutic agents, but also can reduce side effects because malignant cells can be destroyed with much lower doses of chemotherapeutic agents than with conventional chemotherapy.180,181

38.7 SENSING In addition to receiving and transmitting power for communication or imaging and depositing power for heating applications, antennas can be used as sensors. Antennas are used as probes for dielectric properties and electric or magnetic fields. Dielectric Measurement Probes Measurement of the electrical properties of tissues has been done extensively to facilitate research, numerical modeling, etc. High-frequency in vivo and in vitro dielectric measurements of tissue are typically made using an open-ended coaxial probe.182 The coaxial probe is sensitive to material that lies within a fringing capacitance zone adjacent to the probe tip. A two-wire, dipole-type probe has also been used.183 Another application of dielectric property measurements is in vivo measurement of brain fluid.184




FIGURE 38-13 Miniature printed dipole antenna for measurement of electric fields to determine cell phone RF exposure compliance (H. Bassen and G. Smith185 © IEEE 1983)

Electric and Magnetic Field Probes Miniature electric field probe antennas have been designed for assessment of compliance of electromagnetic devices with RF exposure guidelines.185 Measurement of SAR requires evaluation of the localized electric field, which should not be perturbed by the probe. This requires a very tiny electric or magnetic field receiver, such as the one shown in Figure 38-13. Because this probe is inherently sensitive to the polarization of the electric field, three perpendicular probes are used, as shown in the SPEAG probe in Figure 38-14. A magnetic field probe is also shown in this figure, with three perpendicular receiving loops.186

FIGURE 38-14 Electric and magnetic field probes from SPEAG (reprinted with permission, Schmid & Partner Engineering AG, Zurich)



38-19

38.8 FUTURE DIRECTIONS The medical applications of antennas described in this chapter are by no means all inclusive. New technology is rapidly being developed, and creative new ideas continue to emerge. The basic capabilities of antennas to transmit information, deliver heat, sense electrical properties, and receive information for imaging will continue to lead to new applications for antennas in medicine. At the risk of tabloid-type predictions, the following are some expectations for where antennas will continue to grow in medical applications. Communication with medical devices is an area that is rapidly expanding. The first medical devices were cardiac devices that had large battery packs and control systems and minimal data up/downlink requirements that could be managed in a doctor’s office. Today these devices are pressing for higher data rates, real-time communication, and more efficient links. Their size has been continually shrinking and promises to shrink radically, due mainly to advances in battery technology. Emerging nerve stimulation or recording devices require far less power than cardiac devices, and therefore do not require batteries at all. Prosthetic nerve devices hold promise for artificial vision, hearing, smell, balance and muscle control, nerve “repair,” and a new level of treatment of brain malfunction for Parkinson’s disease, depression, epilepsy, and more. New packaging techniques and ultraefficient electronics are driving the need for superminiaturized antennas. This demand is likely to grow dramatically with the success of microscale electronics, microfluidics, and microscale sensors and actuators. Smart pills have shown the capability of “untethered” communication systems in the body, and other applications are likely to utilize this freedom in the future, such as moving through the bloodstream for diagnosis or treatment. Devices used for heating have either been large, external devices that attempt to focus power, or much smaller devices inserted through a vein/artery or laproscopic surgical opening. These devices are also likely to shrink in size with the miniaturization of electronics, providing opportunity for more precise control of heat delivery. Although medical imaging is a relatively mature field, new methods and radical enhancements to mainstream methods continue to emerge. New antenna designs for imaging tend to be larger numbers (arrays) of small antennas. Enhancing the bandwidth while miniaturizing the antenna continues to be a focus in many applications. Arrays that are nonuniform and nonplanar are likely to be important in many applications. With the ability to place antennas in the body for medical implant devices, it would not be surprising to see the development of imaging systems that can be swallowed, injected, or placed in body orifices for better imaging of sensitive structures. Antennas are an integral part of medical devices today, and hold promise to play a significant role in the development of emerging devices for future medical systems. Acknowledgments Several researchers have contributed to this chapter, including Dr. Gianluca Lazzi (inductive coil telemetry), Jeff Johnson and Dr. Pichitpong Soontornpipit (biomedical telemetry), and Dr. Rock Hadley (MRI).

REFERENCES 1. Medical Implant Communications Service (MICS) Federal Register, “Rules and Regulations,” vol. 64, no. 240 (Dec. 1999): 69926–69934. 2. International Telecommunication Union, Recommendation ITU-R SA.1346, 1998.




3. “Planning for Medical Implant Communications Systems (MICS) and Related Devices,” Proposals Paper SPP 6/03, Australian Communications Authority. 4. M. H. Repacholi, Biological Effects and Medical Applications of Electromagnetic Energy, Chap. 2, O. P. Gandhi (ed.) (Englewood Cliffs, NJ: Prentice-Hall, 1990). 5. IEEE Standard for Safety Levels With Respect to Human Exposure to Radiofrequency Electromagnetic Fields, 3 kHz to 300 GHz, IEEE Std. C95.1, 1999, revised to Standard for Safety Levels with Respect to Human Exposure to Radiofrequency Electromagnetic Fields, 3 kHz to 300 GHz, IEEE Standards Coordinating Committee 28.4, 2006. 6. ICNIRP, “Guidelines for Limiting Exposure to Time-Varying Electric, Magnetic, and Electromagnetic Fields (up to 300 GHz),” Health Phys., vol. 74 (1998): 494–522. 7. National Institute of Environmental Health Science, http://www.niehs.nih.gov/emfrapid/ home.htm. 8. O. P. Gandhi, Biological Effects and Medical Applications of Electromagnetic Energy, Chap. 3, O. P. Gandhi (ed.) (Englewood Cliffs, NJ: Prentice-Hall, 1990). 9. C. Gabriel, “Compilation of the Dielectric Properties of Body Tissues at RF and Microwave Frequencies,” Final Technical Report, Occupational and Environmental Health Directorate Radiofrequency Radiation Division, Brooks Air Force Base, TX, 1996. 10. O. P. Gandhi, Biological Effects and Medical Applications of Electromagnetic Energy, Chap. 6 (Englewood Cliffs, NJ: Prentice-Hall, 1990). 11. J. R. Hadley, B. E. Chapman, J. A. Roberts, D. C. Chapman, K. C. Goodrich, H. R. Buswell, A. L. Alexander, J. S. Tsuruda, and D. L. Parker, “A Three-Coil Comparison for MR Angiography,” Journal of Magnetic Resonance Imaging, vol. 11 (2000): 458–468. 12. R. W. P. King, G. J. Fikioris, and R. B. Mack, Cylindrical Antennas and Arrays (Cambridge, UK: Cambridge Univ. Press, 2002). 13. H. Massoudi, C. H. Durney, and M. F. Iskander, “Limitations of the Cubical Block Model of Man in Calculating SAR Distribution,” IEEE Trans. Microwave Theory and Tech., vol. 32 (1984): 746–752. 14. C. T. Tsai, H. Massoudi, C. H. Durney, and M. F. Iskander, “A Procedure for Calculating Fields Inside Arbitrarily-Shaped, Inhomogeneous Dielectric Bodies Using Linear Basis Functions with the Moment Method,” IEEE Trans. Microwave Theory and Tech., vol. 34 (1986): 1131–1139. 15. J. Rock Hadley, “Design of Radio Frequency Coil Arrays for Optimal Signal to Noise Ratio for Magnetic Resonance Angiography,” PhD dissertation, University of Utah Electrical and Computer Engineering Department, 2005. 16. C. M. Furse and M. F. Iskander, “Three-dimensional Electromagnetic Power Deposition in Tumors Using Interstitial Antenna Arrays,” IEEE Trans. on Biomedical Engineering, vol. 36 (Oct. 1989): 977–986. 17. O. H. Schaubert, D. R. Wilton, and A. W. Glisson, “A Tetrahedral Modeling Method of Electromagnetic Scattering by Arbitrarily Shaped Inhomogeneous Objects,” IEEE. Trans. Antennas and Propagation, vol. 32 (1984): 75–82. 18. P. Cherry and M. F. Iskander, “FDTD Analysis of Power Deposition Patterns of an Array of Interstitial Antennas for Use in Microwave Hyperthermia,” IEEE Trans. Microwave Theory and Tech., vol. 40, no. 8 (Aug. 1992): 1692–1700. 19. J. C. Camart, D. Despretz, M. Chive, and J. Pribetich, “Modeling of Various Kinds of Applicators Used for Microwave Hyperthermia Based on the FDTD Method,” IEEE Trans. Microwave Theory and Tech., vol. 44, no. 10 (Oct. 1996): 1811–1818. 20. P. M. Meaney, M. W. Fanning, D. Li, S. P. Poplack, and K. D Paulsen, “A Clinical Prototype for Active Microwave Imaging of the Breast,” IEEE Trans. Microwave Theory Tech., vol. 48 (Nov. 2000): 1841–1853. 21. W. C. Chew and J. H. Lin, “A Frequency-Hopping Approach for Microwave Imaging of Large Inhomogeneous Bodies,” IEEE Microwave Guided Wave Lett., vol. 5 (Dec. 1995): 439–441. 22. O. S. Haddadin and E. S. Ebbini, “Imaging Strongly Scattering Media Using a Multiple Frequency Distorted Born Iterative Method,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. 45 (Nov. 1998): 1485–1496.



38-21

23. Q. Fang, P. M. Meaney, and K. D. Paulsen, “Microwave Image Reconstruction of Tissue Property Dispersion Characteristics Utilizing Multiple-Frequency Information,” IEEE Trans. Microwave Theory Tech., vol. 52 (Aug. 2004): 1866–1875. 24. P. M. Meaney, K. D. Paulsen, A. Hartov, and R. C. Crane, “An Active Microwave Imaging System for Reconstruction of 2-D Electrical Property Distributions,” IEEE Trans. Biomed. Imag., vol. 42 (Oct. 1995): 1017–1026. 25. K. D. Paulsen and P. M. Meaney, “Compensation for Nonactive Array Element Effects in a Microwave Imaging System: Part I—Forward Solution vs. Measured Data Comparison,” IEEE Trans. Med. Imag., vol. 18 (June 1999): 496–507. 26. P. M. Meaney, K. D. Paulsen, M. W. Fanning, and A. Hartov, “Nonactive Antenna Compensation for Fixed-Array Microwave Imaging: Part II—Imaging Results,” IEEE Trans. Med. Imag., vol. 18 (June 1999): 508–518. 27. P. M. Meaney, K. D. Paulsen, A. Hartov, and R. K. Crane, “Microwave Imaging for Tissue Assessment: Initial Evaluation in Multitarget Tissue-Equivalent Phantoms,” IEEE Trans. Biomed. Eng., vol. 43 (Sept. 1996): 878–890. 28. E. C. Fear, S. C. Hagness, P. M. Meaney, M. Okieniewski, and M. Stuchly, “Enhancing Breast Cancer Detection Using Near Field Imaging,” IEEE Microwave Magazine (March 2002): 48–56. 29. S. C. Hagness, A. Taflove, and J. E. Bridges, “Two-dimensional FDTD Analysis of a Pulsed Microwave Confocal System for Breast Cancer Detection: Fixed-Focus and Antenna-Array Sensors,” IEEE Trans. Biomed. Eng., vol. 45 (Dec. 1998): 1470–1479. 30. S. C. Hagness, A. Taflove, and J. E. Bridges, “Three-dimensional FDTD Analysis of a Pulsed Microwave Confocal System for Breast Cancer Detection: Design of an Antenna-Array Element,” IEEE Trans. Antennas Propagat., vol. 47 (May 1999): 783–791. 31. X. Li and S. C. Hagness, “A Confocal Microwave Imaging Algorithm for Breast Cancer Detection,” IEEE Microwave Wireless Comp. Lett., vol. 11 (March 2001): 130–132. 32. E. Fear and M. Stuchly, “Microwave System for Breast Tumor Detection,” IEEE Microwave Guided Wave Lett., vol. 9 (Nov. 1999): 470–472. 33. E. C. Fear and M. A. Stuchly, “Microwave Detection of Breast Cancer,” IEEE Trans. Microwave Theory Tech., vol. 48 (Nov. 2000): 1854–1863. 34. X. Yun, E. C. Fear, and R. H. Johnston, “Compact Antenna for Radar-Based Breast Cancer Detection,” IEEE Trans. Antennas and Propagation, vol. 53, no. 8 (Aug. 2005): 2374–2380. 35. S. C. Hagness, A. Taflove, and J. E. Bridges, “Wideband Ultralow Reverberation Antenna for Biological Sensing,” Electron. Lett., vol. 33, no. 19 (Sept. 1997): 1594–1595. 36. M. A. Hernandez-Lopez, M. Pantoja, M. Fernandez, S. Garcia, A. Bretones, R. Martin, and R. Gomez, “Design of an Ultra-broadband V Antenna for Microwave Detection of Breast Tumors,” Microw. Opt. Tech. Lett., vol. 34, no. 3 (Aug. 2002): 164–166. 37. E. C. Fear and M. A. Stuchly, “Microwave Breast Tumor Detection: Antenna Design and Characterization,” IEEE Antennas Propag. Symp. Dig., vol. 2 (2000): 1076–1079. 38. X. Li, S. C. Hagness, M. K. Choi, and D. W. W. Choi, “Numerical and Experimental Investigation of an Ultrawideband Ridged Pyramidal Horn Antenna with Curved Launching Plane for Pulse Radiation,” IEEE Antennas Wireless Propag. Lett., vol. 2 (2003): 259–262. 39. X. Yun, E. C. Fear, and R. H. Johnston, “Radar-based Microwave Imaging for Breast Cancer Detection: Tumor Sensing with Cross-Polarized Reflections,” IEEE Antennas Propag. Society Symp. Dig., vol. 3 (2004): 2432–2435. 40. C. J. Shannon, E. C. Fear, and M. Okoniewski, “Dielectric-Filled Slotline Bowtie Antenna for Breast Cancer Detection,” Electronics Letters, vol. 41, no. 7 (March 2001). 41. J. M. Sill and E. C. Fear, “Tissue Sensing Adaptive Radar for Breast Cancer Detection: A Study of Immersion Liquid,” Electron. Lett., vol. 41, no. 3 (Feb. 2005): 113–115. 42. J. M. Sill and E. C. Fear, “Tissue Sensing Adaptive Radar for Breast Cancer Detection: Preliminary Experimental Results,” IEEE MTT-S Int. Microwave Symp. Dig., Long Beach, CA, June 2005. 43. J. M. Sill and E. C. Fear, “Tissue Sensing Adaptive Radar for Breast Cancer Detection— Experimental Investigation of Simple Tumor Models,” IEEE Trans. Microwave Theory Tech., vol. 53, no. 11 (Nov. 2005): 3312–3319.




44. D. M. Sullivan, “Three-dimensional Computer Simulation in Deep Regional Hyperthermia Using the FDTD Method,” IEEE Trans. Microwave Theory and Tech., vol. 38, no. 2 (Feb. 1990): 201–211. 45. B. M. Green and M. A. Jensen, “Diversity Performance of Dual-Antenna Handsets Near Operator Tissue,” IEEE. Trans. Antennas and Propagation, vol. 48, no. 7 (July 2000): 1017–1024. 46. O. Gandhi, G. Lazzi, and C. Furse, “Electromagnetic Absorption in the Human Head and Neck for Mobile Telephones at 835 and 1900 MHz,” IEEE Trans. on Microwave Theory and Techn., vol. 44 (1996): 1884–1897. 47. P. Soontornpipit, C. M. Furse, and Y. C. Chung, “Design of Implantable Microstrip Antenna for Communication with Medical Implants,” Special Issue of IEEE Trans. on Microwave Theory and Techn. on Medical Applications and Biological Effects of RF/Microwaves (Sept. 2004). 48. C. E. Reuter, A. Taflove, V. Sathiaseelan, M. Piket-May, and B. B. Mittral, “Unexpected Physical Phenomena Indicated by FDTD Modeling of the Sigma-60 Deep Hyperthermia Applicator,” IEEE Trans. Microwave Theory and Tech., vol. 46, no. 4 (April 1998): 313–319. 49. D. Sullivan, D. Buechler, and F. A. Gibbs, “Comparison of Measured and Simulated Data in an Annular Phased Array Using an Inhomogeneous Phantom,” IEEE Trans. Microwave Theory and Tech., vol. 40, no. 3 (March 1992): 600–604. 50. C. M. Furse, J.-Y. Chen, and O. P. Gandhi, “Use of the Frequency-Dependent Finite-Difference Time-Domain Method for Induced Current and SAR Calculations for a Heterogeneous Model of the Human Body,” IEEE Transactions on Electromagnetic Compatibility (May 1994): 128–133. 51. C. Furse and O. P. Gandhi, “Calculation of Electric Fields and Currents Induced in a MillimeterResolution Human Model at 60 Hz Using the FDTD Method,” Bioelectromagnetics, vol. 19, no. 5 (1998): 293–299. 52. C. M. Furse and O. P. Gandhi, “A Memory Efficient Method of Computing Specific Absorption Rate in CW FDTD Simulations,” IEEE Transactions on Biomedical Engineering, vol. 43, no. 5 (May 1996): 558–560. 53. C. H. Durney, C. C. Johnson, P. W. Barber, H. Massoudi, M. F. Iskander, J. L. Lords, D. K. Ryser, S. J. Allen, and J. C. Mitchell, Radiofrequency Radiation Dosimetry Handbook, 2nd Ed. (Brooks AFB, TX: USAF School of Medicine, 1978). 54. O. P. Gandhi, Y. G. Gu, J. Y. Chen, and H. I. Bassen, “Specific Absorption Rates and Induced Current Distributions in an Anatomically Based Human Model for Plane-Wave Exposures,” Health Physics, vol. 63, no. 3 (1992): 281–290. 55. O. P. Gandhi and C. M. Furse, “Millimeter-Resolution MRI-based Models of the Human Body for Electromagnetic Dosimetry from ELF to Microwave Frequencies,” Voxel Phantom Development: Proc. of an International Workshop, Peter J. Dimbylow (ed.), National Radiological Protection Board, Chilton, UK, July 6–7, 1995. 56. P. J. Dimbylow, “The Development of Realistic Voxel Phantoms for Electromagnetic Field Dosimetry,” Voxel Phantom Development: Proc. of an International Workshop, Peter J. Dimbylow (ed.), National Radiological Protection Board, Chilton, UK, July 6–7, 1995. 57. P. Olley and P. S. Excell, “Classification of a High Resolution Voxel Image of a Human Head,” Voxel Phantom Development: Proc. of an International Workshop held at the National Radiological Protection Board, Chilton, UK, July 6–7, 1995, Peter J. Dimbylow (ed.). 58. M. A. Stuchly, K. Caputa, A. van Wensen, and A. El-Sayed, “Models of Human and Animal Bodies in Electromagnetics,” Voxel Phantom Development: Proc. of an International Workshop, Peter J. Dimbylow (ed.), National Radiological Protection Board, Chilton, UK, July 6–7, 1995. 59. National Library of Medicine, Visible Man Project, MRI scans, CT scans, and photographs available on CD-ROM through Research Systems, Inc., 2995 Wilderness Place, Boulder, CO 80301. 60. C. Gabriel, S. Gabriel, and E. Corthout, “The Dielectric Properties of Biological Tissues: I. Literature survey,” Phys. Med. Biol., vol. 41 (1996): 2231–2249. 61. S. Gabriel, R. W. Lau, and C. Gabriel, “The Dielectric Properties of Biological Tissues: II. Measurements on the Frequency Range 10 Hz to 20 GHz,” Phys. Med. Biol., vol. 41 (1996): 2251–2269.



38-23

62. S. Gabriel, R. W. Lau, and C. Gabriel. “The Dielectric Properties of Biological Tissues: III. Parametric Models for the Dielectric Spectrum of Tissues,” Phys. Med. Biol., vol. 41 (1996): 2271–2293. 63. K. R. Foster and H. P. Schwan, “Dielectric Properties of Tissues and Biological Materials: A Critical Review,” Crit. Rev. Biomed. Eng., vol. 17 (1989): 25–104. 64. S. S. Chaudhary, R. K. Mishra, A. Swarup, and J. M. Thomas, “Dielectric Properties of Normal and Malignant Human Breast Tissues at Radiowave and Microwave Frequencies,” Indian J. Biochem. Biophys., vol. 21 (1984): 76–79. 65. A. J. Surowiec, S. S. Stuchly, J. R. Barr, and A. Swarup, “Dielectric Properties of Breast Carcinoma and the Surrounding Tissues,” IEEE Trans. Biomed. Eng., vol. 35 (April 1988): 257–263. 66. W. T. Joines, Y. Z. Dhenxing, and R. L. Jirtle. “The Measured Electrical Properties of Normal and Malignant Human Tissues from 50 to 900 MHz,” Med. Phys., vol. 21 (1994): 547–550. 67. A. M. Campbell and D. V. Land, “Dielectric Properties of Female Human Breast Tissue Measured in Vitro at 3.2 GHz,” Phys. Med. Biol., vol. 37 (1992): 193–210. 68. C. Gabriel, “Compilation of the Dielectric Properties of Body Tissues at RF and Microwave Frequencies,” Final Report AL/OE-TR-1996-0037 submitted to Occupational and Environmental Health Directorate, RFR Division, 2503 Gillingham Dr., Brooks AFB, TX, June 1996. 69. M. A. Stuchly and S. S. Stuchly, “Dielectric Properties of Biological Substances – Tabulated,” J. Microwave Power, vol. 15, no. 1 (1980): 19–26. 70. S. Rush, J. A. Abildskov, and R. McFee, “Resistivity of Body Tissues at Low Frequencies,” Circ. Research, vol. XII (1963): 40–50. 71. L. A. Geddes and L. E. Baker, “The Specific Resistance of Biological Material—A Compendium of Data for the Biomedical Engineer and Physiologist,” Med. & Biol. Eng., vol. 5 (1967): 271–293. 72. University of Utah Dielectric Database OnLine, http://www.ece.utah.edu/dielectric/. 73. D. Colton and P. Monk, “A New Approach to Detecting Leukemia: Using Computational Electromagnetics,” IEEE Trans. Comput. Sci. Eng., vol. 2 (Winter 1995): 46–52. 74. S. Y. Semenov, A. E. Bulyshev, A. E. Souvorov, R. H. Svenson, Y. E. Sizov, V. Y. Borisov, V. G. Posukh, I. M. Kozlov, A. G. Nazarov, and G. P. Tatsis, “Microwave Tomography: Theoretical and Experimental Investigation of the Iteration Reconstruction Algorithm,” IEEE Trans. Micr. Theory Tech., vol. 46 (Feb. 1998): 133–141. 75. S. Y. Semenov, R. H. Svenson, A. E. Bulyshev, A. E. Souvorov, A. G. Nazarov, Y. E. Sizov, V. G. Posukh, and A. Pavlovsky, “Three-Dimensional Microwave Tomography: Initial Experimental Imaging of Animals,” IEEE Trans. Biomed. Eng., vol. 49 (Jan. 2002): 55–63. 76. K. L. Carr, “Microwave Radiometry: Its Importance to the Detection of Cancer,” IEEE Trans. Microwave Theory Tech., vol. 37, no. 12 (Dec. 1989): 1862–1869. 77. K. L. Carr, “Radiometric Sensing,” IEEE Potentials (April/May 1997): 21–25. 78. L. Dubois, J.-P. Sozanski, V. Tessier, J.-C. Camart, J.-J. Fabre, J. Pribetich, and M. Chiv, “Temperature Control and Thermal Dosimetry by Microwave Radiometry in Hyperthermia,” IEEE Trans. Microwave Theory Tech., vol. 44, no. 10 (Oct. 1996): 1755–1761. 79. S. M. Fraser, D. V. Land, and R. D. Sturrock, “Microwave Thermography—an Index of Inflammatory Disease,” Br. J. Rheumatology, vol. 26 (1987): 37–39. 80. B. Bocquet, J. C. Van de Velde, A. Mamouni, and Y. Leroy, “Microwave Radiometric Imaging at 3 GHz for the Exploration of Breast Tumours,” IEEE Trans. Microwave Theory Tech., vol. 38 (1990): 791–793. 81. J. Robert, J. Edrich, P. Thouvenot, M. Gautherie, and J. M. Escanye, “Millimeter Wave Thermography: Preliminary Clinical Finding on Head and Neck Diseases,” J. Microwave Power, vol. 14 (1979). 82. K. L. Carr, A. M. El Mahdi, and J. Schaeffer, “Dual Mode Microwave System to Enhance Early Detection of Cancer,” IEEE Trans. Microwave Theory Tech., vol. 29 (1980): 256–260. 83. E. A. Cheever, J. B. Leonard, and K. R. Foster, “Depth of Penetration of Fields from Rectangular Apertures into Lossy Media,” IEEE Trans. Microwave Theory Tech., vol. 35 (1987): 865–867.




84. J. Audet, J. C. Bolomey, C. Pichot, D. D. n’Guyen, M. Robillard, M. Chive, and Y. Leroy, “Electrical Characteristics of Waveguide Applicators for Medical Applications,” J. Microwave Power, vol. 15 (1980): 177–186. 85. A. W. Guy, “Electromagnetic Fields and Relative Heating Patterns Due to a Rectangular Aperture Source in Direct Contact with Bilayered Biological Tissue,” IEEE Trans. Microwave Theory Tech., vol. 29 (1971): 214–223. 86. D. V. Land, “Medical Microwave Radiometry and Its Clinical Applications,” IEE Colloquium Application of Microwaves in Medicine (28 Feb. 1995): 2/1–2/5. 87. B. Enander and G. Larson, “Microwave Radiometry Measurements of the Temperature Inside a Body,” Electronic Letters, vol. 10 (1974): 317. 88. J. Edrich and P. C. Hardee, “Thermography at Millimeter Wavelengths,” Proc. IEEE, vol. 62 (1974): 1391–1392. 89. E. A. Cheever and K. R. Foster, “Microwave Radiometry in Living Tissue: What Does It Measure?” IEEE Trans. Biomedical Engineering, vol. 39, no. 6 (June 1992): 563–568. 90. B. Bocquet, J. C. van de Velde, A. Mamouni, Y. Leroy, G. Giauz, J. Delannoy, and D. Delvalee, “Microwave Radiometric Imaging at 3 GHz for the Exploration of Breast Tumors,” IEEE Trans. Microwave Theory Tech., vol. 38, no. 6 (June 1990): 791–793. 91. L. Enel, Y. Leroy, J. C. Van de Velde, and A. Mamouni, “Improved Recognition of Thermal Structures by Microwave Radiometry,” Electron. Lett., vol. 20 (1984): 293–294. 92. Y. Leroy, A. Mamouni, J. C. Van de Velde, B. Bocquet, and B. Dujardin, “Microwave Radiometry for Noninvasive Thermometry,” Automedica (special issue on noninvasive thermometry), vol. 8 (1987): 181–201. 93. C. E. Hayes, W. A. Edelstein, J. F. Schenck, et al, “An Efficient, Highly Homogeneous Radiofrequency Coil for Whole-body NMR Imaging at 1.5 T,” J. Magn. Reson. Imaging, vol. 63 (1985): 622–628. 94. M. C. Leifer, “Theory of the Quadrature Elliptic Birdcage Coil,” Magn. Reson. Med., vol. 38 (1997): 726–732. 95. S. Li, C. M. Collins, B. J. Dardzinski, et al, “A Method to Create an Optimum Current Distribution and Homogeneous B1 Field for Elliptical Birdcage Coils,” Magn. Reson. Med., vol. 37 (1997): 600–608. 96. J. R. Fitzsimmons, J. C. Scott, D. M. Peterson, et al, “Integrated RF Coil with Stabilization for FMRI Human Cortex,” Magn. Reson. Med., vol. 38 (1997): 15–18. 97. L. E. Hendrix, J. A. Strandt, D. L. Daniels, et al, “Three-Dimensional Time-of-Flight MR Angiography with a Surface Coil: Evaluation in 12 Subjects,” American Journal Radiology, vol. 159 (1992): 103–106. 98. P. B. Roemer, W. A. Edelstein, C. E. Hayes, S. P. Souza, and O. M. Mueller, “The NMR Phased Array,” Magn. Reson. Med., vol. 16, no. 2 (1996): 192–225. 99. C. E. Hayes, N. Hattes, and P. B. Roemer, “Volume Imaging with MR Phased Arrays,” Magn. Reson. Med., vol. 18, no. 2 (1991): 309–319. 100. C. E. Hayes and P. B. Roemer, “Noise Correlations in Data Simultaneously Acquired from Multiple Surface Coil Arrays,” Magn. Reson. Med., vol. 16, no. 2 (1991): 181–191. 101. S. M. Wright, R. L. Magin, and J. R. Kelton, “Arrays of Mutually Coupled Receiver Coils: Theory and Application,” Magn. Reson. Med., vol. 17, no. 1 (1991): 252–268. 102. S. M. Wright and L. L. Wald, “Theory and Application of Array Coils in MR Spectroscopy,” NMR Biomed, vol. 10, no. 8 (1997): 394–410. 103. G. R. Duensing, H. R. Brooker, and J. R. Fitzsimmons, “Maximizing Signal-to-Noise Ratio in the Presence of Coil Coupling,” J. Magn. Reson. B., vol. 111, no. 3 (1996): 230–235. 104. D. K. Sodickson and W. J. Manning, “Simultaneous Acquisition of Spatial Harmonics (SMASH): Ultra-fast Imaging with Radiofrequency Coil Arrays,” Magn. Reson. Med., vol. 38 (1997): 591–603. 105. K. P. Pruessmann, M. Weiger, M. B. Scheidegger, and P. Boesiger, “SENSE: Sensitivity Encoding for Fast MRI,” Magn. Reson. Med., vol. 42 (1999): 952–962. 106. Y. Zhu, “Parallel Excitation with an Array of Transmit Coils,” Magn. Reson. Med., vol. 51, no. 4 (2004): 775–784.



38-25

107. K. Y. Kojima, J. Szumowski, R. C. Sheley, et al, “Lower Extremities: MR Angiography with a Unilateral Telescopic Phased-Array Coil,” Radiology, vol. 196 (1995): 871–875. 108. J. W. Monroe, P. Schmalbrock, and D. G. Spigos, “Phased Array Coils for Upper Extremity MRA,” Magn. Reson. Med., vol. 33 (1995): 224–229. 109. C. E. Hayes, C. M. Mathis, and C. Yuan, “Surface Coil Phased Arrays for High-Resolution Imaging of the Carotid Arteries,” J. Magn. Reson. Imaging, vol. 1 (1996): 109–112. 110. C. Yuan, J. W. Murakami, C. E. Hayes, et al, “Phased-Array Magnetic Resonance Imaging of the Carotid Artery Bifurcation: Preliminary Results in Healthy Volunteers and a Patient with Atheroscleroticdisease,” J. Magn. Reson. Imaging, vol. 5 (1995): 561–565. 111. S. H. Faro, S. Vinitski, H. V. Ortega, et al, “Carotid Magnetic Resonance Angiography: Improved Image Quality with Dual 3-inch Surface Coils,” Neuroradiology, vol. 38 (1996): 403–408. 112. H. A. Stark and E. M. Haacke, “Helmet and Cylindrical Shaped CP Array Coils for Brain Imaging: a Comparison of Signal-to-Noise Characteristics,” Proceedings of the International Society for Magnetic Resonance in Medicine (1996): 1412. 113. J. R. Porter, S. M. Wright, and A. Reykowski, “A 16-Element Phased-Array Head Coil,” Magn. Reson. Med., vol. 40 (1998): 272–279. 114. T. Wu and R. King, “The Cylindrical Antenna with Nonreflecting Resistive Loading,” IEEE Trans. Antennas Propag., vol. AP-13, no. 3 (May 1965): 369–373. 115. T. Wu and R. King, “Corrections to ‘The Cylindrical Antenna with Nonreflecting Resistive Loading’,” IEEE Trans. Antennas Propag., vol. AP-13, no. 11 (Nov. 1965): 998. 116. E. C. Fear, J. Sill, and M. A. Stuchly, “Experimental Feasibility Study of Confocal Microwave Imaging for Breast Tumor Detection,” IEEE Trans. Microwave Theory Tech., vol. 51, no. 3 (March 2003): 887–892. 117. Special Issue of IEEE Trans. Microwave Theory Tech., MTT-34, 1986. 118. C. H. Durney and M. F. Iskander, Antenna Handbook, Y. T. Lo and S. W. Lee (eds.) (New York: Springer, 1993). 119. P. K. Sneed and T. L. Phillips, “Combining Hyperthermia and Radiation: How Beneficial?,” Oncology, vol. 5 (1991): 99–108. 120. C. C. Vernon, J. W. Hand, S. B. Field, et al, “Radiotherapy with or Without Hyperthermia in the Treatment of Superficial Localized Breast Cancer: Results from Five Randomized Controlled Trials,” Int. J. Radiat. Oncol. Biol. Phys., vol. 35 (1996): 731–44. 121. F. Monteccia, “Microstrip Antenna Design for Hyperthermia Treatment of Superficial Tumors,” IEEE Trans. BME, vol. 39, no. 6 (June 1992): 580–588. 122. J. Vba, C. Franconi, F. Montecchia, and I. Vannucci, “Evanescent-Mode Applicators (EMA) for Superficial and Subcutaneous Hyperthermia,” IEEE Trans. Biomed. Eng., vol. 40, no. 5 (May 1993): 397–407. 123. M. V. Prior, M. L. D. Lumori, J. W. Hand, G. Lamaitre, C. J. Schneider, and J. D. P. van Dijk, “The Use of a Current Sheet Applicator Array for Superficial Hyperthermia: Incoherent Versus Coherent Operation,” IEEE Trans. Biomed. Eng., vol. 43, no. 7 (July 1995): 694–698. 124. J. Bach Andersen, A. Baun, K. Harmark, L. Heinz, P. Raskmark, and J. Overgaard, “A Hyperthermia System Using a New Type of Inductive Applicator,” IEEE Trans. Biomedical Engineering, vol. 31(1), (1984): 21–27. 125. P. R. Stauffer, M. Leoncini, V. Manfrini, et al, “Dual Concentric Conductor Radiator for Microwave Hyperthermia with Improved Field Uniformity to Periphery of Aperture,” IEICE Trans. on Communicat., vol. E78-B (1995): 826–35. 126. P. F. Maccarini, H. Rolfsnes, D. Neuman, and P. Stauffer, “Optimization of a Dual Concentric Conductor Antenna for Superficial Hyperthermia Applications,” Proceedings of the 26th Annual International Conference of the IEEE EMBS, San Francisco, CA, Sept. 1–5, 2004. 127. H. Tehrani and K. Chang, “Multifrequency Operation of Microstrip Fed Slot Ring Antennas on Thin Low Dielectric Permittivity Substrates,” IEEE Trans. on Antennas and Propagation., vol. 50, no. 9 (Sept. 2002): 1299–1308. 128. S. Jacobsen, P. R. Stauffer, and D. G. Neuman, “Dual-Mode Antenna Design for Microwave Heating and Noninvasive Thermometry of Superficial Tissue Disease,” IEEE Trans. Biomed. Eng., vol. 47 (2000): 1500–9.




129. P. F. Turner, “Interstitial Equal-Phased Arrays for EM Hyperthermia,” IEEE Trans. Micr. Theory and Tech., vol. 34, no. 5 (May 1986): 572–578. 130. R. D. Nevels, G. D. Arndt, G. W. Raffoul, J. R. Carl, and A. Pacifico, “Microwave Catheter Design,” IEEE Trans. Biomed. Eng., vol. 45 (July 1998): 885–890. 131. C. Manry, S. L. Broschat, C.-K. Chou, and J. A. McDougall, “An Eccentrically Coated Asymmetric Antenna Applicator for Intracavity Hyperthermia Treatment of Cancer,” IEEE Trans. Biomed. Eng., vol. 39, no. 9 (Sept. 1992): 935–942. 132. P. F. Turner, “Hyperthermia and Inhomogeneous Tissue Effects Using an Annular Phased Array,” IEEE Trans. Microwave Theory and Tech., vol. 32, no. 8 (Aug. 1984): 874–875. 133. P. Stauffer, J. Schlorff, R. Taschereau, T. Juang, D. Neuman, P. Maccarini, J. Pouliot, and J. Hsu, “Combination Applicator for Simultaneous Heat and Radiation,” Proceedings of the 26th Annual International Conference of the IEEE EMBS, San Francisco, CA, Sept. 1–5, 2004. 134. Y. Kotsuka, E. Hankui, and Y. Shigematsu, “Development of Ferrite Core Applicator System for Deep-Induction Hyperthermia,” IEEE Trans. Micr. Theory and Tech., vol. 44, no. 10 (Oct. 1996): 1803–1810. 135. P. F. Turner, “Sigma 60–24 Prototype Test Results,” BSD Medical Corporation, Internal Rep., Salt Lake City, UT, 1992. 136. J. Nadobny, H. Fähling, M. Hagmann, P. Turner, W. Wlodarczyk, J. Gellermann, P. Deuflhard, and P. Wust, “Experimental and Numerical Investigations of Feed-Point Parameters in a 3-D Hyperthermia Applicator Using Different Models of Feed Networks,” IEEE Trans. Biomed. Eng., vol. 49, no. 11 (Nov. 2002): 1348–1359. 137. J. Nadobny, W. Wlodarczyk, L. Westhoff, J. Gellermann, R. Felix, and P. Wust, “A Clinical Water-Coated Antenna Applicator for MR-Controlled Deep-Body Hyperthermia: A Comparison of Calculated and Measured 3-D Temperature Data Sets,” IEEE Trans. Biomed. Eng., vol. 52, no. 3 (March 2005): 505–519. 138. K. S. Nikita and N. K. Uzunoglu, “Coupling Phenomena in Concentric Multi-Applicator Phased Array Hyperthermia Systems,” IEEE Trans. Microwave Theory and Tech., vol. 44, no. 1 (Jan. 1996): 65–74. 139. F. Bardati, A. Borrani, A. Gerardino, and G. A. Lovisolo, “SAR Optimization in a Phased Array Radiofrequency Hyperthermia System,” IEEE Trans. Biomed. Eng., vol. 42, no. 12 (Dec. 1995): 1201–1207. 140. R. G. Olsen, M. B. Ballinger, T. D. David, and W. G. Lotz, “Rewarming of the Hypothermic Rhesus Monkey with Electromagnetic Radiation,” Bioelectromagnetics, vol. 8 (1987): 183–193. 141. I. D. McRury and D. E. Haines, “Ablation for the Treatment of Arrhythmias,” Proceedings of the IEEE, vol. 84, no. 3: 404–416. 142. A. Labonte, A. Blais, S. Legault, H. O. Ai, and L. Roy, “Monopole Antennas for Microwave Catheter Ablation,” IEEE Trans. on Microwave Theory and Techn., vol. 44, no. 10: 1832–1840. 143. A. Rosen et al, “Percutaneous Transluminal Microwave Balloon Angioplasty,” IEEE Trans. Microwave Theory Tech., vol. 38 (1990): 90–93. 144. A. J. Johansson, “Simulation and Verification of Pacemaker Antennas,” Proceedings of the 25th Annual Int. Conf. of the IEEE EMBS, Cancun, Mexico, Sept. 17–21, 2003. 145. K. Guillory and R. A. Normann, “A 100-Channel System for Real Time Detection and Storage of Extracellular Spike Waveforms,” J. Neurosci. Methods, vol. 91 (1999): 21–29. 146. T. Buchegger et al, “An Ultra-Low Power Transcutaneous Impulse Radio Link for Cochlea Implants,” Joint Ultra Wideband Systems and Technologies (UWBST) and International Workshop on UWBS 2004, IEEE cat. no. 04EX812 (2004): 356–360. 147. K. Gosalia, G. Lazzi, and M. Humayun, “Investigation of Microwave Data Telemetry Link for a Retinal Prosthesis,” IEEE Trans. Microwave Theory and Tech., vol. 52, no. 8 (Aug. 2004): 1925–1932. 148. C. Furse, “Design an Antenna for Pacemaker Communication,” Microwaves & RF (March 2000): 73–76. 149. I. J. Bahl, S. S. Stuchly, J. Lagendijk, and M. Stuchly, “Microstrip Loop Applicators for Medical Applications,” IEEE Trans. MTF (July 1982): 1090–1093.



38-27

150. I. J. Bahl, P. Bhartia, and S. S Stuchly, “Design of Microstrip Antennas Covered with a Dielectric Layer,” IEEE Trans. Antennas Propagat., vol. AP-30, no. 2 (March 1982): 314–318. 151. R. D. Nevels, D. Arndt, J. Carl, G. Raffoul, and A. Pacifico, “Microwave Antenna Design for Myocardial Tissue Ablation Applications,” IEEE Antennas and Propagat. Soc. Mt. Symp., vol. 3 (1995): 1572. 152. C. T. Charles, “Electrical Components for a Fully Implantable Neural Recording System,” Master’s thesis, Electrical and Computer Engineering, University of Utah, Salt Lake City, Utah, 2003. 153. W. G. Scanlon, N. E. Evans, G. C. Crumley, and Z. M. McCreesh, “Low-Power Radio Telemetry: the Potential for Remote Patient Monitoring,” Journal of Telemedicine and Telecare, vol. 2, no. 4 (Dec. 1996): 185. 154. W. G Scanlon, N. E. Evans, and Z. M. McCreesh, “RF Performance of a 418 MHz Radio Telemeter Packaged for Human Vaginal Placement,” IEEE Trans. BME, vol. 44, no. 5 (May 1997): 427–430. 155. W. G Scanlon, N. E. Evans, and J. B. Burns, “FDTD Analysis of Close-Coupled 418 MHz Radiating Devices for Human Biotelemetry,” Physics in Medicine and Biology, vol. 44, no. 2 (Feb. 1999): 335–345. 156. W. G. Scanlon, J. B. Burns, and N. E. Evans, “Radiowave Propagation from a Tissue-Implanted Source at 418 MHz and 916.5 MHz,” IEEE Trans. Biomedical Engineering, vol. 47, no. 4 (April 2000): 527–534. 157. G. C. Crumley, N. E. Evans, J. B. Burns, and T. G. Trouton, “On the Design and Assessment of a 2.45 GHz Radio: Telecommand System for Remote Patient Monitoring,” Medical Engineering and Physics, vol. 20, no. 10 (March 1999): 750–755. 158. C. P. Yue and S. S. Wong, “On-Chip Spiral Inductors with Patterned Ground Shields for Si-based RF IC’s,” Center for Integrated Systems, Stanford University, 1998. 159. G. S. Brindley and W. S. Lewin, “The Sensations Produced by Electrical Stimulation of the Visual Cortex,” J. Physiol., vol. 196 (1968): 479–493. 160. G. E. Loeb, C. J. Samin, J. H. Schulman, and P. R. Troyk, “Injectable Microstimulator for Functional Electrical Stimulation,” Med. Biol. Eng. Comput., vol. 29 (1991): NS 13–NS 19. 161. B. Ziaie, M. D. Nardin, A. R. Coghlan, and K. Najafi, “A Single Channel Implantable Microstimulator for Functional Neuromuscular Stimulation,” IEEE Trans. Biomed. Eng., vol. 44 (Oct. 1997): 909–920. 162. Weimin Sun et al, Implantable Medical Device Microstrip Telemetry Antenna (Jan. 19, 1999): U.S. Pat. 5,861,019. 163. P. R. Troyk and M. A. K. Schwan, “Closed Loop Class E Transcutaneous Power and Data Link for Microimplants,” IEEE Trans. Biomed. Eng., vol. 39 (June 1992): 589–598. 164. G. M. Clark, Y. C. Tong, J. F. Patrick, P. M. Seligman, P. A. Crosby, J. A. Kuzma, and D. K. Money, “A Multi-channel Hearing Prosthesis for Profound-to-Total Hearing Loss,” J. Med. Eng., Technol., vol. 8 (Jan. 1984): 3–8. 165. N. de N. Donaldson and T. A. Perkins, “Analysis of Resonant Coupled Coils in Design of Radio Frequency Transcutaneous Links,” Med. Biol. Eng. Comput., vol. 21 (Sept. 1983): 612–626. 166. D. C. Galbraith, S. Mani, and R. L. White, “A Wide Band Efficient Inductive Transdermal Power and Data Link with Coupling Insensitive Gain,” IEEE Trans. Biomed. Eng., vol. BME-34 (April 1987): 265–275. 167. C. R. Pfaltz (ed.), “The University of Melbourne Nucleus Multi-Electrode Cochlear Implant,” Adv. Oto-Rhino-Laryngol., vol. 38 (1987): 63–81. 168. R. S. Mackay, Bio-Medical Telemetry, 2nd Ed. (New York: IEEE Press, 1993). 169. A. Johansson, “Wireless Communication with Medical Implants: Antennas and Propagation,” PhD dissertation, Lunds Universitet, 2004. 170. M. D. Amundson, J. A. Von Arx, W. J. Linder, P. Rawat, and W. R. Mass, Circumferential Antenna for an Implantable Medical Device (2002): U.S. Pat. 6,456,256 171. L. Griffiths, “Analysis of Wire Antennas for Implantation in the Body,” Master’s thesis, Utah State University, Logan, UT, 2002.




172. M. Sun, M. Mickle, W. Liang, Q. Liu, and R. J. Sclabassi, “Data Communication Between Brain Implants and Computer,” IEEE Trans. on Neural Systems and Rehabilitation Engineering, vol. 11, no. 2 (2003): 189–192. 173. P. Soontornpipit, C. M. Furse, and Y. C. Chung, “Design of Implantable Microstrip Antennas for Communication with Medical Implants,” IEEE Trans. MTT, vol. 52, issue 8 (Aug. 2004): 1944–1951. 174. J. Kim and Y. Rahmat-Samii, “Implanted Antennas Inside a Human Body: Simulations, Designs, and Characterizations,” IEEE Trans. MTT, vol. 52, issue 8 (Aug. 2004): 1934–1943. 175. K. Gosalia, J. Weiland, M. Humayun, and G. Lazzi, “Thermal Elevation in the Human Eye and Head Due to the Operation of a Retinal Prosthesis,” IEEE Trans. Biomedical Engineering, vol. 51, no. 8 (Aug. 2004). 176. M. Ghovanloo and G. Lazzi, “Transcutaneous Magnetic Coupling of Power and Data,” Wiley Encyclopedia of Biomedical Engineering, M. Akay (ed.) (New York: John Wiley & Sons, 2006). 177. C. Furse, G. Lazzi, and O. P. Gandhi, “Dipoles, Monopoles, and Loop Antennas,” (invited chapter), Modern Antennas, C. Balanis (ed.) (Springer, est. publish date 2007). 178. P. Soontornpipit, C. M. Furse, and Y. C. Chung, “Miniaturized Biocompatible Microstrip Antenna Using Genetic Algorithm,” IEEE Trans. Antennas and Propagat. (June 2005): 1939–1945. 179. C. A. L. Bassett, “Bioelectromagnetics in the Service of Medicine,” Electromagnetic Fields and Interactions, Martin Blank, (ed.), American Chemical Society (1995): 262–275. 180. J. C. Weaver, R. Langer, and R. O. Potts, “Tissue Electroporation for Localized Drug Delivery,” Electromagnetic Fields and Interactions, Martin Blank (ed.), American Chemical Society (1995): 301–316. 181. F. Sterzer, “Microwave Medical Devices,” IEEE Microwave Magazine (March 2002): 65–70. 182. E. C. Burdette, F. L. Cain, and J. Seals, “In Vivo Probe Measurement Technique for Determining Dielectric Properties at VHF Through Microwave Frequencies,” IEEE Trans. Microwave Theory Tech., vol. MTT-28 (April 1980): 414–427. 183. T. K. Bose, A. M. Bottreau, and R. Chahine, “Development of a Dipole Probe for the Study of Dielectric Properties of Biological Substances in Radio Frequency and Microwave Region with Time-Domain Reflectometry,” IEEE Trans. on Instr. and Meas., vol. IM-35, no. 1 (March 1986): 56–60. 184. L. Aamodt, M. Manwaring, and K. Manwaring, “In Vivo Brain Tissue Water Measurement,” Proceedings 12th IEEE Symposium on Computer-Based Medical Systems (June 18–20, 1999): 130–135. 185. H. Bassen and G. Smith, “Electric Field Probes—A Review,” IEEE Transact. on Antennas and Propagat., vol. AP-31, no. 5 (Sept. 1983): 710–718. 186. Bassen, “Electric Field Probes for Cell Phone Dosimetry,” Proc. 19th International Congress IEEE/EMBS Society, Chicago, Illinois, Oct. 30–Nov. 2, 1997.



Chapter 39

Automobile Antennas Louis L. Nagy Delphi Research Labs

CONTENTS 39.1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39-2

39.2 AM ANTENNAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39-2

39.3 AM/FM MAST ANTENNAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39-7

39.4 AM/FM WINDOW GLASS ANTENNAS . . . . . . . . . . . . . . . . . . . . 39-10 39.5 AM/FM PANEL ANTENNAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39-14 39.6 FM DIVERSITY ANTENNAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39-15 39.7 RKE AND TPMS ANTENNAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39-16 39.8 UHF/VHF TV ANTENNAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39-17 39.9 MOBILE TELEPHONE ANTENNAS . . . . . . . . . . . . . . . . . . . . . . . . 39-18 39.10

GPS ANTENNAS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39-19

39.11 SDARS ANTENNAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39-26 39.12 EMERGING TECHNOLOGIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39-30 39.13 ANTENNA MEASUREMENTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . 39-31


Automobile Antennas 39-2

CHAPTER THIRTY-NINE

39.1 INTRODUCTION Automobile communications are rapidly expanding with the emergence of many new telecommunication technologies (e.g., XM Radio and Sirius Satellite Radio, WiFi, Bluetooth, mobile TV, Dedicated Short Range Communication, etc.), as illustrated in Figure 39-1. This chapter reviews the antenna technologies that are being used for these various communication systems. In addition, this chapter reviews the antenna requirements as established by vehicle manufacturers in terms of radiation/reception characteristics, cost, aesthetics, reliability, functionality, repairability, design flexibility, RFI/EMC immunity, and buildability (i.e., the ability to place the antenna into the production process of an automobile assembly plant). Remote Function Access: RKE Remote Start etc. Tire Pressure Monitor

AM/FM/DAB Radio

Dedicated Short Range Communication Internet Access Electronic Toll

Blind Spot Radar

Adaptive Cruse Control Radar

Backup Detection Radar Mobile Phone: Cellular GMS 900 Satellite Radio: Mobile TV: GSM 1800 XM UHF/VHF PCS Sirius Satellite UMTS

Collision Avoidance Radar

Satellite Navigation (GPS)

Intra-Vehicle Communication: Bluetooth WI-FI etc.

FIGURE 39-1 Mobile communication services

39.2 AM ANTENNAS The AM broadcast frequency band is from 0.530 to 1.710 MHz with its wavelengths varying from 566 to 175 m. These wavelengths are considerably longer than the maximum dimension of any automobile antenna system (i.e., antenna element and vehicle body structure). Thus, the AM car antenna is considered to be a very small antenna.1 Illustrated in Figure 39-2a is the geometry of an automobile at a height h (where h  l) above the earth’s surface, and its simple electromagnetic model is shown in Figure 39-2b. The AM impedance path through each wheel-bearing and tire structure (relative to the earth’s surface) is modeled by a lump Earth’s surface

Tire-wheel-bearing impedance (Z) Vehicle-earth capacitance (C)

h

Z

Z

Vehicle image (a) Automobile above earth’s surface

(b) Simple electromagnetic model

FIGURE 39-2 AM frequency model for vehicle above earth’s surface



AUTOMOBILE ANTENNAS

impedance element (Z) while the vehicle-earth capacitance is modeled by a lump capacitor element (C). This capacitor can be approximated by using image theory and representing the undercarriage of the vehicle as a flat planar conductor at a height h above the earth’s surface. The dielectric characteristics for various soil conditions (ranging from dry to 16.8 percent moisture) were measured at 1 MHz and found to range from a relative dielectric constant (er) of 2.5 to 20 with a loss tangent (tand ) of 0.017 to 4.0.2 Dry soil has a low loss tangent and can be treated as a lossy dielectric, whereas moist soil has a high loss tangent and can be treated as a lossy conductor. The vehicle-earth capacitance is known to have a significant effect on the AM reception of the vehicle. This capacitance, for a medium-sized automobile, was measured to be approximately 250 pF.3 Early Automobile Antennas Factory-installed vehicle radio systems began to appear in 1923 with the crude addition of a rather large receiver, antenna, battery, and speaker.4 The actual integration of the radio system into the vehicle had to wait until the early 1930s when vehicle manufacturers provided an in-dash radio with a factory-installed antenna. At that time, many of the exterior vehicle panels were made of composite materials (wood, canvas, glass, etc.) that allowed the development of a variety of hidden antenna systems: the Under-Car, Spare Wheel, Running Board, Rooftop Screen (commonly known as the Chicken Wire Antenna), Rooftop Wire (a long wire threaded into the fabric of the roof), and Retractable Mast Antenna (see Figure 39-3). Of these early systems, the rooftop screen antenna gave best reception due to its large size, height above ground, and location away from the car engine. Antenna screen

Insulated radio feed wire

Insulated antenna wire

(a) Under car antenna

(b) Roof screen antenna

Antenna wire

Retractable mast

(c) Roof wire antenna

(d) Retractable mast antenna

FIGURE 39-3 Early AM vehicle antenna systems



CHAPTER THIRTY-NINE

However, a major setback for this antenna came in 1934 in the form of the all metal “Turret Top” roof manufactured by General Motors, which was quickly adopted by other automobile manufacturers.5 The advent of the all-metal roof resulted in the mast antenna becoming the dominant vehicle antenna until the end of the 20th century. The length of these early mast antennas generally ranged from three to five feet. The longer the antenna element, the higher the AM gain, but the more unsightly it became. Mast Antennas A very short monopole antenna mounted on a perfectly conducting infinite ground plane has a radiation resistance given by the following expression (i.e., when the wavelength is much longer than the antenna length, antenna length is much longer than its radius, and current distribution is assumed to be triangular in shape6): RA = 40 (p/l)2

(39-1)

where  is the length of the antenna, l is the wavelength, and  < l /30. This formula can be used to obtain a good approximation for RA for antenna lengths up to l /8. The reactance for a very short monopole is capacitive with values given by the following expressions7: XA ≈ −60[ln(/a) –1]/ [2p/l] (Ω)

(39-2)

CA ≈ [2p/l]/ (60w [ln(/a) –1]) (F)

(39-3)

and

where XA is the antenna reactance, CA is the antenna capacitance, a is the radius of the antenna element, w is the angular frequency (2pf ), f is the RF operating frequency, and a  . A typical automobile mast antenna has an average height () of about 30 inches (0.76 m), and an average diameter (2a) of about 1/8 inch (3.2 mm). When this mast is mounted on a perfectly conducting infinite ground plane and operated at 1 MHz, it will have an /l of 0.0025, a/ of 0.0021, radiation resistance of 0.0025 Ω, reactive impedance of −19.7 KΩ, and capacitance of 8.1 pF. It should be noted that the reactive impedance is several orders of magnitude greater than the radiation resistance. Thus, this antenna can be viewed as a capacitor probe. The radiation efficiency for a small monopole antenna is given by the following equation: e = RA /(RA + ROhmic )

(39-4)

where e is the antenna radiation efficiency, and ROhmic is the ohmic loss of the antenna element. The typical automobile mast element is constructed of materials having high conductivity (e.g., Ni-Cr stainless steel with a conductivity ≈106/Ω m) that results in the antenna currents being confined to the thin outer layer of the conductor. For this current condition, the following equation can be used to determine the ohmic loss of the antenna element8: ROhmic = Rs /3P

(39-5)

where Rs = (wm/2s)1/2 is the high-frequency surface resistance, P is the circumference (2pa), m is the permeability, and s is the conductivity of the antenna element. A Ni-Cr stainless steel mast antenna with a height of 30 inches, diameter of 1/8 inch, operated at 1 MHz,



AUTOMOBILE ANTENNAS

100 Efficiency

Efficiency (%)

50 % efficiency at 7.3 MHz

50

0 0.1

1

10

Frequency (MHz) FIGURE 39-4 Efficiency of a very small monopole antenna

and mounted on a perfectly conducting infinite ground plane has an RA of 0.0025 Ω, Rs of 0.002 Ω, ROhmic of 0.050 Ω, and radiation efficiency of 4.7 percent. A plot of the efficiency versus frequency is contained in Figure 39-4 for the frequency range from 0.1 to 10 MHz. It should be noted that this simple antenna model does not take into account any of the effects associated with the vehicle body structure. The selection of the coaxial cable, from the antenna to the radio, is important in achieving quality AM reception. Shown in Figure 39-5 is a simple electrical model for the automobile’s AM antenna-radio system (i.e., a Norton equivalent circuit with the negligible resistance and inductance components removed), where IAN = inducted short-circuit antenna current, IR = current passing through the radio capacitor, CA = antenna capacitance (including vehicle structure effects), CC = coaxial cable capacitance, and CR = radio capacitance. The three modeled capacitors form a current divider circuit with the radio current given by IR = IAN (CR /(CR + CC + CA)). The input capacitance of the radio is about 75 pF, capacitance of the 30-inch mast antenna is about 10–15 pF, and antenna base-mount capacitance is about 15 pF. The length of coaxial cable for a front-mounted fender mast antenna is approximately 5 feet, while the length for a rear-mounted fender mast can be over 20 feet. Norton equivalent antenna circuit

IR

CR

CC

CA

IAN

FIGURE 39-5 Norton equivalent circuit for vehicle AM antenna-radio system



CHAPTER THIRTY-NINE

Because of the increase in cable capacitance, the front-mounted mast antenna system (including its associated coaxial cable) has an effective AM gain several dB greater than a rear-mounted mast antenna system (assuming the same type coaxial cable is used for both antennas). In order to avoid excessive AM signal loss, the rear-mounted antenna requires the use of a very low-capacitance/length cable ( 1/2, and when the handedness is opposite, the efficiency is < ½ yielding the polarization isolation. The second order (variance) statistics become VP = (1 / 2π )∫

2π 0

(η p − Ep )2 d ∆

= B2/2 so that the standard deviation equals B/√2. Since these statistics have a non-zero mean value, the second order statistics are generally expressed as +1s about the mean value, also known as the rms spread about the mean. Notice that the polarization efficiency statistics are not Gaussian. The peak to peak bounds are +B for these statistics while the rms spread is +0.707 B. Thus, the bounds equal 1.41 times the standard deviation. Example statistics for isolation and polarization mismatch loss in Figure 44-1 illustrate the variations, assuming an incident field with a 1 dB axial ratio and receiving antennas with axial ratio values as great as 4 dB. The isolation has a finite value even with an ideally polarized receiving antenna resulting from the cross polarization of the incident field and also has a possibility to achieve ideal cancellation when received by an orthogonal polarization, 1 dB axial ratio receiving antenna. The polarization mismatch has minimal values for the range of receiving axial ratio values and again, zero loss is probable with an ideally matched 1 dB axial ratio receiving antenna. It is also apparent these statistics do not have a Gaussian distribution. This example illustrates that achieving high values of polarization isolation require significant design attention to the polarization purity of the link antennas and minimizing depolarizing effects in the link. Military communication systems generally have spot coverage requirements for serving military theaters and this coverage is mechanically repositioned to respond to changing geopolitical requirements. Interference threats and issues of privacy are also concerns to military systems. These concerns are addressed through spread spectrum modulation formats, error correction techniques, data interleaving, and encryption, as well as the protection that can be achieved through sidelobe control and adaptive antenna processing8 to negate 0

Polarization Mismatch Loss, dB

−10 Polarization Isolation, dB

rms Spread Minimum −20 Mean

Maximum

−30

Minimum

−0.1

rms Spread

−0.2

1 dB Incident Axial Ratio −40

Mean

Maximum

1 dB Incident axial ratio −0.3

0

1

2 Receive Axial Ratio, dB (a) Polarization Isolation

FIGURE 44-1

3

4

0

1

2 Receive Axial Ratio, dB

3

4

(b) Polarization Mismatch Loss

Polarization statistics (after R. B. Dybdal7 © IEEE 1999)


Satellite Antennas 44-8

CHAPTER FORTY-FOUR

uplink interference. Sidelobe protection reduces interference power beyond the design coverage area. Adaptive processing using a cluster of uplink beams with a reflector design9 provides the ability to negate interference within the coverage area. Reflector antenna technology,10 for adaptive uplink antennas, is generally preferred to array technology both for performance reasons and SWaP. Rather than using a fixed combination of beams to conform to a specified coverage shape, the adaptive design provides amplitude and phase control to dynamically produce antenna pattern nulls in the direction of interference sources. Because these systems generally use spread spectrum modulation as well, for interference protection, the interference power must significantly exceed the user power if the interference is to be effective. Adaptive antennas must: 1. 2. 3. 4.

Distinguish desired signals from interference Respond to interference that can degrade system performance Combine antenna elements with adaptive weighting circuitry to reduce interference Provide a control processor to derive the adaptive weight values to satisfy an optimization criteria 5. Monitor and respond to changes in the interference. Since spread spectrum techniques are generally used, interference is distinguished from desired signals on the basis of the spreading technique. The adaptive design is generally activated when a threshold power level is exceeded, indicating interference signals that are sufficiently strong to exceed the capabilities of other interference protection techniques. Correlation techniques are used to determine the amplitude and phase weighting of the individual beams. Iterative correlation measurements are typically processed by an algorithm based on least mean square measures. This algorithm is derived to satisfy an optimization criterion such as maximum S/(N + I ) or signal to noise plus interference ratio. An interference scenario is specified to guide the adaptive antenna development and defines the number, geographic distribution, and RF performance of the interference sources. Adaptive system design11 proceeds by developing a simulation program based on a candidate RF design, the RF circuitry, and the correlation and control electronics associated with the design. Two types of performance measures are required, one specifying the transient time in responding to interference initiation and the second measuring steady state performance. When nulls are produced within the coverage area, portions of the coverage area have insufficient performance to permit link closure for the users. The percent coverage area is a quantitative measure of performance and is defined as the fraction of the original design coverage area remaining after interference cancellation that permits user operation. A threshold S/(N + I) is determined where acceptable communications as measured by a BER value can occur. This threshold S/(N + I) level provides a means to define a coverage boundary and hence determine the percent coverage area. These measures depend on specific interference source locations. The simulation program varies the interference locations, power levels, spectral contents, and so on, as defined by the scenario in a Monte Carlo manner, in order to obtain statistical answers to the transient time and percent coverage characteristics as needed to evaluate adaptive system performance. Selected cases that are representative of the scenario are used to develop a test program, since it is impractical to test on a Monte Carlo basis. These test cases are then used to validate the simulation and the statistical measures of performance are obtained from the validated simulation. The minimum angular separation between users and interference sources is an important parameter for adaptive uplink antenna design. Statistical answers for this parameter


Satellite Antennas SATELLITE ANTENNAS

44-9

can be obtained from the adaptive system simulation. The separation9 depends on the minimum beamwidth that can be formed by the antenna, the performance above a threshold level for a user, and the incidence angle on the earth’s surface. An estimated separation distance in Figure 44-2 illustrates these factors as a function of the minimum beamwidth of the aperture and is based on aligning the antenna pattern null between the main beam and first sidelobe with the interference source. These systems, particularly at the higher EHF frequencies, are margined to maintain link availability during rain, and hence, under clear conditions, the margin can be reduced in the interest of interference negation and still maintain operation at the threshold S/(N + I). Two values of antenna gain loss, 3 and 10 dB, are illustrated in Figure 44-2 ; if 10 dB of margin loss can be tolerated, the system can maintain threshold performance closer to the interference than if only 3 dB of margin loss can be accommodated, for example. Finally, the minimum separation differs between the subsatellite point and a lower 20° elevation angle because the beams spread further over the earth’s surface at the elevation angles. These values provide a first order estimate of the required aperture size for meeting the desired angular resolution between users and interference, and are supported by more detailed simulation results. A further parameter of interest is the achievable amount of interference suppression. Interference cancellation is inherently a subtraction process, and the cancellation depth depends on the ability to match the amplitude and phase responses in the adaptive channels over the required bandwidth. The tolerance requirements in Figure 44-3 become increasingly stringent as the required interference suppression increases. The protection of high data rate services12 in a military area benefits from using a high gain space segment antenna and from the operation of EHF frequencies with wide bandwidth frequency allocations. Antenna pointing must be maintained and the attitude stability of the satellite becomes a significant contributor to the antenna pointing loss. One solution13 lies in actively tracking a ground beacon to maintain pointing; for this a coded beacon signal and correlation processing on the satellite can be used. This technique can also be used to reduce interference. The narrow beamwidth limits interference from ground-based sources removed from the ground terminal location. Because system transfers high data rate signals, only limited interference protection is afforded by spread spectrum modulation. Protection from interference located close to the ground terminal is provided by the narrow beamwidth.

FIGURE 44-2 Interference/User angular separation (after K. M. Soo Hoo and R. B. Dybdal9 © IEEE 1989)



CHAPTER FORTY-FOUR

FIGURE 44-3 Amplitude and phase tolerance for adaptive antennas

Additional protection can be obtained by offsetting the satellite antenna from the ground terminal in a direction away from the interference source, as illustrated in Figure 44-4. The ratio of the signal to interference power increases from R to R’ as indicated as the beam is repositioned, with some loss in the received S/N. Because a significant rain margin is typically required for link availability, some loss in peak gain can generally be sacrificed in order to gain protection from interference close to the ground terminal. As the beam Example Interference Direction 0

Relative Level, dB

R −5

R’ Boresight Reposition

−10

−15

−20

−1

−0.5

0

0.5

1

Angle, beamwidths FIGURE 44-4 Adaptive beam repositioning illustration (after R. B. Dybdal and S. J. Curry13 © IEEE 1996)



SATELLITE ANTENNAS

0

Margin Loss, dB

−5 −10 −15 −20 −25 −30

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Signal/Interference Separation, beamwidths FIGURE 44-5 Margin loss with beam repositioning (after R. B. Dybdal and S. J. Curry13 © IEEE 1996)

is offset from the ground terminal away from the interference source, the interference is reduced more than the desired signal because of the increased slope of the antenna pattern. As the signal margin permits, it is desirable to offset the narrow beam in order to align the null between the main beam and the first sidelobe with the interference source to improve the S/(N + I). The required margin is shown in Figure 44-5 as a function of the angular separation between the user and interference source. The implementation of adaptive beam repositioning follows the functional diagram in Figure 44-6. The desired signal is accompanied by a pseudo-random code for tracking purposes. The correlation of the data (Σ channel) and the tracking (∆ channel) signals produces outputs for tracking that are not affected by interference, because the interference signal is not correlated with the coded desired signal. Thus, independent of fluctuations in the desired received signal level that may result from weather events, the desired signal location is always determined. When interference is present, the cross correlation of the data

Data Receiver

– Antenna –

Correlator

Signal and Interference Components

Code Generator

Correlator

Correlator

Control System

FIGURE 44-6 © IEEE 1996)

Interference-free Tracking Components

Processor

Functional diagram of adaptive beam repositioning (after R. B. Dybdal and S. J. Curry13



CHAPTER FORTY-FOUR

and tracking channels contains components that include both the desired signal and interference; additionally, power meters on the data and tracking channels can also indicate interference initiation. The cross correlation output level indicates the interference strength and open loop commands can be initiated to reduce the interference as indicated in the cross correlation output. During such commanding, the desired signal level in both the data and tracking channels can be monitored independent of the interference levels, and the signal reduction relative to the margin and the signal growth in the tracking channel can be observed. Antennas for spot beam applications are typically reflector technology, since a limited field of view is required and reflector antennas have attractive SWaP. In many cases, the number of antennas that can be used is limited by the available space on the satellite. One approach14 for systems using multiple frequency bands integrated several such bands into the same aperture to provide more compact designs. This design example provides both shaped and multiple beam capabilities. In many cases, frequency selective surface technology can be used to isolate frequency bands in multiple frequency designs. More commonly, the coverage requirements must be maintained for both the uplink and downlink services that use different frequency bands, the required aperture sizes for the uplink and downlink antennas differ, and separate uplink and downlink antennas are often used to maintain the same coverage characteristics at both frequencies to reduce design complexity. Multiple Beam Antennas In many applications, the system design requires covering large portions of the field of view with high G/ T and ERP beams for link closure reasons and/or achieving greater data capacity than the frequency allocation can support, if a single coverage beam is used. Multiple beam antenna technology provides the means to meet these objectives for such applications. The field of view is served by a collection of beams where the narrow antenna beamwidth is used to meet G/ T and ERP requirements. Isolation between individual beams permits frequency subbands to be reused in separated coverage areas and this frequency reuse increases the system data capacity. Multiple beam antennas typically use reflector antennas with a cluster of feed elements in the focal region to generated independent beams in different directions. The beams are arranged on an equilateral triangular grid. Commonly, a central beam is surrounded by a ring of six feeds surrounded again by a ring of 12 feeds, then by a ring of 18 feeds, and so on. This 91 Beam Design results in the sequence, 7, 19, 37, 61, 91, … and so on for the number of beams to service a coverage area. Figure 44-7 illustrates an example beam pattern for 91 beams. The required number of beams greatly increases as the beamwidth of the individual beam decreases, and the peak gain for each beam also increases. Figure 44-8 illustrates a typical variation of these parameters for the complete field of view coverage of geosynchronous satellites, along with the required aperture size at selected frequencies. The required aperture size variation is also indicated in this figure for selected frequencies. As the beamwidth of the individual beams becomes narrower, not only is the system complexity increased by the number of beams, at lower frequencies, deployment of the reflector surface is also required FIGURE 44-7 Multiple beam arrangement for 91 beams at lower operating frequencies.



SATELLITE ANTENNAS 10

60

100

1000

Peak Antenna Gain, dB

Number of beams

Number of beams 50

40

Gain 10000

30 4

2

1

0.8

0.6

Size (inches) required at

0.4

Half power beamwidth, deg 24

48

6

12 6

72 18 12

120 24 18

8 GHz

168 36

24

48 36

30 GHz

60

44 GHz 48

FIGURE 44-8 Multiple beam parameters

A design conflict15 results between the desire for beams with low sidelobes to isolate beam positions and the desire for the beams to crossover at a high pattern level to maximize the minimum gain within the coverage area. Reduced sidelobe levels require an aperture distribution with amplitude tapering that in turn requires directive antenna feeds. However, the antenna feed size is limited by the inability for the antenna feeds to physically overlap. Two different approaches have been investigated for addressing this conflict between beam overlap and sidelobe levels. One approach16 defocuses the feed cluster and under-illuminates the reflector. A second approach17 develops low sidelobes by combining a feed with a weighted sum of the six adjacent feeds, which results in a complex beamformer. The scan performance of reflector antennas18 is limited by phase aberrations for feed positions removed from the reflector’s focal point. Reflectors with a long f /D ratio and dual reflector designs benefiting from the magnification ratio are means for reducing the phase aberrations and producing high quality beams over a field of view. Multiple beams commonly use frequency reuse techniques to extend system capacity. Frequency reuse imposes isolation requirements between beam positions that share a common frequency subband. The isolation between beam positions depends on the sidelobe response of the closest beams sharing the same subband. The required isolation depends on the susceptibility of the modulation to cochannel interference and the dynamic range of the users. In practice, these two factors are used to devise a frequency reuse plan defining the subband assignments to individual beams within the multiple beam collection. Since the beams are typically arranged on an equilateral triangular grid, the minimum number of subbands to avoid the same subband in adjacent beam is three. More commonly, a frequency reuse plan based on seven frequency subbands is selected to provide adequate isolation. Multiple beam antennas can also been used in a limited portion of the earth’s field of view19 to provide high data rate, high capability communications to a limited geographical area. As the individual beamwidths of the design become narrower, their width can become



CHAPTER FORTY-FOUR

comparable to the attitude stability of the spacecraft. In such cases, actively tracking ground beacons20 can be used to maintain beam pointing. Such designs are attractive for military communications to theater operations. Not only are the high data capacity advantages provided by such designs, protection from interference is afforded by the narrow beamwidths and the effectiveness such interference is limited to a few beam positions. In multiple beam operation, capacity demands are often not uniformly distributed over the collection of beams and can dynamically vary. One approach21 to provide high capacity in a limited number of beam positions would use a frequency reuse plan providing a basic capacity to the individual beams within the collection along with separate frequency subbands that can be switched to beam positions with higher capacity demands as required. Because these frequency subbands do not overlap the subbands in the basic service, the necessary isolation is provided. As memory technology expands and becomes space qualified, storing non-time critical data for transmission at a later time when capacity is available may also provide a means to lower the peak capacity requirements. While such approaches add flexibility and capacity to the system design, they also add complexity and the need for network control. The preponderance of multiple antenna technology for GEO satellites is reflector designs because the earth subtends a relatively small field of view. When multiple beam designs are used at lower microwave frequencies, the required aperture size requires the use of deployable reflector technology. When multiple beam designs are used for satellites in lower LEO orbits, beam generation over a wider field of view is required. For such applications, array technology is used to produce relatively broad beams over the larger field of view. Such array designs22–23 have a manageable number of elements, whereas applying array technology at the higher GEO altitudes results in an excessive number of elements. Array designs require active receive and transmit elements to offset losses in their corporate feed networks resulting in a prime power penalty. Transmit arrays require attention to the transmitter operating points because intermodulation of carrier components when multiple beams are routed through the transmit modules degrade performance. Failure of active elements has a minor effect on the peak gain levels of array antennas, but the sidelobe response24 does not have the same graceful degradation. If a small number of elements fail, rephasing can maintain sidelobe performance, but as the number of failed elements increase, rephasing cannot maintain sidelobe performance. Further attention is also required to devise on-orbit calibration techniques to identify and verify array element excitation errors. Design attention is required to establish performance projected to the endof-life levels accounting for the anticipated number of element failures. Further, the array elements must be commanded to produce the required beam patterns and beam steering, and such control also incurs additional weight and power impacts. As digital technology matures, digital beamforming techniques will evolve. Development experience for this technology25 is described for a multiple beam design, albeit for a mobile user terminal. Crosslink Antennas Communication links between satellites provide a means to relay information between system users that are not in the field of view of a single satellite, and are referred to as crosslinks. Crosslinks permit global linkage of users without using a relay through ground stations. The antennas used in crosslinks are point-to-point configurations in contrast to other spacecraft antennas that serve a coverage area on the earth’s surface. Antennas for crosslink applications are typically reflector designs mechanically pointed at the corresponding antennas on other satellites within the constellation. Such designs use antenna tracking26 to maintain beam alignment with their destination satellite. For GEO satellites, the antenna must have the capability to track other satellites within the equatorial plane and accommodate the elevation variations that depend on how well the satellite is stationkept; at GEO, orbital perturbations result in satellites following a figure eight motion



44-15

about the equatorial plane over a 24-hour period. Spacecraft thrusters control the size of this figure eight maintaining the satellite’s position on the equatorial plane. This adjustment of the satellite’s orbit is referred to as stationkeeping. Crosslink antennas, therefore, are required to provide principally azimuth positioning to link with different satellites within the constellation and a smaller elevation positioning requirement dependent on how well the satellites maintain stationkeeping. Crosslink systems often operate at 60 GHz where wide bandwidth allocations exist to transfer high data rates. The 60 GHz frequency range corresponds to the resonances of oxygen absorption spectra27,28 and provides broad bandwidth, high level attenuation of terrestrial signals. Thus, operation at 60 GHz also isolates the crosslinks from ground based interference.29 Satellite crosslinks often transfer high data rate data between satellites. The EHF operation also results in antennas with a compact physical size—an important factor in satellite systems. These antennas achieve high gain performance and their corresponding narrow beamwidths require antenna tracking capabilities to maintain beam alignment with adjacent satellites. TT&C Antennas All satellites require a TT&C subsystem to provide information on the satellite’s orbital position, monitor the health and status of the spacecraft, and to inject and validate commands for the satellite’s operation. The spacecraft’s TT&C subsystem communicates to a dedicated mission control station that maintains the operation of the satellite. Antenna systems for TT&C applications require two types of antenna coverage. During launch operations, complete spherical coverage is required so if the spacecraft begins to tumble, a command can be injected to attempt to control the spacecraft independent of the satellite’s orientation. Once on orbit, an earth coverage pattern is required so that alternative mission control stations can perform the TT&C operations. However, because the data rate requirements for TT&C applications are modest and ground terminal designs with ample performance can be configured, many satellite programs simply use the earth facing hemispheric coverage antenna to reduce design complexity. Two antennas that provide hemispheric coverage are typically used to satisfy the spherical coverage requirement. One antenna covers the forward hemisphere and is also used in the on orbit position, and the second covers the rear hemisphere. During launch operations, the antenna with the higher signal power is used. Attempting to coherently combine the two antennas results in significant pattern ripple where their patterns overlap because of the physical separation between the two antennas. These broad hemispheric coverage requirements have the generic problem of isolating the TT&C antenna from the satellite blockage and scattering. Such systems typically operate at lower microwave frequencies with a relatively long wavelength. The long wavelength further limits the ability to control antenna backlobes that interact with the satellite’s structure with a compact design. The receive and transmit frequencies can also be a relatively large percentage bandwidth. To satisfy these diverse requirements, generally a frequency independent antenna is used and the antennas are mounted on booms to provide a clear field of view and to minimize the interactions with the spacecraft structure. Design attention to reducing the backlobes of TT&C antennas30 and the development of analysis and measurement techniques for projecting and verifying the performance of TT&C antennas in the presence of the spacecraft structure would be useful capabilities. Remote Sensing Antennas Remote sensing satellites can use RF sensors31 to collect data. Two types of sensors have been developed: active or radar sensors and passive or radiometric sensors. Such sensors



CHAPTER FORTY-FOUR

are used in low orbiting satellites for global coverage characteristics and high resolution of surface features. Radar sensors32 for remote sensing applications use array antennas operated in a synthetic aperture mode. An example design is the SIR-C radar flown on the space shuttle. This design uses two active array antennas, one operating a L-band and the other operating at C-band. Additionally, the program also uses an X-band slotted waveguide array. Such a design approach allows the evaluation of image quality at different frequencies. The arrays are also capable of operation using orthogonally polarized beams so that the polarization contrast can be judged. Like other spaceborne antennas, SWaP has a significant impact on the design. Further development of this technology can be anticipated in the future; as the remote sensing data gathered by this effort provides a proof of concept and an example data base with which to refine estimates of remote sensing extraction algorithms. Two categories of radiometric designs are used in passive RF remote sensing applications.33 One category uses narrow antenna beamwidths that are mechanically scanned to obtain data across a terrain swath beneath the satellite. The beam is scanned across the swath and the spacecraft orbital trajectory fills in the swath in subsequent scans. A portion of this scan looks at reference temperatures to calibrate the radiometer, a “hot” reference is provided by an ambient temperature and a “cold” reference is provided by the deep space background with a 3 K temperature. Radiometric sensors receive data from the emission of the surfaces and contrasts of the surface emission properties allow the generation of an image. Such sensors normally look obliquely at the earth’s surface and the combination of the polarization dependence of the surface emissivity and multiple frequency data allows estimates of remote sensing parameters. The resolution of the emission temperatures depends on the ∆ T of the radiometer that is defined by ∆T = Ts /(Bt)1/2 where Ts is the total system temperature, B is the IF bandwidth, and t is the integration time. Physically ∆T is the standard deviation of the output temperature. For specification purposes of the radiometric sensitivity, a 290 K antenna temperature is used as a reference value and the integration time is 1 second. In practice, the resolution of the emission temperature depends on the integration time at each pixel (picture element). Design attention to minimizing RF loss and to utilizing low noise amplifiers is necessary to achieve good resolution in emission temperature as measured by a low ∆T. The second performance parameter for radiometric system concerns the spatial resolution of the design. The spatial resolution is measured by the antenna’s solid beam efficiency. The solid beam efficiency is the ratio of the integrated antenna pattern over a specified solid angle for a particular polarization to the total integrated pattern for all polarizations. The sense of this definition is to examine the collected emission over a specified solid angle in comparison to the emission collected over the entire pattern. The solid beam efficiency measured the ability of the antenna to respond to its intended pixel coverage and not the surrounding radiometric background. High solid beam efficiency values require both high polarization purity and attention to reducing the antenna sidelobes by tapering the aperture’s amplitude distribution and maintaining a uniform phase distribution. A second type of radiometric is referred to as a sounder, which uses simple broad coverage antennas. Sounders use frequencies corresponding to the molecular absorption frequencies. The center frequency corresponds to the peak absorption value and several IF bandwidths are used. As the IF bandwidth increases, the absorption at the band edges decreases. Using different IF bandwidths results in a height profile of the radiometric response results, meaning that wider IF bandwidths respond to absorption characteristics at lower altitude levels.



44-17

44.3 USER SEGMENT ANTENNAS Like the space segment, the user segment of satellite systems has a wide array of antenna technology. The early user segments required relatively large ground terminal antennas because of the limited satellite performance, the broad coverage space segment antennas necessitated by uncertain satellite attitudes, and the high receiver noise temperatures of early designs. The large ground terminals generally served as a hub for distribution by terrestrial segments. Improvements in satellite capabilities and user technology availability results in satellite services presently being extended to the personal level with examples being cost effective handheld GPS receivers and direct broadcast services. Such trends can be anticipated to continue in the future. These trends prompt the development of cost effective designs to control total system costs and appeal to individual users. These trends also pose interesting challenges in testing on a production level. At a system level, like the space segment antennas, the user segment antennas have G/ T and ERP requirements to meet the specified link performance. Unlike the space segment antennas, which are required to meet G/ T and ERP performance over a specified coverage area, the user segments must align their main beams with the satellite, so that the peak antenna gain values burdened with the antenna pointing loss determines the link performance. While the user segment exploits a diversity of antenna technology, common issues exist. These issues include antenna temperature characterization, interference protection, diplexing, and production issues. Antenna Temperature Unlike space segment antennas, the antenna temperature for the user segment depends on the propagation path and its loss. Propagation loss, in common with ohmic loss in RF circuitry and antenna feeds, both attenuates the received signal and generates additional noise. In both cases, the increased noise is a consequence of blackbody radiation. At microwave frequencies, propagation loss depends on the water vapor and oxygen absorption characteristics in the atmosphere.34 At higher EHF frequencies, the propagation loss is weather dependent.35–36 The principal impact results from raindrops within the propagation path, because their size becomes comparable to the EHF wavelength. The loss results from the absorption cross section of rain drops. Other hydrometeors, such as fog, are smaller, and then there are ice crystals and snow, which are both even smaller and have reduced absorption. Statistical projections of path loss at EHF frequencies are commonly made using climate models of the geographical locations based on cumulative rainfall rate statistics. The specific attenuation (dB/km) of the path has an aRb dependence,37 where R is the rain rate (mm/hr), and the a and b constants depend on the operating frequency. Using the cumulative rain rate statistics and a freezing height, above which rain drops become ice crystals with little loss, projections can be made of link availability. In addition to loss in the propagation path, antennas enclosed by radomes also suffer performance degradation when radome surfaces are wet38–39 and must be factored into projections of link availability. These link availability statistics are stationary only in a long-term multi-year sense—for example, some years are drier than others. While such analyses provide a basis of establishing rain margin requirements, the availability of a given system at a particular time is subject to meteorological conditions prevailing at that time. Like other loss components, such as ohmic loss, the propagation loss not only reduces the signal level but also increases the system noise level. The increase in the system noise temperature for user antenna contrasts with uplink satellite receiving antennas where the 290 K earth background antenna is not increased by noise contributions from loss. Practical antenna systems likewise contain ohmic loss and other necessary components such as filters



CHAPTER FORTY-FOUR

unavoidably contain loss. Today’s availability of low noise receiver technology often results in receiver noise temperatures that are less than the total antenna noise temperature. Because the user’s G/ T dictates the received signal sensitivity, attention must be paid to the total system noise temperature T that equals the sum of the antenna temperature and the receiver temperature. In the past, the achievable receiver noise temperature greatly exceeded the antenna noise temperature and consequently, simply reducing the receiver noise temperature almost directly reduced the system noise temperature. This is not the case with available low noise receiver technology. Antenna temperature for user systems depends not only on the operating frequency but also the propagation path. The loss increases as the elevation angle to the satellite decreases, because the path traveling through atmosphere where loss occurs is longer. The antenna temperature depends on the emission temperature of the environment surrounding the antenna. Finally, practical antennas are not ideally lossless and matched to the receiver input impedance, and corrections must be made for those effects. The total system noise temperature must be referenced to a specified elevation angle because the antenna temperature depends on the path length through the atmospheric loss. At low elevation angles, the antennas are limited by obscura from natural terrain and manmade structures specific to the site. As a consequence, a moderate elevation angle, for example, 20°, is often selected for specification purposes. Such a choice allows compliance measurements without the measurement uncertainties at the lower elevation angles. In practice, however, the elevation angles used operationally must be addressed and at the lower elevation angles, the antenna temperature can depend on the azimuth angle because of obscura; in such cases, blockage may be a further problem. These concerns must be addressed when a site selection is made for user antennas. The antenna and receiver noise temperatures comprise the total system noise temperature. The user terminal’s G/ T requires determining the antenna gain and the system noise temperature at the same terminal reference plane. In many cases, the input to the receiver’s LNA is a convenient terminal plane because the receiver temperature can be measured at the LNA input, using standard Y factor hot/cold load measurements. Additionally in many cases, the LNA input is convenient to terminate in a matched load to provide a stable noise temperature value for measuring the antenna temperature. The total system temperature for a terminal is comprised of the antenna temperature at the reference plane and the receiver noise temperature at the reference plane. The system temperature at this reference plane40 Ts is given by Ts = (1−Γ2) [TantL + 290(1 – L)] + Trec where Γ is the magnitude of the reflection coefficient (for well matched systems, (1 – Γ2) is very close to 1), Tant is the antenna temperature, L is the loss, and Trec is the receiver noise temperature. Physically, the noise power received by a lossless antenna is reduced by that loss but the noise generated by the loss must be included. If significant impedance mismatch loss (1 – Γ2) exists, the noise incident on the LNA must be reduced by the mismatch loss. The antenna temperature can be established through measurement. If such measurements are not available, the antenna temperature of a lossless antenna can be calculated from the pattern characteristics and the emission background levels. Such a procedure has been done41 and verified by measurement of a Cassegrain reflector antenna, as shown in Figure 44-9. The antenna temperature differs from the sky temperature values,34 particularly at the lower elevation (high zenith) angles, because of ground emission coupling through the antenna sidelobes. The sky temperature values correspond to the noise received by an ideally lossless antenna with an infinitely small beamwidth. The measured and calcu-



44-19

FIGURE 44-9 Calculated and measured antenna temperature (after K. M. Lambert and R. C. Rudduck41© Radio Science 1992)

lated antenna temperatures were performed using very low loss antenna. Practical antenna systems must be corrected for loss in filters and other RF components needed for system operation. Often, data for only the antenna is available and when determining the G/ T the gain must be reduced by the RF loss and the antenna temperature likewise compensated by the RF loss as indicated. Interference Issues The utilization of satellite and terrestrial systems has greatly expanded in recent years, a trend that can be expected to continue. As a consequence, interference between systems increases both as a possibility and as a concern. Two distinct types of interference problems exist. One type results from interference between satellite programs and the second type is interference to/from user terminals and other terrestrial systems. The dense population of satellites in geostationary orbits prompted concerns of interference between users and satellites, for example, user illumination of unintended satellites create interference at the unintended satellite and likewise, user reception, from not only the desired satellite but also unintended satellites, creates interference. Accordingly, constraints were placed on the sidelobe envelope for user antennas, and are commonly followed by antenna vendors. A typical envelope requirement42 is G = 32 – 25 log q, dBi, 1° < q < 48° = –10 dBi, 48° < q < 180° for D/l > 100



CHAPTER FORTY-FOUR

and G = 52 –10 log (D/l) – 25 logq, dBi, (100D/l)° < q < 48° = 10 –10 log (D/l), dBi, 48° < q < 180° for D/l < 100 where D/l is the antenna aperture size in wavelengths and q is measured from the antenna’s boresight axis. The spectral crowding at the lower microwave frequencies results in potential interference between user satellite antennas and other terrestrial antennas. While frequency regulations are imposed to isolate users, the proximity of terrestrial systems to users is much closer than satellites. While emissions from terrestrial users are isolated in frequency, the higher power levels from nearby terrestrial systems can degrade user systems by saturating the receiver at out-of-band frequencies. Such interference is unintentional. Military systems also have concerns of intentional interference or jamming, and have developed effective means for negating such interference. User systems can be protected from interference by a variety of techniques. In many cases, site alternatives can be effective in avoiding interference; and spectrum analyzer instrumentation43 is commonly used to determine potential interference sources at a given site. Another operational alternative that can be effective is determining the feasibility of restrictions on the operational use of the interference source (sector blanking for nearby radars, for example), not transmitting in a limited angular sector, is sometimes effective in controlling interference from radar systems. In addition to such operational alternatives, four basic techniques may be applied to user systems to negate interference. The gain and filtering distribution of the receiver should be examined to filter interference so that the receiver remains in its linear operating range is a basic technique that should be addressed in the terminal design. Spread spectrum modulation and data interleaving are commonly used to dilute and distribute the effects of interference. In addition to the receiver architecture and the modulation alternatives, two antenna techniques are also available to negate interference. One technique is the passive control of antenna sidelobes.44–46 In many cases, the interference arrives through the antenna sidelobes. Because the radiation mechanisms for the wide angle sidelobes do not contribute to the aperture distribution, these sidelobe control techniques can be employed without significantly impacting the antenna’s main beam gain performance. A second technique is the adaptive cancellation of interference. For user antennas, adaptive sidelobe cancellation47–48 is an appropriate design choice. In this design, auxiliary antenna elements with gain levels that exceed the sidelobe gain are used to sample the interference power. Correlation of the auxiliary antenna elements with the main antenna response is used as a means to dynamically determine the adaptive weight settings. Since the main antenna output is made up of both desired signal and interference components, whereas the auxiliary antenna elements have predominantly interference. The control circuitry for the adaptive weight values is based on minimizing the correlation of the main antenna and the auxiliary antenna elements. The gain of the auxiliary antenna elements and the circuitry is purposely constrained to avoid canceling the main beam of the main antenna containing the desired signal power. One problem with such designs is that the frequency response of the main antenna does not match the response of the auxiliary antenna elements nor are their phase centers coincident. Effective interference over a bandwidth requires equalization of the responses of the main and auxiliary antennas. The differences in these antenna responses have been examined49 and generally, an adaptive transversal equalizer is used to equalize these antenna responses. The number of taps and the required tap spacings for the adaptive transversal filter design have been derived for a specified interference cancellation performance.



SATELLITE ANTENNAS

44.4 SATELLITE ANTENNA TESTING Spacecraft antennas require rigorous RF and environmental testing because of the reliability needed to perform over their long operational lifetimes. While antenna measurement techniques are well established,50 spacecraft antennas have some specialized test requirements that are not commonly used in other applications. User segment antennas also pose antenna test challenges, particularly testing in a production basis as services are extended to a personal level. Larger ground terminals also present test challenges, particularly from excessive far field requirements. Radio source techniques are commonly used in such situations. Spacecraft Antenna Test Phases There are three distinct phases in testing spacecraft antennas, as illustrated in Figure 44-10. The first phase is developmental testing, where the compliance of candidate design(s) with system requirements are established; requirements include both RF performance and testing for compatibility with the launch and on-orbit environments. The second phase is qualification testing, which verifies that the flight hardware conforms to the RF performance established during development testing and that the flight hardware can withstand launch and on-orbit environments. The third phase is on-orbit testing that has the objectives of establishing compliance with system specifications, providing baseline data for subsequent trending, and complementing satellite telemetry to diagnose on-orbit performance shortcomings that may arise during the satellite’s lifetime. Each test phase differs in both scope and requirements. Additionally, the test parameters in each of these phases evolve from component level parameters in the development phase to system level parameters used in the on-orbit phase. Detailed testing can be conducted using general purpose instrumentation and facilities in the development phase, while specialized ground terminals that are generally part of the program’s mission control terminal are required for on-orbit testing and the satellite in its orbital location rather than in a test facility. Development Testing Development testing can generally be accomplished with general purpose test facilities and instrumentation, and it uses engineering model hardware. Much of the RF testing is Development (Component Level)

Qualification (Subsystem Level)

RF Parameters Antenna Gain Pattern Polarization Impedance Power Handling Environmental Piece Parts Facilities Far Field Range Compact Range Near Field Sampling GP Instrumentation

RF Parameters G/T ERP Polarization Environmental EMI/EMC Thermal Vacuum Acoustic Vibration Facilities Compact Range Near Field Sampling Payload Test Set

On-Orbit (Integrated Satellite) Initial On-Orbit Key Performance Parameters Operational Monitoring Data Trending Diagnostic As Required Facility Mission Test Terminal

FIGURE 44-10 Satellite test phases



CHAPTER FORTY-FOUR

conducted at the component level and detailed measurements on the engineering model are performed. Typically, the projected performance of the antenna is evaluated using analytic codes; agreement between these analytic projections and engineering model measurements provides confidence in the results. In some cases, more than one candidate design exists at the initiation of the program, and design tradeoffs are conducted during the development phase to select a final design. In other cases, requirements for other issues such as electromagnetic interference/electromagnetic compatibility (EMI/EMC) result in test requirements at out-of-band frequencies or in specialized models needed to emulate the spacecraft structure and/or payload antennas. Examination of the candidate designs for identifying potential workmanship issues is a prudent practice for assisting in flight hardware testing. Finally, present testing is often conducted using software test procedures and experience with these procedures provides benefits for subsequent qualification testing. At the initiation of the program, a requirements verification matrix is generally constructed flowing down the specific antenna requirements from the system specification and identifying how these requirements are to be verified at each phase of the testing. Attention to the test requirements identified in this matrix is necessary early in the program’s initiation in order to identify instrumentation and facility needs, as well as necessary test points within the antenna hardware. In some cases, development is required to conduct the qualification testing. Additionally, the requirements verification matrix helps to identify the scope of the test procedures and planning. Such steps are necessary in determining program needs and schedule requirements. Qualification Testing Flight hardware must be space qualified to demonstrate its capability for properly performing over its orbital lifetime. Qualification testing is performed on flight hardware that must be properly protected. RF testing is therefore performed in protected test facilities to avoid damage to flight hardware. Compact ranges and near field sampling facilities are in common use for this testing. Qualification testing also requires demonstrating that the flight hardware is capable of withstanding launch and on-orbit environments. These environments typically require acoustic, thermal and vacuum, and vibration testing. Such testing is often performed at a more stringent level, on engineering development hardware that emulates the flight hardware in order to determine design margins. The flight hardware is then tested at a somewhat lower qualification level following environmental values anticipated during launch and on-orbit operations. One of the important features of the qualification level testing is the identification of workmanship errors (such as cable integrity) and specific payload assembly errors that would limit on-orbit reliability. Such testing benefits from identification of potential workmanship errors in the development test phase. Qualification testing is generally conducted at the component, subsystem, and integrated spacecraft levels. The RF testing is distributed over these levels with attention at the component and subsystem levels using common antenna measurement parameters and at the integrated spacecraft level, transitioning to system level parameters such as ERP, G/ T, and BER values for the overall transponder. Redundant electronics paths are also evaluated in this testing, but this involves not just the performance of these redundant paths but the command paths needed to select the redundant components as well. Testing of this type particularly within the test environments precludes the normal RF test facilities. Testing in these environments often involves “hat couplers” that are comprised of RF sealed enclosures containing RF probes to allow test signal injection and reception. Such hat couplers permit testing within environmental facilities such as thermal vacuum chambers that preclude conventional test facilities. The hat coupler designs are specific to the antenna technology used, and require development and calibration prior to their use in



44-23

qualification testing. Likewise, testing to evaluate both the antennas and payload electronics is typically desired after shipment to launch sites and spacecraft assembly areas. The development of suitably portable RF test facilities would benefit testing in these areas, and as the spacecraft antennas become more complex and integrated with RF electronics, test requirements for the assembled payload will evolve from a “light bulb” evaluation of the heath to more insightful evaluations of the antenna systems functionality. The evaluation of EMI/EMC compliance is a spacecraft-wide issue. At the initiation of the spacecraft design, the frequency plans should be examined to avoid conflicts. All frequency components and their harmonics need to be examined and the coupling between these components and other potentially susceptible components throughout the entire spacecraft need to be examined. Estimates of these potential coupling paths need to be examined and those paths with limited susceptibility margins need to be identified and verified by test. It is not unusual for the number of potential coupling paths to exceed 10,000, and after examination the number of paths with limited margin will be substantially less. The spacecraft antennas are generally well isolated from the majority of the payload electronics and their coupling can be estimated and/or augmented by measurement particularly at outof-band frequencies that are not normally characterized. Some of the coupling paths can be compromised by workmanship errors and call for examination of the flight hardware. Some EMI/EMC testing is performed at the component and subsystem levels. Generally, the assembled flight hardware is tested in an EMI/EMC facility to assure compliance. The natural background levels within these facilities needs to be established first, and these levels compared with the required sensitivities. Emissions from the test equipment and any anticipated leakage of the cabling that supports the testing bears scrutiny prior to the spacecraft examination. The EMI/EMC testing should concentrate on those paths previously identified as well as a more general survey of the emissions. Measurements of potential emissions are often sniffed using waveguide probes at microwave frequencies to detect leakage components that could degrade system operation. Satellite qualification testing includes some specialized evaluations that are not normally a concern in other antenna applications. One concern for spacecraft transmitting antennas is multipaction51 that results in vacuum conditions when sufficient electrical field strength exists to strip electrons from metallic surfaces. A resonant condition produces an electron avalanche damaging the surfaces. Multipaction susceptibility depends on the product of gap size between surfaces and the RF frequency the spacing between surfaces, vacuum conditions, and the secondary electron emission properties of the materials as well as the RF power levels. Multipaction generally occurs at the smallest separation between surfaces in components. Testing components with potential multipaction concerns is generally performed under vacuum conditions, and since a 6 dB margin in power handling capability is commonly used, the transmitter power for such testing is accordingly higher than that used operationally. Space segment antennas serve a variety of users and thus typically have modulations with multiple carrier components. In addition to transmitter linearity requirements, concerns also exist regarding the linearity of the antennas and RF components, particularly in the case where the same antenna aperture is used to both receive and transmit; the receive and transmit bandwidths are isolated by the filtering in a diplexer. Mechanical joints within the RF and antenna components can exhibit nonlinear behavior creating intermodulation products that interfere with signal reception. This phenomenon is referred to as (passive intermodulation) PIM.52–53 Electroforming techniques can minimize the required number of joints in RF components and reduce potential PIM levels. The generation of PIM products is often determined by injecting two CW tones into the components. The frequencies are chosen to produce PIM products within the received passband. An examination of various orders of potential PIM products is typically performed to understand what problems can arise and guide the evaluation of potential PIM generation.



CHAPTER FORTY-FOUR

Another concern for satellite antennas is the potential of electrostatic discharge (ESD), which can damage or interfere with RF electronics. Satellite antennas are generally protected from the on-orbit thermal extremes by thermal covers. Such covers necessarily must have low insertion loss and are typically dielectric materials. On-orbit, such covers can accumulate sufficient charge to produce a discharge54–55 and while the spectra of the discharge has strong low frequency components, sufficient energy can couple into RF electronics at levels that can cause damage or interference. In other cases, the same thermal control materials are used to protect other spacecraft components such as solar arrays, and while the discharge is not associated with the antenna itself, nonetheless, spectral coupling from the discharge into the antenna must be considered to evaluate potential interference. The potential for ESD is particularly high during geomagnetic storms. Thermal covers, therefore, require sufficient conductivity in either their composition or through special ESD paints and ground paths to control the charge buildup. Generally, ESD characteristics are measured on material samples that are charged with electron beams to determine the spectral characteristics of the discharge. Such tests are often performed on samples of the materials and in some cases, on the assembled antenna to verify the integrity of the grounding paths. A common problem involves testing antennas that have broad coverage requirements such as the hemispheric coverage required by TT&C antennas. Because these designs often use lower microwave frequencies and must be compact, isolating such antennas from the surrounding spacecraft becomes a challenge. The antenna performance on the spacecraft differs from its free space performance and the interaction between the antenna and spacecraft results in pattern ripples in the coverage area that can lead to noncompliance. Antenna backlobe reduction helps to reduce the interaction. Testing on the spacecraft has its challenges not only from the physical size of the spacecraft but also the increased far field distance requirements compared to antenna itself. Development attention to address analysis and measurement techniques for such situations is recommended. The environmental testing on spacecraft components includes acoustic, vibration, and thermal vacuum evaluations. Such testing is performed on the antennas by themselves and repeated when the antennas are integrated with the payload. In cases where the design margins relative to the environmental tests are required, testing performed on engineering model hardware is often performed at more stringent protoqualification levels. The test levels depend on the launch vehicle characteristics and structural details of the payload and the thermal characteristics projected in orbit. These tests are useful in ascertaining workmanship defects such as cable shortfalls. Testing within the environmental facility, particularly for the assembled payload, often require an end-to-end evaluation, such as BER measurements. Since conventional antenna test facilities cannot be configured within the environmental facility, a means of injecting test signals and measuring their output is required. In some cases, test couplers are integrated into the antenna hardware for such purposes and in other cases, hat coupler are used to inject or receive test signals from the antenna. Both injection methods require attention to calibration and measurement repeatability to obtain meaningful results particularly for evaluating the noise characterization for uplink receiving antennas. On-Orbit Testing On-orbit testing verifies compliance with the key system level parameters for the spacecraft initially, and provides performance monitoring capabilities of the satellite over its lifetime. These measurements are at the system level, such as ERP, G/ T, BER, and others. Such testing is performed with a specialized ground terminal. The requirements for these terminals include rigorous calibration and diagnostic capabilities to ensure accurate results, and other capabilities that are specific to the programs requirements. For example, systems that use



44-25

polarization reuse place stringent polarization purity requirements on the ground terminal to accurately measure the satellite’s polarization isolation. Likewise, some satellite systems have requirements for antenna sidelobe control (such as isolation characteristics) between multiple beams, and the sensitivity required to verify such sidelobe requirements dictates the ground terminal aperture size to perform accurate measurements. Other monitoring requirements include the satellite user loading characteristics versus the satellite transmitter operating point to control the level of intermodulation products. On-orbit testing requires a dedicated ground terminal to monitor the on-orbit satellite operation. Such terminals are generally a part of the program’s mission control facilities. These test terminals require particular attention to calibration to obtain accurate measurements. Additionally, these terminals require effective diagnostic capabilities to assure indicated shortfalls result from the spacecraft rather than the test terminal. Generally, the performance requirements for the test terminal exceed those of the user terminals. For example, measurement of the isolation between antenna beams requires a significantly larger test terminal aperture than the data reception requirements for a user terminal. User Antenna Testing User antennas for satellite systems also have requirements that differ from other applications. One problem that will become more commonplace is the ability to test at production levels. This problem results as satellite services are extended more broadly at the personal level and the number of user terminals increases. Testing each individual antenna is often both cost and schedule prohibitive. One approach for testing at a production level examines those areas where workmanship shortfalls can lead to noncompliance. One example of such shortfalls is user terminals where polarization reuse is used and the polarization purity must be maintained to isolate signals on orthogonal polarizations. An example of workmanship shortfall leading to noncompliance of these isolation requirements is the failure to maintain the required mechanical tolerance in the internal antenna feed dimensions that can increase cross-polarization levels. In this case, simple mechanical tolerance testing and random RF testing can provide assurance of compliance with polarization purity requirements without testing each individual feed. Compliance testing for user antennas is generally conducted using general purpose instrumentation and test facilities. The normal antenna characterizations of gain, polarization, bandwidth, and impedance parameters are conducted using standard measurement practices. Other requirements, such as sidelobe envelope compliance, must address the capabilities needed for the sensitivity for accurately measuring lower power levels from the sidelobes and attention to the facility multipath levels that can degrade measurement accuracies. In such cases, attention to leakage levels in the RF test instrumentation and microwave components are needed to assure the apparent sidelobe levels are from the antenna rather than leakage components. The correlation of these sidelobe responses with analytic projections of the design performance provides increased measurement credibility. Radio Source Measurements A measurement technique using radio sources has proved useful for characterizing the G/ T of user antennas. Originally, the technique was applied to large ground terminal antennas,56–57 which have a physical size that cannot be accommodated in general purpose test facilities and which have far field distances that are excessive. Because of available low noise receivers in system use, the technique can be applied to much smaller diameter antennas58 and, as RF electronics are integrated into the antenna design, it can provide a means of testing designs that have no convenient antenna terminals.



CHAPTER FORTY-FOUR

The measurement technique59 consists of measuring the noise power from a radio source and measuring the cold sky background noise at the same elevation angle; the ratio of these noise power measurements is a Y factor. The noise flux densities of radio sources are very well known. For large antennas, radio sources such as Cassiopeia A are used. If the antenna does not have the sensitivity to measure the noise from these sources, the moon is commonly used. The strongest radio source is the sun, and measurements can be made on much smaller antennas using this source. The sun has some limitations. Solar flare activity results in variations in the noise flux densities, but measurements of the solar flux are made daily by a global network, so that variations in the solar flux are determined. The solar disk subtends about a 1/2° angular width so that it does not appear to be a point source for antennas whose beamwidths approach the 1/2° width. Techniques to correct for the angular width have been devised. However, because of solar flares, the solar flux density is not uniformly distributed over the solar disk, and experience has shown that particularly during active solar periods, measurements of antennas with beamwidths smaller than 1/2° experience experimental results that are not always repeatable. By contrast, the correction for the angular width of the moon is not limited because the lunar flux results from solar reflection and does not have the variability of flare activity. The G/ T from these measurements is derived from G/ T = 8p k (Y – 1)/Sl2 where k is Boltzmann’s constant, Y is the noise power ratio of antenna pointed at the source and the antenna pointed at a cold sky background, S is the radio source flux density, and l is the wavelength. In practice, the G/ T is referenced to a specified elevation angle because of the antenna noise temperature dependence on elevation angle. Since the radio source is generally at a different elevation angle, the G/ T must be compensated for the differences in the cold sky noise measurement where the source is located, as well as the cold sky noise measurement at the specified elevation angle. The radio source flux density does not include the propagation loss of the atmospheric path and compensation for this path loss must be made. At microwave frequencies, the path loss can be estimated, but at the higher EHF frequencies, the path loss depends on weather conditions. A separate measurement of the antenna temperature versus elevation angle is made as a means of determining weather effects at the time of the G/ T measurement. The antenna temperature can be measured from the ratio of the noise powers when the receiver is terminated by an ambient load and when the receiver is terminated by the antenna pointed at the desired elevation angle. This ratio equals the sum of the ambient and receiver noise temperatures divided by the sum of the antenna and receiver noise temperatures. The antenna temperature is determined from this ratio and the receiver noise temperature. The G/ T measurements need to assure the antenna is properly pointed at the noise source and that the measurements are not degraded by interference. In many cases, the antenna gain rather than the G/ T is required. The total system temperature is the sum of the antenna noise temperature and the receiver noise temperature. The antenna temperature can be measured and the receiver noise temperature is typically measured using the standard Y factor technique with hot and cold loads. The antenna gain is determined from the product of the measured G/ T and the system noise temperature comprised of the sum of the antenna temperature at the elevation angle where the radio source was located for the G/ T measurement and the receiver noise temperature. The sensitivity of the antenna temperature to elevation angle is of interest not only to obtain the correct value for determining the antenna gain but also in understanding how the user antenna system will perform at different elevation angles. The antenna gain can then be determined by multiplying the antenna’s G/ T by the system noise temperature. In addition, a common reference plane must be used for both the antenna’s G/ T and system noise temperature values. In some cases, the antenna gain within the transmit band is desired. If an LNA



44-27

operating in the transmit band is substituted for the transmitter, a G/ T measurement using a radio source can be performed. Determining the system noise temperature T for this case allows the determination of the transmit antenna gain.

Diplexer User antennas commonly require the same antenna for both receive and transmit operation. A diplexer is required to separate the transmitting and receiving frequencies. While a variety of techniques may be used for diplexing, the requirements for diplexing are common to all designs. A diplexer provides the filtering necessary to isolate the receiver and transmitter and a means to connect the receiver and transmitter to the antenna. Diplexer loss is to be minimized to avoid reduced received and transmitted signal levels. The requirements of filtering are derived in the following manner. The transmit filtering must pass the transmitted spectrum with low loss and with sufficient amplitude and phase flatness to avoid transmit signal distortion. The transmit filter must sufficiently suppress the out-of-band emissions from the transmitter that fall within the receive bandwidth sufficiently to avoid degrading receiver performance. This filtering requirement requires knowledge of the out-of-band spectral characteristics of the transmitter and generally, the required filter rejection in the receive bandwidth reduces the out-of-band transmitted spectrum to a level 10 dB below the receiver noise spectra density. This requirement also forms a basis to test the rejection by observing the receiver noise level with the transmitter off and the increase in the receiver noise level with the transmitter on. Testing for high transmit power levels also involves examining the power handling capabilities. Two requirements exist for high transmit power levels and are reflected in the testing. The first requirement is to ensure breakdown does not occur, a peak power limitation. The second requirement concerns average power limitations and involves demonstrating the increased temperature resulting from ohmic loss stays beneath a safe operating temperature limit. In such cases, the transmit diplexing filtering might be distributed with some of the filtering being located on the transmitter and other filtering located near the antenna’s diplexer combiner to distribute the thermal load. The receive filter must pass the receive spectrum with low loss, as well as sufficient amplitude and phase flatness to avoid receive signal distortion. The filtering must also suppress the transmitted spectrum sufficiently to avoid compressing the receiver at out-of-band frequencies. This filtering requirement involves determining the spectral coupling of the transmitter into the receiver port, understanding the gain and filtering distribution of the receiver, and determining the power levels needed to saturate the receiver at out-of-band frequencies. Again, the filtering rejection is sized with a margin, commonly 6 dB, to avoid receiver compression throughout the receiver architecture. Testing involves measuring the receiver’s amplifiers output levels at the out-of-band frequencies where saturation could potentially occur. In practice, a significant amount of filter rejection can be required. Attention should be given to leakage around filter sections to assure the overall filtering rejection is not compromised. Leakage components can be detected by using waveguide probes to assure component joints and tuning screws do not have excessive leakage levels. Electroforming techniques and distributing the filters at different locations can be helpful. The overall filter rejection may be sufficiently high to limit verification within the dynamic range of standard instrumentation such as network analyzers. The network analyzers provide the means to measure in-band insertion loss and amplitude and phase flatness as necessary but determining the rejection can require further sensitivity. Using additional amplification may provide the added sensitivity and in the case of distributed filtering, the individual components can be measured. In any event, measurements of the assembled system are necessary to assure the diplexer performance complies with requirements.



CHAPTER FORTY-FOUR

Radiation Hazard Large ground terminal antennas often require significant power levels to meet link requirements for high data rate transfer. When high transmitting power levels, concerns exist that the power densities near the antenna can potentially harm personnel. Generally, the safe levels for RF exposure50 are 2 mW/cm2 and the debate on the safe level is ongoing. When such concerns exist, measurements of the antenna fields are performed. These measurements are typically within the antenna’s near field and are performed using a probe antenna, as well as spectrum analyzer instrumentation. Because the power densities of concern have extremely high levels, measurements are generally performed using a signal generator in place of the actual transmitter to avoid concerns of potential harm to personnel performing the measurements and generally adequate sensitivity exists when the signal generator is used. Such measurements are guided by first using available analysis codes to examine the near field characteristics to identify regions where the near field power densities could potentially exceed the safe levels of exposure. One region is the near field of the aperture distribution and since in operation, the antenna is pointed at the satellite, this region is not readily accessible to personnel on the ground. Other regions, for example, pillover and scattering by struts supporting the feed and/or subreflector, are potential near field locations where the safe levels can be exceeded. Using the near field analyses provides insight of where measurements should be performed and an indication of what their levels should be. Other regions need to be examined if their effects are not included in the analyses. For example, leakage through the reflector panel gaps in large reflector antennas and openings in the reflector surfaces for antennas feeds need to be separately probed.

REFERENCES 1. D. H. Martin, Communication Satellites-Fifth Edition (El Segundo. CA: The Aerospace Press and AIAA, 2007). 2. E. G. Njoku and E. K. Smith, “Microwave Antenna Temperature of the Earth From Geostationary Orbit, Radio Science, vol. 20 (May–June 1985): 591–599. 3. C. T. Brumbaugh, A. W. Love, G. M. Randall, D. K. Waineo, and S. H. Wong, “Shaped Beam Antenna for the Global Positioning Satellites,” 1976 IEEE AP-S Symposium Digest (October 1976): 117–120. 4. T. Katagi and Y. Takeichi, “Shaped Beam Horn Reflector Antenna,” IEEE Trans Antennas and Propagation, vol. AP-23 (November 1975): 757–763. 5. A. R. Cherrette, S. W. Lee, and R. J. Acosta. “A Method for Producing a Shaped Contour Radiation Pattern Using a Single Shaped Reflector and a Single Feed,” IEEE Trans Antennas and Propagation, vol. 37 (June 1989): 698–706. 6. O. M. Bucci, G. D’Elia, G. Mazzarello, and G. Panariello, “Antenna Pattern Synthesis: A New General Approach,” Proc IEEE, vol 82 (March 1994): 358–371. 7. R. B. Dybdal, “Polarization Efficiency Statistics,” 1999 IEEE MILCOM Symposium Digest (November 1999): 203–207. 8. R. A. Monzingo and T. W. Miller, Introduction to Adaptive Arrays (New York: Wiley, 1980). 9. K. M. Soo Hoo and R. B. Dybdal, “Resolution Performance of an Adaptive Multiple Beam Antenna,” IEEE MILCOM ’89 Symposium Digest (October 1989). 10. J. T. Mayhan, ”Area Coverage Adaptive Nulling from Geosynchronous Satellites: Phased Arrays Versus Multiple Beam Antennas, IEEE Trans Antennas and Propagation, vol. AP-27 (March 1986): 410–419. 11. R. B. Dybdal, D. J. Hinshilwood, and K. M. SooHoo, “Development Considerations in the Design and Simulation of Adaptive MBAs for Satellite Communications,” 1993 IEEE MILCOM Symposium Digest (October 1993).



44-29

12. R. B. Dybdal and H. J. Wintroub, “An Antenna Based High Data Rate Concept,” IEEE MILCOM ’95 Symposium Digest (November 1995). 13. R. B. Dybdal and S. J. Curry, “Adaptive Beam Pointing,” 1996 IEEE MILCOM Symposium Digest (October 1996): 882–886; see also, R. B. Dybdal and S. J. Curry, Adaptive Receiving Antenna for Beam Repositioning, (April, 14, 1998): U.S. Patent 5,739,788. 14. K. Ueno, T. Itanami, H. Kumazawa, and I. Ohtomo, “Design and Cahracteristics of a Multiband Communication Satellite Atnenna System,” IEEE Trans Aerospace and Electronic Systems, vol. 31, 600–606 15. R. J. Mailloux, Phased Array Antenna Handbook, Chap. 8 (Norwood, Massachusetts: Artech, 2005). 16. P. Ingerson and C. A. Chen, “The Use of Non-Focusing Aperture for Multibeam Antenna,” 1983 IEEE AP-S Symposium Digest (May 1983): 330–333. 17. K. S. Rao, G. A. Morin, M. Q. Tang, S. Richard, and K. K. Chan, “Development of a 45 GHz Multiple Beam Antenna for Military Satellite Communications,” IEEE Trans Antennas and Propagation, vol. 43 (October 1995): 1036–1047. 18. A. W. Love (ed.), Reflector Antennas (New York: IEEE Press, 1978). 19. R. B. Dybdal, S. J. Curry, and M. A. King, “An Uplink Multiple Beam Concept for Theater Coverage,” 2002 IEEE MILCOM Symposium Digest (October 2002). 20. R. B. Dybdal and S. J. Curry, “An Uplink Antenna for Electronic Beam Steering and Interference Reduction,” 2002 IEEE AP-S Symposium Digest, San Antonio TX, vol. 1 (June 2002): 590–593. 21. R. B. Dybdal, “Adaptive Control of Multiple Beam Satellite Transponders,” 1997 IEEE MILCOM Symposium Digest (November 1997): 252–255; see also, R. B. Dybdal, Adaptive Control of Multiple Beam Communication Transponders, (April 25, 2000): U. S. Patent 6,055,431. 22. F. J. Dietrich, P Metzen, and P Monte, “The Globalstar Cellular Satellite System,” IEEE Trans Antennas and Propagation, vol 46 (June 1998): 933–942. 23. J. J. Schuss, J. Upton, B Myers, T. Sikina, A. Rohwer, P. Makridakas, R. Francois, L. Wardle, and R. Smith, “The IRIDIUM Main Mission Antenna Concept,” IEEE Trans Antennas and Propagation, vol 47 (March 1999): 416–424. 24. T. J. Peters, “A Conjugate Gradient-based Algorithm to Minimize the Sidelobe Level of Planar Arrays with Element Failures,” IEEE Trans Antennas and Propagation, vol 39 (October 1991): 1497–1504. 25. R. Miura, T. Tanaka, I. Chiba, A. Horie, and Y. Karasawa, “Beamforming Experiment with a DBF Multibeam Antenna in a Mobile Satellite Environment,” IEEE Trans Antennas and Propagation, vol. 45 (April 1997): 707–714. 26. See Chap. 42. 27. H. J. Liebe, “Modeling Attenuation and Phase of Radio Waves in Air at Frequencies Below 1000 GHz,” Radio Science, vol. 16 (November–December 1981): 1183–1199. 28. H. J. Liebe, “An Updated Model for Millimeter Wave Propagation in Moist Air,” Radio Science, vol. 20 (September–October 1985). 29. R. B. Dybdal and F. I. Shimabukuro, “Electronic Vulnerability of 60 GHz Crosslinks,” 1984 IEEE MILCOM Symposium Digest, Paper 27.6 (October 1984). 30. D. E. Ping, J. T. Shaffer, L. U. Brown, and R. B. Dybdal, “A Broadband Rolled Edged Cavity Antenna,” 2004 IEEE AP-S Symposium Digest (June 2004). 31. R. Tomiyasu, “Remote Sensing of the Earth by Microwaves,” Proc IEEE, vol. 62 (January 1974): 86–92. 32. …, Special Issue of SIR-C/X SAR, IEEE Trans on Geoscience and Remote Sensing, vol. 33 (July 1995). 33. …, Special Issue on the Windsat Spaceborne Polerimetric Radiometer, IEEE Trans Geoscience and Remote Sensing, vol. 44 (March 2006). 34. E. K. Smith, “Centimeter and Millimeter Wave Attenuation and Brightness Temperature Due to Atmospheric Oxygen and Water Vapor,” Radio Science, vol. 17 (November–December 1982): 1455–1464.



CHAPTER FORTY-FOUR

35. R. K. Crane, Electromagnetic Wave Propagation Through Rain (New York: Wiley, 1996). 36. L. J. Ippolito, “Propagation Effects Handbook for Satellite Systems Design,” ITT Industries Doc. AF01-0006 (September 2000). 37. R. L. Olsen, D. V. Rogers, and D. B. Hodge, “The aR b Relation in the Calculation of Rain Attenuation,” IEEE Trans Antennas and Propagation, vol. AP-26 (March 1978): 318–329. 38. F. J. Dietrich and D. B. West, “An Experimental Radome Panel Evaluation,” IEEE Trans Antennas and Propagation, vol. AP-36 (November 1988): 1566–1570. 39. C. E. Hendrix, J. E. McNalley, and R. A. Monzingo, “Depolarization and Attenuation Effects of Radomes at 20 GHz,” IEEE Trans Antennas and Propagation, vol. AP-37 (March 1989): 320–328. 40. R. E. Collin and F. J. Zucker, Antenna Theory, Chap. 4 (New York: McGraw Hill, 1969). 41. K. M. Lambert and R. C. Rudduck, “Calculation and Verification of Antenna Temperature for Earth-Based Reflector Antennas,” Radio Science, vol. 27 (January–February 1992): 23–30. 42. CCIR Recommendation 465–563. 43. R. B. Dybdal, G. M. Shaw, and T. T. Mori, “A RFI Measurement System for Field Sites,” 1995 AMTA Symposium Digest (November 1995). 44. R. B. Dybdal, “Millimeter Wave Antenna Technology,” IEEE Trans Selected Areas on Comm, vol. SAC-1 (September 1983): 633–644. 45. H. P. Coleman, R. M. Brown, and B. D. Wright, “Parabolic Reflector Offset Fed with a Corrugated Horn,” IEEE Trans Antennas and Propagation, vol. AP-23 (November 1975): 817–819. 46. A. B. Crawford, D. C. Hogg, and L. E. Hunt, “A Horn Reflector Antenna for Space Communications,” BSTJ, vol. 40, 1095–1116. 47. P. W. Howells, Intermediate Frequency Sidelobe Canceller, (1965): U. S. Patent 3,202,990. 48. K. M. SooHoo and W. Masenten, “Adaptive Sidelobe Canceller Designs for Large Earth Terminals,” 1984 IEEE MILCOM Symposium Digest, paper 40.5, (Classified). 49. R. A. Dell Imagine and K. M. SooHoo, “Adaptive Sidelobe Canceller Designs for Large Earth Terminals,” Proc 1980 Adaptive Antenna Symposium, RADC Doc. TR-80-378. 50. …, IEEE Standard Test Procedures for Antennas (New York: Wiley-Interscience, 1979). 51. A. D. Woode and J. Petit, “Design Data for the Control of Multipactor Discharge in Spacecraft Microwave and RF Systems, Microwave Journal (January 1992): 142–155. 52. J. W. Boyhan, H. F. Lenzing, and C. Koduru, “Satellite Passive Intermodulation: Systems Considerations,” IEEE Trans Aerospace and Electronic Systems, vol. 32 (July 1996): 1058–1063. 53. J. W. Boyhan, “Ratio of Gaussian PIM to Two Carrier PIM,” IEEE Trans Aerospace and Electronic Systems, vol. 36 (October 2000): 1336–1342. 54. C. Bowman, A Bogorad, G. Brucker, S. Seehra, and T. Lloyd, “ITO-Coated RF Transparent Materials for Antenna Sunshields-Space Environment Effects,” IEEE Trans Nuclear Science, vol. 37 (December 1990): 2134–2137. 55. H. C. Koons and T. S. Chin, “Broadband RF Spectrum for Electrostatic Discharges on Spacecraft,” Aerospace Corp Tech Rept, TR-93(3940)-6, (May 1993). 56. J. W. M. Baars, “The Measurement of Large Antennas with Cosmic Radio Sources,” IEEE Trans Antennas and Propagation, vol. AP-21, 461–474. 57. D. F. Wait, “Precision Measurement of Antenna System Noise Using Radio Stars,” IEEE Trans Instrumentation and Measurement, vol. IM-32 (March 1983): 110–116. 58. R. B. Dybdal, “G/ T Measurement of Small Antennas,” 1997 AMTA Symposium Digest (November 1997): 37–42. 59. R. B. Dybdal, “On G/ T Radio Source Measurements,” 2000 AMTA Symposium Digest (October 2000): 187–191.



Chapter 45

Earth Station Antennas William A. Imbriale Jet Propulsion Laboratory California Institute of Technology


45-2

45.2 SINGLE-BEAM EARTH STATION ANTENNAS . . . . . . . . . . . . . . . .

45-4

45.3 MULTIPLE-BEAM EARTH STATION ANTENNAS. . . . . . . . . . . . . . 45-13 45.4 ANGLE-TRACKING TECHNIQUES. . . . . . . . . . . . . . . . . . . . . . . . . . 45-16 45.5 POLARIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45-17 45.6 MEASUREMENT TECHNIQUES . . . . . . . . . . . . . . . . . . . . . . . . . . . 45-18


Earth Station Antennas 45-2

CHAPTER FORTY-FIVE

45.1 INTRODUCTION Almost all Earth station and ground station antennas are used for either telecommunications or science, and frequently the same antenna (as, for example, the 70-m antenna of the National Aeronautics and Space Administration [NASA] Deep Space Network [DSN]1) is used for both purposes. The terms Earth station antenna and ground station antenna seem to be used interchangeably. However, most often, antennas used for supporting Earth-orbiting satellites are called Earth station antennas and those used primarily for science or deepspace satellites are termed ground station antennas. This chapter primarily deals with telecommunications, and Chapter 49 covers the science applications. However, much of the material included in this chapter (e.g., analysis, design, and measurement techniques) applies equally well to both applications and will not be repeated in Chapter 49. This chapter also uses some material from previous editions of the book. Reflector antennas have existed since the days of Heinrich Hertz (1857–1894) and are still one of the best solutions to requirements for cost-effective, high-gain, high-performance antenna systems. Consequently, virtually all ground station antennas are reflector antennas of one type or another. Requirements for ground station antennas can be grouped into several major categories: ●

●

●

●

●

●

Electrical or radio-frequency (RF) requirements Control-system requirements Structural requirements Pointing- and tracking-accuracy requirements Environmental requirements Miscellaneous, such as those concerning radiation hazards, primary-power distribution for de-icing, etc.

Only the electrical or RF requirements will be dealt with in this chapter. The primary electrical antenna specifications are gain, noise temperature, voltagestanding-wave ratio (VSWR), power rating, receive–transmit group delay, radiation pattern, polarization, axial ratio, isolation, and G/T (antenna gain divided by system noise temperature). All parameters except the radiation pattern are determined by system requirements. For commercial Earth station antennas, the radiation pattern must meet the minimum requirements set by the International Radio Consultative Committee (CCIR) of the International Telecommunications Union (ITU) and/or national regulatory agencies such as the U.S. Federal Communications Commission (FCC). For antennas used for deep-space communications, the primary design requirement is for G/T, and, in general, there are no requirements for a specific radiation pattern. Earth station antennas operating in the field of international satellite communications must have sidelobe performance as specified by INTELSAT standards or by CCIR Recommendation 580-1 (see Figure 45-1). The CCIR standard specifies the pattern envelope in terms of allowing 10 percent of the sidelobes to exceed the reference envelope and also permits the envelope to be adjusted for antennas whose aperture is less than 100 wavelengths (100l). The reference envelope is given by G = [49 − 10 log ( D /L ) − 25 log θ ] dBi = ( 29 − 25 log θ ) dBi

D ≤ 100 λ D > 100 λ


Earth Station Antennas EARTH STATION ANTENNAS

FIGURE 45-1

45-3

Sidelobe envelope as defined by CCIR Recommendation 580-1 and the FCC

This envelope takes into consideration the limitations of small-antenna design and is representative of measured patterns of well-designed dual-reflector antennas. For antennas used for deep-space communications, there is no specific sidelobe requirement. However, since the prime driver for deep-space missions is to receive extremely weak signals over vast interplanetary distances, a key element of telecommunications link performance is the received power signal-to-noise ratio (SNR), which is given by

S /N ≈

PT GT GR 4π PT AT AR = 2 2 4π R 2 N λ R kBTs

(45-1)

where PT = spacecraft transmit power GT = gain AT = effective area of the transmit (spacecraft) antenna GR = gain AR = effective areas of the receive ground antennas N = total noise R = distance to the spacecraft Ts = receive system noise temperature l = wavelength B = bandwidth k = Boltzman’s constant To do its part effectively, the ground antenna system must maximize the ratio of received signal to the receiving system noise power, which is measured by an antenna figure of merit (FM), defined as the ratio of antenna effective area (or equivalent gain) to system noise temperature.



CHAPTER FORTY-FIVE

The receive temperature consists primarily of the antenna feed system and amplifier contributions. For assessing the antenna FM, it is desirable to draw an imaginary reference plane between the receiver system and the antenna system, thus placing all noise contributions in one or the other of these categories. If the receiver contribution (including the feed system losses) is given by TR and the antenna noise contribution by TA, then the FM will be given by FM =

GR TR + TA

(45-2)

Antenna noise temperature properties are very significant contributors to FM, especially for cases of low receiver noise temperature systems. Thus, to maximize the FM for a given antenna size and frequency of operation, it is necessary to both maximize the antenna gain and minimize the total system noise temperature. Since the individual contributions to noise temperature are additive and essentially independent of each other, it is necessary to individually minimize each contribution. If any one of the noise contributions is large, minimizing the others only marginally improves the FM. However, by using cryogenic amplifiers, the receiver noise temperature contribution can be small (as low as 2 to 3 K for the best performing masers), and it then becomes imperative to minimize both the antenna and feed system contributions. For ambient conditions, it should be noted that feed system losses contribute to noise temperature at the rate of 7 K per 0.1-dB loss. Ground station antennas can be grouped into two broad categories: single-beam antennas and multiple-beam antennas. A single-beam antenna is defined as an antenna that generates a single beam that is pointed toward a satellite by means of a positioning system. A multiple-beam Earth station antenna is defined as an antenna that generates multiple beams by employing a common reflector aperture with multiple feeds illuminating that aperture. The axes of the beams are determined by the location of the feeds. The individual beam identified with a feed is pointed toward a satellite by positioning the feed, generally without moving the reflector. An important subclass of singe-beam antennas is the beam-waveguide (BWG) antenna, which is composed of one or more feedhorns with a series of flat and curved mirrors arranged so that power can be propagated from the horn through the mirrors to the main reflector with minimum loss.

45.2 SINGLE-BEAM EARTH STATION ANTENNAS Single-beam antenna types used as Earth stations are paraboloidal reflectors with focalpoint feeds (prime-focus antenna), dual-reflector antennas such as the Cassegrain, Gregorian, and dual-shaped configurations, BWG antennas, offset-fed paraboloidal antennas, and offset-fed multiple-reflector antennas. Each of these antenna types has its own unique characteristics, and the advantages and disadvantages have to be considered when choosing one for a particular application. Axisymmetric Dual-Reflector Antennas The predominant choice of designers of large Earth station antennas has been the axially symmetrical dual-reflector antenna (Cassegrain or Gregorian, classical or shaped). 1. The classical Cassegrain geometry2,3 employs a paraboloidal surface for the main reflector and a hyperboloidal surface for the subreflector (see Figure 45-2). The paraboloidal



FIGURE 45-2

45-5

Geometry of the Cassegrain antenna system

reflector is a point-focus device with a diameter, Dp, and a focal length, fp. The hyperboloidal subreflector has two foci. For proper operation, one of the two foci is the real focal point of the system and is located coincident with the phase center of the feed; the other focus, the virtual focal point, is located coincident with the focal point of the main reflector. The parameters of the Cassegrain system, as shown in Figure 45-2, are related as follows:

φ p = 2 tan −1 ( 0.25D p /Fp )

(45-3)

fs /ds = 0.5(cot φ p + cot φs )

(45-4)

 s / fs = 0.5(1 − {sin[0.5(φ p − φs )] / sin[0.5(φ p + φs )]})

(45-5)

In a typical design, the parameters Fp, Dp, and fs are chosen, and the remaining three parameters are then calculated. Granet4 has given design formulas and procedures for both Cassegrain and Gregorian starting with different design parameters and design considerations such as minimum blockage. The contours of the main reflector and subreflector are given by Main reflector: ym2 = 4 Fp xm

(45-6)

Subreflector: ( ys /b )2 + 1 = ( xs /a + 1)2

(45-7)

where a = ( f / 2e)

b = a e2 − 1

e = sin [0.5(φ p + φs )] / sin[0.5(φ p − φs )] The quantities a, b, and e are half of the transverse axis, half of the conjugate axis, and the eccentricity parameters of the hyperboloidal subreflector, respectively.



CHAPTER FORTY-FIVE

FIGURE 45-3 reflectors

Circularly symmetric dual-shaped

2. An extremely important generalization of the Cassegrain geometry is that it consists of a special-shaped quasi-paraboloidal main reflector and a shaped quasi-hyperboloidal subreflector.5–8 Green9 observed that in dual-reflector systems with high magnification—essentially a large ratio of main-reflector diameter to subreflector diameter—the distribution of energy (as a function of angle) is largely controlled by the subreflector curvature. The path length or phase front is dominated by the main reflector (see Figure 45-3). Kinber5 and Galindo6,7 found a method for simultaneously solving for the main-reflector and subreflector shapes to obtain an exact solution for both the phase and the amplitude distributions in the aperture of the main reflector of an axisymmetric dual-reflector antenna. Their technique, based on geometrical optics, involves solving two simultaneous, nonlinear, first-order, ordinary differential equations. Figure 45-4 gives the geometry showing the path of a single ray.

FIGURE 45-4

Dual-shaped reflector geometry



45-7

The feed phase center is located as shown, and the feed is assumed to have a power radiation pattern (q1). The parameters a and b represent, respectively, the distance of the feed phase center from the aperture plane and the distance between the feed phase center and the back surface of the subreflector. The constraints to the dual-reflector system are as follows: a. The phase distribution across the main-reflector aperture plane will be uniform, or r1 + ρ 2 + ρ 3 + C p (θ1 ) = constant

(45-8)

over the range 0 ≤ θ1 ≤ θ1max . C p (θ1 ) represents the phase distribution across the primary-feed radiation pattern in units of length. b. The feed energy, or ray bundles intercepted and reflected by the subreflector, is conserved and redistributed according to a specified aperture distribution, or I (θ1 )sin(θ1 ) dθ1 = C ⋅ I ( X 2 ) X 2 dX 2

(45-9)

where I(X2) represents the power radiation distribution across the main-reflector aperture, and C represents a constant that is determined by applying the conservationof-power principle. θ1max

∫0

I (θ1 )sin θ1 dθ1 = C ∫

X2max X2 min

I ( X 2 ) X 2 dX 2

(45-10)

The lower limit of integration over the main reflector can be arbitrarily chosen so that only an annular region of the main reflector is illuminated. c. Snell’s law must be satisfied at the two reflecting surfaces; applying it yields dY1 θ − θ  = tan  1 2  dX1  2 

(45-11)

dY2 θ  = − tan  2  dX 2  2

(45-12)

Solving Eqs. 45-9 through 45-12 simultaneously results in a nonlinear, first-order differential equation of the form dY1 = f (θ1 ,θ 2 ,α , β , etc.) dX 2

(45-13)

which leads to the cross sections of each reflector when subject to the boundary condition Y1 ( X 2 = X 2max ) = 0 , where X2 is the independent variable. Equation 45-13 can be solved numerically by using an algorithm such as a Runge-Kutta, order 4. The above procedure is based on geometrical optics (GO), but it is evident that the assumptions of GO are far from adequate when reflectors are small in terms of wavelengths. An improvement in the design approach is to include the effects of diffraction. Clarricoats and Poulton10 reported a gain increase of 0.5 dB for a diffraction-optimized design over the GO design with a 400l-diameter main reflector and 40l-diameter subreflector.



CHAPTER FORTY-FIVE

Prime-Focus-Fed Parabolic Antennas Prime-focus-fed parabolic reflector antennas are also often employed as Earth station antennas. For moderate to large aperture sizes, this type of antenna has excellent sidelobe performance in all angular regions except the spillover region around the edge of the reflector. The CCIR sidelobe specification can be met with this type of antenna. See Chapter 15 on reflector antennas for more information on parabolic antennas. Offset-Fed Reflector Antennas The advantage of offset-fed reflector antennas is the elimination of feed and feed-support blockage. This can be extremely important for radio-telescope antennas. Because they are asymmetric, offset antennas are more complicated to build per unit aperture than circularly symmetric antennas, but the additional cost of the offset aperture generally isn’t warranted for telecommunications. Consequently, their use is generally reserved for smaller antennas, where eliminating blockage is important and the cost differential for producing the asymmetric aperture is small. Offset-fed reflector antennas can employ a single reflector or multiple reflectors, with two-reflector types the more prevalent of the multiple-reflector designs. The offset frontfed reflector, consisting of a section of a paraboloidal surface (see Figure 45-5), minimizes diffraction scattering by eliminating the aperture blockage of the feed and feed-support structure. Sidelobe levels of (29−25 log q ) dBi can be expected from this type of antenna (where q is the far-field angle in degrees) with aperture efficiencies of 65 to 80 percent. The increase in aperture efficiency compared with that of axisymmetric prime-focus-fed antennas is due to the elimination of direct blockage. For a detailed discussion of this antenna, see C. A. Mentzer.13

FIGURE 45-5

Basic offset-fed paraboloidal antenna



45-9

FIGURE 45-6 Offset dual-reflector geometries: (a) double-offset geometry (feed phase center and paraboloidal vertex at 0), and (b) open Cassegrainian geometry (feed phase center located at 0; paraboloidal vertex at 0)

Offset-fed dual-reflector antennas exhibit sidelobe performance similar to that of front-fed offset reflectors. Two offset-fed dual-reflector geometries are used for Earth station antennas: the double-offset geometry shown in Figure 45-6a, and the open Cassegrain geometry introduced by Cook et al11 of Bell Laboratories and shown in Figure 45-6b. In the double-offset geometry, the feed is located below the main reflector, and no blocking of the optical path occurs. In contrast, the open Cassegrain geometry is such that the primary feed protrudes through the main reflector; thus, it is not completely blockage-free. Nevertheless, both of these geometries have the capability of excellent sidelobe and efficiency performance. The disadvantage of offset-fed dual-reflector antennas is that they are asymmetric, which results in increased manufacturing cost. Also, offset-geometry antennas, when used for linear polarization, have a depolarizing effect on the primary-feed radiation and produce two crosspolarized lobes within the main beam in the plane of symmetry. For circular polarization, this depolarization effect introduces a beam squint whose direction is dependent upon the sense of polarization. The beam squint is approximately given by12 ys = arc sin [l sin (q0) /4p F], where q0 is the offset angle, l is the free-space wavelength, and F is the focal length. The effects of beam squint versus polarization for circular polarization and the relatively high off-axis cross-polarization performance for linear polarization must be considered by the antenna designer, since these characteristics may present a problem for the overall Earth station operation. Offset dual-reflector geometry has the capability of employing a feed tilt to correct for the polarization problems associated with the offset geometry resulting in the equivalent polarization performance of an axially symmetric antenna.13,15 Galindo-Israel, Mittra, and Cha16 have studied the offset-shaped dual-reflector geometry for high-aperture-efficiency applications. Their analytical techniques are reported to result in efficiencies in the 80 to 90 percent range.17 See Galindo-Israel18 for a discussion of the synthesis of offset-shaped reflector antennas.



CHAPTER FORTY-FIVE

Beam-Waveguide Systems A BWG feed system is composed of one or more feedhorns with a series of flat and curved mirrors arranged so that power can be propagated from the horn through the mirrors to the reflector system with minimum losses. Horns and equipment can then be located in a large, non-tipping enclosure at an accessible location. Feeding a large, low-noise ground antenna via a beam BWG system has several advantages over placing the feed directly at the focal point of a dual-reflector antenna. For example, significant simplifications are possible in the design of high-power, water-cooled transmitters and low-noise cryogenic amplifiers. These systems do not have to rotate as in a normally fed dual reflector. Also, since a BWG system can transmit power over considerable distances at very low losses, BWG optics is often used in the design of very highfrequency feed systems. Various design techniques19 have been used, but most frequently they are designed using either GO or Gaussian beams. Geometrical Optics The design for a BWG system using GO is based upon criteria introduced by Mizusawa and Kituregawa20,21 that guarantee a perfect image from a reflector pair. Mizusawa’s criterion can be briefly stated as follows: For a circularly symmetric input beam, the conditions on a conic reflector pair necessary to produce an identical output beam are 1. The four loci (two of which may be coincident) associated with the two curved reflectors must be arranged on a straight line; and 2. The eccentricity of the second reflector must be equal to the eccentricity or the reciprocal of the eccentricity of the first reflector. Figures 45-7a to 45-7c show some curved reflector pair orientations that satisfy Mizusawa’s criteria. Mizusawa’s criteria were used in the design of NASA’s first DSN BWG antenna, which is shown in Figures 45-8 and 45-9 and described in detail in Chapter 7 of Large Antennas of the Deep Space Network.1 The design of the center-fed BWG consists of a beam magnifier ellipse in a pedestal room located below ground level that transforms a 22-dB gain feedhorn into a high-gain 29-dB pattern for input to a Mizusawa four-mirror (two flat and two paraboloid-case) BWG system. The system was initially designed for operation at 8.45 GHz (X-band) and 32 GHz (Ka-band) and has less than a 0.2-dB loss at X-band

FIGURE 45-7

Examples of two curved-reflector BWG configurations



45-11

FIGURE 45-8 Typical center-fed beam-waveguide antenna

FIGURE 45-9 First DSN beam-waveguide antenna named DSS-13



CHAPTER FORTY-FIVE

(determined by comparing the gain of a 29-dB gain horn feeding the dual-shaped reflector system with that obtained using the BWG system) Gaussian beam. While GO is useful for designing systems with electrically large mirrors (>50l diameter with −20 dB or greater edge taper), some BWGs may be operated at low frequencies where the mirrors may be as small as 20l in diameter. Due to diffraction effects, the characteristics of a field propagated between small BWG mirrors ( 25 dB to produce useful DF. Both Root MUSIC and classic MUSIC can resolve highly correlated, closely spaced signals given sufficient ASNR.


Direction Finding Antennas and Systems 47-26

CHAPTER FORTY-SEVEN

FIGURE 47-19 DF performance for two closely spaced signals that are 95 percent correlated

47.5 GEOLOCATION ALGORITHMS In this section, we present three geolocation algorithms. The Stansfield algorithm is the oldest and has been used successfully in both worldwide HFDF systems and line-of-sight angle-of-arrival systems as well. The Wangsness algorithm uses LOBs drawn on a spherical earth and likewise can be used globally or locally for geolocation solutions. A simple 2D TDOA geolocation algorithm is also presented. These three approaches are compared using confidence ellipse containment. Stansfield AOA Geolocation Algorithm The Stansfield algorithm37 was developed during World War II for AOA geolocation using a flat (Cartesian) geometry, illustrated in Figure 47-20. The LOB from a receiving site has a bearing angle b from north (y-axis). Using an arbitrary coordinate system, the LOB is also described by d, the perpendicular distance to the coordinate system origin; m is the LOB miss distance from point (px, py)T. The miss distances from a set of K lines of bearing with common Cartesian origin created by some number of DF sites is written in matrix form as m = Ap + d

(47-47)

where m is the miss vector, A is a transformation matrix, p is the position vector, and d is the LOB origin offset parameter:  − cos(β1 ) sin(β1 )   m1   d1   − cos(β2 ) sin(β2 )   m2  d   px  m =  , A =   , p =  p  , and d =  2      y      dK  mK  − cos(β K ) sin(β K )


Direction Finding Antennas and Systems DIRECTION FINDING ANTENNAS AND SYSTEMS

FIGURE 47-20

47-27

Stansfield geolocation geometry

The goal is to find a position pˆ that minimizes mT m = Σ1K mi2, which can be done in an iterative approach using the gradient form of Eq. 47-47, The iterative improvement in position is

and

dp = inv( AT A) ⋅ ( AT m )

(47-48)

pˆ new = pˆ old + dp

(47-49)

Inspection of the components of Eq. 47-48 reveals that the ATA and ATm matrices are the classic unweighted least squares normal equations. These should be used on the first iteration from an arbitrary starting point po. Subsequent iterations should use a weighted 2 version of ATA and ATm that account for the LOB a priori DF variance σ LOB and distance ri from the ith bearing site to emitter. Using a covariance matrix R whose diagonal elements 2 ri2 ) , Eq. 47-48 is modified to be contain the squared weights wi2 = 1 / (σ LOB dp = inv( AT RA) ⋅ ( AT Rm )

(47-50)

The weighted normal equations are written as K  cos 2 (βi )  ∑ σ 2 r2 LOB i (AT RA) =  K i =1  cos(βi )sin(βi ) − ∑ 2 σ LOB ri2  i =1

cos(βi )sin(βi )   2 σ LOB ri2 i =1  K 2  siin (βi ) ∑ σ LOB  2 2 r i i =1  K

−∑

(47-51)



CHAPTER FORTY-SEVEN

TABLE 47-2

Ellipse Confidence Factor

Ellipse Confidence Factor

Scaling Factor ( jo)

68%

1

80%

1.79

90%

2.15

95%

2.45

 K m i cos(βi )  − ∑ σ 2 r 2  LOB i T  (A Rm) =  Ki =1  m i sin(βi )   ∑ σ 2 r2   i =1 LOB i 

and

(47-52)

Convergence to pˆ requires only three or four iterations from an arbitrary starting point, making the Stansfield geolocation algorithm fairly efficient. It is customary to estimate a confidence containment ellipse about the location pˆ , which can be determined from the eigenvalues of the inv(ATRA) matrix. Letting the inverse matrix be noted as Q, the eigenvalues are found by solving the determinate |lI − Q| = 0. Two eigenvalues result, designated l+ (maximum) and l− (minimum). The confidence ellipse is scaled by the number of LOBs, K, and by a scaling factor, jo, which establishes the statistical confidence level (see Table 47-2). The confidence ellipse is then described by the semi-major and semi-minor axes: s maj = joλ +

K K and s min = joλ− K −1 K −1

(47-53)

The unambiguous ellipse orientation (see Table 47-3) requires testing of the terms of the Q matrix. A Monte Carlo simulation of LOBs from three bearing sites illustrates the Stansfield algorithm. LOBs with 2° rms error were generated from three sites, producing a best estimate position and 95 percent confidence containment ellipse (see Figure 47-21a). To evaluate 300 trials, the confidence ellipse, normally centered on each estimated position, is shifted to the true emitter location, and allows comparison of the theoretical 68%, 80%, and 95% confidence ellipse containment with the trial estimated positions (see Figure 47-21b). TABLE 47-3

Confidence Ellipse Orientation

Compare Diagonal Terms Q1,1 Q2,2

where

Orientation 90 −

1  y term  atan 2  2  x term 

1  y term  atan 2  2  x term 

y term = Q1,2 + Q2,1 x term = Q1,1 − Q2,2



47-29

FIGURE 47-21 (a) Stansfield LOBs, geolocation, and 95% confidence ellipse and (b) 300 trials with 68%, 80%, and 95% confidence ellipse centered on the true emitter

Wangsness AOA Geolocation Algorithm In 1973 Dennis Wangsness proposed an elegant solution to geolocation on a spherical earth.38 It has the advantage of being a fast, non-iterative eigenvector approach to produce both geolocation and a confidence containment ellipse. Extending the spherical solution onto a geoid allows the algorithm to be WGS-84 compliant. Wangsness recognized that any LOB describes a great circle around the earth and is anchored through the bearing site. These great circles are described by perpendicular normal vectors, n, which collectively produce an eigenvector solution for best estimate of position and error ellipse. First, the LOB normal vectors are described in terms of the Earth Centered Earth Fixed (ECEF) coordinate system and are collected into transformation matrix A for K LOBs:  nx1 A=   nxK

ny1  nyK

nz1  T   = n1  n i  n K   nzK 

(47-54)

nxi  where ni = nyi  is the ith LOB normal vector in ECEF coordinates.   nzi  Once the ATA matrix is formed, the Wangsness algorithm solves for the three eigenvalues and associated eigenvectors using |lI − ATA| = 0. The normalized eigenvector associated with the smallest eigenvalue is a unit vector that points toward the best estimate of emitter position. This unit vector can be transformed into spherical latitude and longitude and further transformed into geodetic latitude and longitude using a standard geoid such as WGS-84. Although the eigen solution is not iterative, once an approximate emitter position is found, a weighted solution using the covariance error matrix R = diag(1 / ri2 ) , where ri is the arc distance between the ith-LOB bearing site and the estimated location. The ATA matrix now becomes the weighted matrix ATRA.



CHAPTER FORTY-SEVEN

The three eigenvalues of |lI − ATRA| = 0 are sorted smallest to largest (l1, l2, and l3) to provide the confidence ellipse equations: s maj =

ae ⋅ jo K −2

⋅

λ1 λ2

and

s min =

ae ⋅ jo K −2

⋅

λ1 λ3

(47-55)

where K is the number of LOBs, jo is the ellipse scaling factor for confidence of containment, and ae is the earth’s radius to convert arc distances into meters. Ellipse orientation is derived from the arc cosine of the dot product formed by the eigenvector associated with l 2 (the eigenvalue associated with semi-major axis) and a unit vector at the estimated location that points east. A similar Monte Carlo simulation was run for the Wangsness algorithm with the same LOB error and site geometry of the Stansfield simulation. LOB, best position, and confidence ellipse were computed (see Figure 47-22a) and 300 trials were made (see Figure 47-22b). Again, the confidence ellipse was shifted to the true emitter location, for comparison of the theoretical 68%, 80%, and 95% confidence containment ellipses. TDOA Geolocation Algorithm When the LORAN hyperbolic location system was first developed by MIT in 1942, it was computationally intensive and highly classified.39 Now TDOA geolocation can be done with small radio tuners and inexpensive computer chips. The algorithm presented here is consistent with the two-dimensional Stansfield approach, but can be easily extended to a three-dimensional solution on a spherical earth or WGS-84 geoid.40,41 The time difference of a signal arriving at two sites s1 and s2 from an emitter at so can be written in terms of the two distances D1 and D2 between the sites and emitter, scaled by the speed of light, c: TDOA ik =

Di − Dk c

(47-56)

where Di is the distance between site si and emitter at so

FIGURE 47-22 (a) Wangsness LOBs, geolocation, and 95% confidence ellipse and (b) 300 trials with 68%, 80%, and 95% confidence ellipse centered on the true emitter


Direction Finding Antennas and Systems DIRECTION FINDING ANTENNAS AND SYSTEMS 1

Di = s i − s o = {(x i − x o )2 + ( yi − y o )2} 2

47-31

(47-57)

The hyperbolic TDOA solution of location is written in an iterative, gradient form similar to the Stansfield approach of Eqs. 47-48 and 47-49:

and

ds o = (AT A) ⋅ d (TDOA)

(47-58)

sˆ new = sˆ o + ds o

(47-59)

The matrix d(TDOA) is a K × 1 vector of K observed minus calculated TDOA values from the i-k site pair, where the calculated TDOAik and its derivative are based on the current estimate of emitter position sˆ o . The ATA matrix is a K × 2 transformation matrix of partial derivatives of TDOAik with respect to sˆ o . The fundamental d(TDOA) vector and A matrix are written as  TDOA ik ( measured )1 − TDOA ik ( calculated )1   TDOA ( measured ) − TDOA ( calculated )  2 2 ik ik d ( TDOA) =      TDOA ik ( measured ) K − TDOA ik ( calculated ) K 

and

  ∂TDOA ik  dxo    ∂x o 1     ∂TDOA ik  dxo   A =   ∂xo 2    ∂TDOA  ik  dx o   ∂xo K

   ∂TDOA ik dyo    ∂y o 1      ∂TDOA ik dyo    ∂y o 2         ∂TDOA ik dyo   ∂y o K  

(47-60)

(47-61)

The partial differential of TDOAik with respect to the current estimate of emitter position is further expanded using the definitions of time delay from Eqs. 47-56 and 47-57:

and

∂TDOA ik 1  ∂Di ∂Dk  1  ( xo − xi ) ( xo − xk ) =  − = − c  ∂xo ∂xo  c  Di Dk  ∂xo

(47-62)

∂TDOA ik 1  ∂Di ∂Dk  1  ( yo − yi ) ( yo − yk ) =  − = − c  ∂yo ∂yo  c  Di Dk  ∂yo

(47-63)

At least two different site pairs are required for geolocation convergence, and care must be taken to avoid ambiguities, since, unlike LOB observations, the hyperbolic TDOA isochrones have left-right symmetry about a line between the measuring i-k site pair.



CHAPTER FORTY-SEVEN

FIGURE 47-23 GDOP for TDOA geolocation from three sites. Lightest color indicates lowest GDOP and area of best geolocation.

A useful concept in evaluating the geolocation geometry is a quantity called the Geometric Dilution of Precision (GDOP), defined as GDOP = trace(inv(AT A))

(47-64)

A low value of GDOP indicates good geometry for the geolocation solution, while a high GDOP indicates poor geometry (see Figure 47-23). A Monte Carlo simulation with a geometry similar to Stansfield and Wangsness was run for TDOA geolocation (see Figure 47-24). Each TDOA has s TDOA = 50-ns rms error.

FIGURE 47-24 (a) TDOA hyperbolic isochrones, geolocation, and 95% confidence ellipse and (b) 300 trials with 95% confidence ellipse centered on the true emitter



47-33

Although the TDOA algorithm estimates position entirely in the time domain, the simulation also drew hyperbolic isochrones for comparison with the LOB geolocation approach (compare Figure 47-24a with Figures 47-21a and 47-22b). Four TDOA measures were made from each site pair (shown as dotted lines connecting the sites). The calculation of the confidence error ellipse closely parallels that of the Stansfield algorithm described by Eq. 47-53 and Table 47-3. In Figure 47-24b, the confidence ellipse is shifted to the true emitter location to examine the 95% containment of 300 geolocation trials. The TDOA geolocation error is distributed very differently from the LOB geolocation error (compare Figure 47-24b with Figures 47-21b and 47-22b).

REFERENCES 1. F.E. Terman, Radio Engineering (New York: McGraw-Hill, 1947): 817. 2. National Telecommunications and Information Administration (NITA), Manual of Regulations and Procedures for Federal Radio Frequency Management, Section 6.1.1 (Jan. 1993): 6 –12. 3. F. B. Gross, Smart Antennas for Wireless Communications with MatLab (New York: McGrawHill, 2005). 4. S. C. Hooper, “Navy History—Radio, Radar, and Sonar,” Recordings, Office of Naval History, Washington, D.C., 31R76. [Quoted references may be found in L. S. Howeth, History of Communications-Electronics in the United States Navy, Chapter XII, “The Radio Direction Finder”: http://earlyradiohistory.us/1963hw22.htm.] 5. L. S. Howeth, History of Communications-Electronics in the United States Navy (Washington, D.C.: United States Government Printing Office, 1963): http://earlyradiohistory.us/1963hw.htm. 6. F. H. Engel and F. W. Dunmore, “A Directive Type of Radio Beacon and Its Application to Navigation,” Scientific Papers Bureau of Standards, vol. 19 (1923): 281–295. 7. G. Spinger, “Radio Systems in the Early A6M Zero,” http://www.jaircraft.com/research/ gregspringer/radios/radio_systems.htm. 8. “History or Radio Research at Ditton Park,” http://www.dittonpark-archive.rl.ac.uk/histTime .html. 9. D. Kahn, The Codebreakers (New York: Macmillan Co., 1969): 8/504. 10. V.A. Kotelnikov, “Determination of the Elements of the Orbit of a Satellite Using the Doppler Effect,” Radiotechnika and Electronika, vol. 7 (July 1958): 873–991. 11. C. Soanes, “Orbital Navigation Systems—Present and Future Systems,” Dynamic Positioning Maritime Technology Society Conference, London, Sept. 28–30, 2004. [Paper available at http:// dynamic-positioning.com/dp2004/sensors_soanes.pdf.] 12. P. Daly and G. E. Perry, “Recent Developments with the Soviet Union’s VHF Satellite Navigation System,” Space Communications and Broadcasting, vol. 4 (1986): 51–61. 13. F. W. Lehan and G. L. Brown, Discovery and Location System (Nov. 6, 1962): U.S. Pat. 3,063,048. 14. http://www.cospas-sarsat.org/ and http://www.sarsat.noaa.gov/. 15. R. O. Schmidt, “Multiple Emitter Location and Signal Parameter Estimation,” IEEE Trans. on Antennas and Propagation, vol. AP-34(3) (1986): 276–280. 16. R. Roy and T. Kailath, “ESPRIT—Estimation of Signal Parameters via Rotational Invariance Techniques,” IEEE Trans. Acoustics, Speech, and Sig. Proc., vol. ASSP-37 (July 1989) 984 –995. 17. A. Moffet, “Minimum-Redundancy Linear Arrays,” IEEE Trans AP, vol. 16(2) (1968): 172–175. 18. S. Stein, “Algorithms for Ambiguity Function Processing,” IEEE Trans. Acoustics, Speech, and Signal Processing, vol. ASSP-29 (June 1981): 588–599. 19. C. Knapp and G. C. Carter, “The Generalized Correlation Method for Estimation of Time Delay,” IEEE Trans. Acoustics, Speech, and Sig. Proc., vol. ASSP-24 (Aug. 1976). 20. P. J. D. Gething, Radio Direction Finding and Superresolution, 2nd Ed. (London: Peter Peregrinus Ltd. on behalf of IEE, 1991): 94–113.



CHAPTER FORTY-SEVEN

21. C. W. Earp and R. M. Godfrey, “Radio Direction Finding by Measurement of the Cyclical Difference of Phase,” J. IEE (London), part IIIA, vol. 94 (March 1947): 705. 22. F. Adcock, Improvements in Means for Determining the Direction of a Distant Source of ElectroMagnetic Radiation (1919): British Pat. 130,490. 23. R. Poisel, Introduction to Communication Electronic Warfare Systems (Boston: Artech House, 2002): 352–359. 24. D. H. Johnson, “The Application of Spectral Estimation Methods to Bearing Estimation Problems,” Proc. IEEE, vol. 70(9) (Sept. 1982): 1018–1028. 25. J. P. Burg, “Maximum Entropy Spectrum Analysis,” Ph.D. dissertation, Dept. of Geophysics, Stanford University, Stanford, CA, 1975. 26. V. F. Pisarenko, “The Retrieval of Harmonics from a Covariance Function,” Geophys. J. Royal Astronomical Soc., vol. 33 (1973): 347–366. 27. R. Kumaresan and D. W. Tufts, “Estimating the Angles of Arrival of Multiple Plane Waves,” IEEE Trans. Aerospace Electron. Sys., vol. AES-19 (Jan. 1983) 134–139. 28. I. Ziskind and M. Wax, “Maximum Likelihood Localization of Multiple Sources by Alternating Projection,” IEEE Trans. Acoustics, Speech, and Sig. Proc., vol. ASSP-36 (Oct. 1988): 1553–1560. 29. E. Anderson et al, LAPACK Linear Algebra Package, http://math.nist.gov/lapack++/. 30. J. Volder, “The CORDIC Computing Technique,” IRE Trans. Computers, vol. EC-8 (Sept. 1959): 330–334. 31. A. Ahmedsaid, A. Amira, and A. Bouridane, “Improved SVD Systolic Array and Implementation on FPGA,” Proc. 2003 IEEE Inter. Conf. on Field-Programmable Technology (FPT) (Dec. 15–17, 2003): 35–42. 32. G. W. Stewart, Matrix Algorithms Vol. II: Eigensystems (Philadelphia: Society for Industrial and Applied Mathematics, 2001): 56–70. 33. H.L. Van Trees, Optimum Array Processing (New York: Wiley Interscience, 2002): 925–932. 34. H.L. Van Trees, Optimum Array Processing: 827–845. 35. H. Akaike, “A New Look at the Statistical Model Identification,” IEEE Trans. Automatic Control, vol. AC-19 (June 1974): 716–723. 36. M. Wax and T. Kailath, “Detection of Signals by Information Theoretic Criteria,” IEEE Trans. Acoustics, Speech, and Sig. Proc., vol. ASSP-33 (April 1985): 387–392. 37. R. G. Stansfield, “Statistical Theory of DF Fixing,” J. IEE (London), part IIIA, vol. 94 (1947): 762–770. 38. D. Wangsness, “A New Method of Position Estimation Using Bearing Measurements,” IEEE Trans. Aerospace and Electronic Sys. (Nov. 1973): 959–960. 39. D. A. Grier, When Computers Were Human (Princeton, NJ: Princeton University Press, 2005): 249–250. 40. A. D. Stewart, “Comparing Time-Based and Hybrid Time-Based/Frequency Based Multi-Platform Geolocation Systems,” Thesis, Naval Postgraduate School, Monterey, CA, Sept. 1997. 41. H. C. Schau and A. Z. Robinson, “Passive Source Location Employing Spherical Surfaces from Time-of-Arrival Differences,” IEEE Trans. Acoustics, Speech, and Sig. Proc., vol. ASSP-35 (Aug. 1987) 1223–1225.



Chapter 48

ESM and ECM Antennas Nikolaos K. Uzunoglu George Geroulis National Technical University of Athens

CONTENTS 48.1 INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48-2 48.2 COMMUNICATION ESM AND ECM SYSTEM ANTENNAS. . . . . . . 48-3 48.3 RADAR ESM AND ECM SYSTEM ANTENNAS . . . . . . . . . . . . . . . . 48-9


ESM and ECM Antennas 48-2

CHAPTER FORTY-EIGHT

48.1 INTRODUCTION ESM and ECM Systems The use of electronic countermeasures (ECM) against opponents’ communications systems started with the use of wireless communications in military operations at the beginning of the 20th century. Extensive electronic monitoring/interception of nonfriendly signals and countermeasures against both communications links (primarily high frequency [HF] at that time) and newly developed radar systems began during World War II (1939–45). The use of the electromagnetic spectrum in military operations was defined as “electronic warfare.” Soon these activities were classified as electronic support measures (ESM). ESM consisted of passive measurements of emitted signals by nonfriendly emitters, and ECM systems, which included the emission of signals to confuse or deceive the communications and radar sensor systems of opposing forces. Extensive use of passive ECM techniques, such as chaff (a large number of dipole elements spread into the air), began toward the end of WWII. Electronic counter-counter measure (ECCM) techniques are aimed to overcome or decrease the ECM threats of friendly communications and sensor systems. To this day, ECM and ECCM techniques are in a continuous competition. The design and development of antennas intended to serve the needs of ESM and ECM systems poses unique requirements, as compared to communications or radar antennas, such as ●

●

●

●

Ultrawide spectral bandwidths Low weight and size Full polarization control Very high isolation between transmit and receive in collocated antennas

The last requirement is of particular significance, since the ability to receive weak signals while transmitting countermeasure (jamming) signals is of primary importance. Achievement of this goal provides superior ECM abilities. Indeed, the capability to regenerate threat signals improves the efficiency of ECM systems to a high degree. General Properties of ESM and ECM Antennas In treating the ESM/ECM antennas the spectral region of interest has primary importance. In tactical communications systems the frequency region 1–2500 MHz (HF, VHF) is used predominantly, while radars use microwave (1–30 GHz) and millimeter wavelength (30–100 GHz) spectral regions. The wideband antenna technologies, which started to appear during the mid-1950s, are widely used in both ESM and ECM applications concerning both communications and radar systems. The extensive use of multiple communications networks at the same geographical place and jamming-resistant wideband systems forces the use of wideband systems, thus requiring ultra wideband (UWB) ESM and ECM antennas. Furthermore, especially in low-frequency ESM and ECM systems, the relatively large size of antennas necessitates the use of the same antenna aperture for various needs such as signal interception and emission at various frequency bands almost simultaneously. The development of monolithic integrated circuits (based on GaAs or Si technologies) after the 1980s enabled the use of phased-array principles in ESM and ECM systems.


ESM and ECM Antennas ESM AND ECM ANTENNAS

48-3

The availability of true time-delay, digitally controlled, two-port networks allowed the development of wideband phased-array antennas, which provide superior operational characteristics such as the following: ●

●

●

●

Highly dynamic beam agility in one or two planes Ability to modify outgoing wavefront to effect monopulse radars Use of nulling at the direction of arrival energy jamming signals without significant degradation of radiation pattern at other angles Highly effective irradiated power toward threats

In specifying and designing ESM and ECM antennas, the standard quantities used in general antenna theory are used, such as directivity (D), gain (G), bandwidth (BW), and radiation patterns at various levels. In ECM and ESM system antennas, the polarization properties of antennas are of paramount importance to performance. The polarization behavior of ECM systems, and to some extent ESM systems, has essential effects on the ability of radars to resist jamming and deception phenomena (ECCM). Furthermore, the level of isolation between receiver and transmitter antennas in an ECM system, either for communications or radar, is of fundamental importance because it affects to a significant degree the countermeasure capabilities. In this chapter, the ESM and ECM systems antennas are reviewed independently concerning the communications and radar systems. The presentation reflects the practical use of antennas in the corresponding applications based on generic concepts. In the case of communications systems, attention is focused on the low frequencies (1–2500 GHz), while for radar applications the microwave (1–30 GHz) and millimeter wavelength (30–100 GHz) bands are considered. In all the considered cases, the mechanical, environmental, and radar cross-section implications are also considered in practical terms. In both cases, the directionfinding antenna principles are also discussed and investigated based on various monopulse principles. At the end of the chapter, the important issue of transmit-receive antenna isolation in ECM systems is analyzed and general rules to achieve maximum isolation are suggested.

48.2 COMMUNICATION ESM AND ECM SYSTEM ANTENNAS Mission of Communication ESM and ECM Systems Communication ESM Systems The operational mission of communication ESM systems is to detect-receive incident, usually very weak, signals and convey them to sensitive wideband receivers. There are two fundamental operational missions of ESM systems: (a) acquisition and analysis of incoming signals and (b) measuring direction of arrival (direction finding, or DF) at the horizontal plane with respect to the Earth’s surface. In the former case, signal parameters are measured without determining the angle of arrival in the horizontal plane. Usually, omnidirectional antennas are utilized to receive signals without measurement of angle of arrival. Near the earth’s surface, at the low radio frequencies (RF), 1–500 MHz, only vertical polarization is a matter of interest, while at higher frequencies, reception horizontal polarization should also be taken into account.



CHAPTER FORTY-EIGHT

In the case of DF systems, monopulse antenna principles based on the simultaneous measurement and comparison of received signal amplitudes or phases is utilized (see Chapter 47). The DF systems are required to measure wideband signals such as the case of frequency-hopping and/or spread-spectrum modulated carriers, which pose serious difficulties in achieving azimuth angle measurement accuracies on the order of a few degrees. Furthermore, a dynamic range of at least 80 dB is required. The extensive use of digital signal processing techniques during recent years has enabled simultaneous measurement of DF and signals tracking of frequency-hopping and spread-spectrum signals. The combination of DF measurement techniques and use of signal processing techniques such as FFT and wavelet transform has provided the capability of efficient multiple signals processing and de-interleaving. Therefore, the modern trend is to have combined DF and signal acquisition-analysis ESM receivers, which require a single antenna for both missions. Communication ECM Systems The primary role of an ECM antenna in communications applications is to direct the radiated countermeasure signals toward the enemy receivers to achieve degradation or complete loss of signal reception at the opponent’s receiver units. This means that the azimuthal orientation of the victim receiver (or receivers) should be known with an accuracy of at least 10° in tactical communications using VHF (30–90 MHz) or UHF (200–2500 MHz) frequencies. In such cases, mobile ECM systems are usually used, placed inside a shelter for land-based systems. Ships or airplanes (manned or unmanned) are also used to a lesser degree. A significant difficulty arises because of the large antennas needed for ECM tactical systems required to cover the low VHF region, corresponding to this wavelength l = 3–10 m. Concerning HF communication ECM systems, usually land-based systems, antenna arrays with wideband properties are used. In this case, large-size fixed-array installations are used. In mobile, land-based HF ECM systems, usually omnidirectional or mediumdirectivity antennas with wideband characteristics are used. Omnidirectional Antennas To achieve wideband reception of signals omnidirectionally, the most commonly used antennas are based on conical type structures, as shown in Figure 48-1, which also shows the corresponding critical design parameters.

θ Conical Monopole ~90°

Biconical Antenna

h

α Coaxial Line

h

Discone Antenna

FIGURE 48-1 Family of conical antennas


ESM and ECM Antennas ESM AND ECM ANTENNAS

48-5

In practice, the geometric size restrictions pose difficulties in achieving lowfrequency coverage. The use of conical monopole or discone antennas usually is preferable since the required antenna heights are h = 2 λ 8 and h = l/8, respectively. In the case of VHF and UHF frequencies, to reduce the antenna weight and wind load, radial conductors are used as shown in Figure 48-2. The number or radials should be at least 16 peripherally to achieve sufficiently dense current distributions. A significant issue in omnidirectional antennas is the ability to receive and transmit both polarizations simultaneously, especially in the case of microwave signals and radar signals. The simplest approach is to polarize the incident-outgoing waves by using a grid of parallel wires, as shown in Figure 48-3. The distance between the two parallel running wires should be less than lM / 20, where lM is the maximum wavelength of the radiation. Usually, printed strips of approximately lM / 100 on a low-permittivity substrate are used. The drawback of this method is the 3 dB insertion loss, concerning either vertical or horizontal polarization. In the case of HF (wavelength l = 100–10 m) and VHF (l = 10–1 m) frequencies, reception of both polarizations rarely is needed. Therefore, in the case of UHF (l = 1–0.1 m) and SHF, EHF (l = 10–1 cm) frequencies, in order to receive both polarizations with an omnidirectional radiation pattern, one can use independent antennas and then use either independent analog summation of the two signals or independent signal processing channels. Figure 48-4 depicts two techniques to achieve reception and transmission for both polarization antenna systems. In Figure 48-4a the printed and slot triangular butterfly antennas

FIGURE 48-2 Discone antenna made of radial conductors



CHAPTER FORTY-EIGHT

45°

FIGURE 48-3 Polarized discone conical antenna

are placed collinearly to simultaneously radiate both polarizations. Another example is shown in Figure 48-4b where orthogonal conical dipoles are used to receive at all possible polarizations and incident angle signals. Directional Antennas The antenna type that has been used, and still is widely used, in ESM and ECM communications systems is the logarithmic periodic antenna (LPA), which

Electric Dipole Butterfly Antenna (Vertical Polarization)

Horizontal Conical Dipoles

Butterfly Slot (Magnetic Dipole Antenna (Horizontal Polarization)

(a)

(b)

FIGURE 48-4 Combination of electric-magnetic wideband dipoles



ESM AND ECM ANTENNAS

was invented by Raymond H. Du Hamel.1–3 The following are the critical design parameters in LPAs: ●

●

●

Maximum and minimum operation frequency The required directivity and gain Polarization characteristics

The basic structure of an LPA is shown in Figure 48-5. The basic design equations of the structure of an LPA are also given. Although numerous variations of LPAs have been developed and used during the last 50 years, the structure with multiple dipole elements has been the most commonly used ECM and ESM wideband antenna in practice. LPAs could provide antenna gains on the order of 7-10 dBi depending on the number of elements (N) and apex angle (a). If more directivity is required, use of multiple LPAs proved to be a useful technique, such as shown in Figure 48-6, where the two LPAs are placed parallel with their axis forming the angle b and the distance between the largest dipole is equal to the largest dipole (half wavelength lM / 2 at the largest frequency). ln + 1 hn + 1 = =τ ln hn tan ( a / 2 ) =

ln − ln + 1 hn

Maximum operation wavelength = lM = 4l1 (largest dipole double length)  β λ 4 sin   = Μ L  2 where N

L = ∑ hn n =1

N = number of elements

L h1

l1

h3

h2

l2

l3

λM/2

l4 α

l1

l2

l3

λM/2

λM/2 λM/2 λM/2

l4

β

λM/2

FIGURE 48-5 Array of log-periodic arrays



CHAPTER FORTY-EIGHT

Communication Systems Direction Finding Antennas The simultaneous reception of signals with different radiation pattern antennas—a principle known as monopulse—is used to measure the angle of incidence of communications and radar signals. Usually the azimuth angles of the incident signals are measured. In this case, an even number of wideband antennas, such as n = 4, 6, 8, etc. directive antennas, are used, as shown in Figure 48-6. For low frequency communications systems, only vertical polarization reception is typically required, simplifying the antenna element construction. A commonly used wide antenna type is the transverse electromagnetic mode (TEM) horn antenna, shown in Figure 48-6a, in which log-periodic elements are also used (see Figure 48-6b). Analog signal processing techniques consisting of hybrid circuits (sum and difference) and down converter units are employed to drive the A/D conversion circuits for direction finding (see Chapter 47) using the two signals X(j) and Y(j). Assuming each individual antenna in Figure 48-6 has a radiation pattern function described by an even function F(j) = F(−j) with respect to the j = 0 axis, the X(j) and Y(j) functions are computed as follows: X (ϕ ) = F (ϕ ) − F (ϕ − 180°) Y (ϕ ) = F (ϕ − 90°) − F (ϕ − 270°) By comparing the two signals X(j), Y(j) we can then obtain the j angle as a function of the X and Y signals on the assumption of a monotonically decreasing function. Calibration techniques should be used to account for any frequency dependence of the F(j) radiation pattern function. Increased bandwidths and conversion rates of the A/D converters provides opportunity for eliminating the necessity of using analog processing units to obtain the X(j) and Y(j) signals by digital processing. Instead of using directive antenna elements, many systems use a more simplified technique, known as the Watson-Watt technique, shown in Figure 48-6c, where four dipole elements are used to obtain the X(j) and Y(j) functions. The distance between the two antipodal dipole elements should be less than half a wavelength at the maximum operation frequency. This condition guarantees the unique dependence of j to X and Y voltages. To cover wide bandwidths, three or four sets of collinearly placed quadrant dipole or monopole elements are used along the vertical axis.

3 α

2 3

4

2

4 1

2

3 4

1

1

Four TEM Horn Elements

Four Log-Periodic Antenna

α

Antenna engineering handbook

Antenna Engineering Handbook

Antenna handbook

Antenna handbook

Smart Antenna Engineering

Practical Antenna Handbook

Antenna Handbook - Applications

Phased Array Antenna Handbook

Antenna Handbook - Theory

Antenna Handbook - Related Topics

Practical antenna handbook

Modern antenna handbook

Antenna Handbook (1999)(en)(196s)

Antenna Handbook MCRP 3-40.3C

Small Antenna Design (Communications Engineering Series)

Electromagnetic Band Gap Structures in Antenna Engineering

Electromagnetic Band Gap Structures in Antenna Engineering

Antenna handbook. Vol.4, Related topics

Antenna Handbook - Fundamentals and Mathematical Techniques

Clinical Engineering Handbook (Biomedical Engineering)

Clinical Engineering Handbook (Biomedical Engineering)

Antenna Toolkit

Antenna Toolkit

Antenna Technology

The ocean engineering handbook

Optomechanical Engineering Handbook

Optomechanical Engineering Handbook

Engineering Design Reliability Handbook

The Computer Engineering Handbook

The ocean engineering handbook

Adaptive Optics Engineering Handbook

Antenna engineering handbook