Assembly Automation Volume 23, Number 2, 2003
ISSN 0144-5154
Adhesives + welding + joining Guest Specialist: Dr Robert W. Messler
Contents 119 Access this journal online 120 Abstracts & keywords 122 Editorial
Viewpoint 123 Joining research and education: Where must it go? How must it change? Robert W. Messler
Company news 125 A review of the latest developments in the industry
Keynote 130 Joining comes of age: from pragmatic process to enabling technology Robert W. Messler
147 Productivity and quality improvements through orbital forming Werner R. Stutz 153 Robotic ‘‘layup’’ of composite materials David Groppe 159 Advances in resistance welding for body-in-white Brian Rooks 163 IMTS leads with technological innovation Dick Bloss 166 Robotics and assembly automation at TEAM Brian Rooks 172 A novel indexing mechanism for paper cutting machines Rajanish K. Kamat
Features 144 Clinching with a superimposed movement – a method for force reduced joining Rolf Dieter Schraft, Stefan Schmid and Achim Breckweg
Research articles 174 The automated filling of bonded joints – Part 2 three dimensional joints Ken Young and Ian Pearson
Access this journal electronically The current and past volumes of this journal are available at
www.emeraldinsight.com/ft You can access over 100 additional Emerald journals, each with a comprehensive searchable archive of articles (many dating back to 1989), a detailed classification system and links to referenced material. See page following contents for full details of what your access includes.
Contents
(continued)
181 Optimum assembly design utilizing a behavioral modeling concept Y.J. Lin and R. Farahati 192 A simulation method and distributed server balancing results of networked industrial robots for automotive welding and assembly lines Paul G. Ranky 198 Mini features .
Welding of body scanner vessels improved using robot vision system
207
.
A big linear step forward. . . for small applications
.
Linear motor driven stage brings ‘‘nano-positioning’’ capability to large-travel applications
.
Worldwide adhesive coating manufacturer uses online measurement system to improve production efficiency and quality control
.
Araldite SMC bonds Scania truck grills
New products Updated information on the latest entrants in the automated assembly field
.
Weld procedure development with OSLW
.
Simultaneous engineering the key to success
212
Internet page
New automation system keeps manufacturing in the UK at Medic-aid
214
Book reviews
.
Axxom visualizes warehouse locations graphically
216
Patent abstracts
.
Practicality rules when choosing a servo bus system for machine building
218
Diary
.
Conferences, seminars and exhibitions
Productivity and quality improvements through orbital forming Werner R. Stutz
Abstracts & keywords
Keywords Assembly, Metal forming
Joining comes of age: from pragmatic process to enabling technology Robert W. Messler Keywords Joints, Biotechnology, Information technology, Microelectronics
Orbital forming is an efficient and precise process to assemble component parts. It provides strength, an attractive finished appearance, and batch-to-batch uniformity. Orbital forming machines can produce high-torque assemblies and also freely swinging joints, and any degree of built-in resistance in between. These machines quietly flare and form all malleable materials, including many engineering thermoplastics, and work safely on delicate and brittle parts. The machine controls provide infinitely variable cycle times (speed), forming pressure and tool stroke on the micrometer dial with resolution to 0.001 in.
Robotic ‘‘layup’’ of composite materials David Groppe
From when we, as humans, first lashed a pointed stone to a split straight stick to make a more effective spear for hunting to now when we fasten and bond ablative ceramic tiles to the frail metal skin of the Space Shuttle to allow safe re-entry from manned excursions into space, joining has been a pragmatic, albeit critically important, fabrication process. As we move beyond the Industrial Age to the ages of Information Technology, Nanotechnology, and Biotechnology, joining must move from a secondary process for manufacturing objects or articles from pre-synthesized and pre-shaped materials to a primary process for combining materials into fundamental structures as these structures and even materials are being synthesized; where the boundary between the materials and the structure becomes blurred. This paper attempts to catch a glimpse of the future where joining comes of age to become an enabling technology practiced as much or more by technicians or physicians than as a trade practiced by helmeted welders or hard-hatted riveters.
Clinching with a superimposed movement – a method for force reduced joining Rolf Dieter Schraft, Stefan Schmid and Achim Breckweg Keywords Assembly, Joining, Metals Clinching is, due to its characteristics, a joining method with several advantages. The high joining forces, which require heavy process equipment are a major disadvantage. The Fraunhofer Institute has developed clinching methods which reduce the joining forces considerably to make clinching applicable for further developments and new application areas. Assembly Automation Volume 23 · Number 2 · 2003 · pp. 120–121 q MCB UP Limited · ISSN 0144-5154
120
Keywords Composite materials, Robots, Automation This paper describes the history and current technology behind composite manufacturing and the development of a precision feed endeffector (PFE). The PFE is used on the end of a robot arm and performs many functions associated with the handling of prepreg and semipreg materials. The PFE helps to achieve higher levels of accuracy and productivity for automated layup systems. Advances in resistance welding for body-in-white Brian Rooks Keywords Welding, Assembly This paper describes the developments in the control of spot welding from Bosch Rexroth with particular reference to body-in-white applications. The Bosch Rexroth MF system uses 1 kHz rather than conventional 50 Hz in the control of spot welding and DC current rather than AC at the weld gun. The several benefits of this arrangement are discussed including lower power and energy losses, lighter cabling and a more compact weld transformer. Also described is a new ultrasonic adaptive control system developed by Bosch Rexroth which enables the growth of the weld nugget to be monitored and recorded for traceability. IMTS leads with technological innovation Dick Bloss Keywords Assembly, Motion control, Systems Innovation drives suppliers even as demand is soft. Modular assembly stations, compact part machining/turning systems, linear motor powered motion and smart end effectors lead the way as the industry drives ahead to provide added benefits to users and system integrators.
Abstracts & keywords
Assembly Automation Volume 23 · Number 2 · 2003 · 120–121
Robotics and assembly automation at TEAM Brian Rooks
The pressure of the adhesive at the joint inlet (gate) was recorded (data logger), and an analysis of this has been used to determine the point when adhesive injection can be arrested and the joint correctly filled.
Keywords Robots, Machine vision, Packaging, Consumer goods, Automotive Ten events made up the new total engineering and manufacturing (TEAM) exhibition, one of which, factory automation (FA) is the focus of this article. A main sponsor of FA was Bosch Rexroth, and a report is given on the state of the company since the acquisition of Rexroth by Bosch. Descriptions are given of a new range of Rexroth Bosch articulated arm robots and an application of the company’s Turboscara robots assembling lambda exhaust control probes. Other exhibitors featured are ABB with a new heavy duty robot, Orwin Automation who demonstrated an automatic bagger, part of its flexible intelligent packaging systems (FIPS) range and DT Assembly and Test who promoted its project and value engineering expertise. Finally, reference is made to GE Panametrics hand held ultrasonic weld testing system.
A novel indexing mechanism for paper cutting machines Rajanish K. Kamat Keywords Paper industry, Cutting This article describes a novel indexing mechanism to improve the productivity of the paper cutting machines. The mechanism is based on dual light sources and an optical detector.
Optimum assembly design utilizing a behavioral modeling concept Y.J. Lin and R. Farahati Keywords Assembly, Modelling, Computer aided design This paper presents a versatile and economical knowledge-based assembly design of blade and shell assemblies by employing behavioral modeling concepts. Behavioral modeling is a new generation CAD concept aimed at achieving ultimately optimum results with the efforts made in the early stage of the product development cycle. As a result, the assembly process of any odd-configured parts such as torque converter blades, can be accurately planned, and made adaptable to all potential in-process alterations due to either changes of components design or that of the assembly kinematics. Optimum assembly design is achieved when the volumetric interference meets a desired value based on an expert’s determination. Experimental verification of the proposed optimum assembly design conducted in Luk, Inc. with two different blades’ assemblies demonstrates satisfactory results.
The automated filling of bonded joints – Part 2 three dimensional joints Ken Young and Ian Pearson
A simulation method and distributed server balancing results of networked industrial robots for automotive welding and assembly lines Paul G. Ranky
Keywords Adhesives, Bonding, Joints
Keywords Automotive, Software
Building automobile bodies from lightweight materials using space-frame construction techniques is increasingly popular because of exhaust emission legislation. One proposed method of achieving this is by using plug and socket joints, which are injected with adhesive after assembly. A method for controlling this process, irrespective of component tolerances, is proposed here. A test rig representing a plug and socket joint was injected with the adhesive and a method for successfully filling the butt-jointed end of the joint found. The addition of a restriction to the joint’s open end gave a method of filling the cavity without creating any air gaps. The use of neoprene O-ring seals for creating the restriction was investigated.
The fundamental purpose of building simulation models is because it is often impossible to experiment with real-world systems. In terms of our networked robots in the automotive welding and assembly lines, simulating possible scenarios, both for design as well as for operation control purposes, is very important for the design team, as well as for management, due to the per-minute-cost of every failed robotic operation. In order to support, both the design as well as the management community of such systems, in this article, we discuss a generic simulation methodology, using the IT-Guru OPNET simulation program, as well as show practical results, that demonstrate the benefits of simulation methods.
121
only concerned themselves with better ways of sharpening quills.
Editorial
Natural evolution No problem? The themes for this issue are ‘‘Adhesives + welding + joining’’ and I am very pleased with the breadth and depth of our contributions. It is evident from these articles and papers, and in particular from the contributions of our Guest Specialist, Dr Robert W. Messler, that we live in interesting times. Every issue that we publish is intended to provide useful information that can be of real commercial and practical benefit to our readers. This issue is no exception; however, the depth of content impresses on me that an additional very important aim of the journal is to make engineers question the way that they are currently doing things. If you have a new problem to solve then the way ahead is clear. You research the possibilities that are open to you; try a few out, and then (hopefully) arrive at a solution. But what should you do if basically you have not got a problem in the first place? Why spend time and money looking into alternatives if existing methods are tried and trusted? Old adages such as ‘‘If it ain’t broke, don’t fix it’’ and ‘‘But we have always done it this way’’, spring to mind. The answer, of course, is that it is only by exploring new methods that we can improve our products and steal a lead on our competitors. The ball point pen would never been invented if the pen manufacturers had
One of the most significant advances over the last few decades has been in the area of new materials. These includes metals and plastics, high strength fibres and those used in nanotechnology. This in turn has opened up a door into a whole new area of design that we earlier had probably never imagined we would want to open, let alone be able to. There is now a very common trend towards making products that mimic nature. Not just because we can, or for academic interest, but because for many applications it provides the best solution. For example, in the field of robotics many mechanisms now use tendon like materials and actuators rather than gears and motors. Robert Messler describes nanotechnology parts that will self-assemble, and there are lots of materials that are designed to selfheal. All attributes without which we would not have a leg to stand on. It is interesting that this trend is in direct opposition to the mechanisation that led to the industrial revolution and mass production. Then manual labour was ruthlessly discarded in favour of cast iron machines that mindlessly produced previous unprecedented quantities of goods. The machines are here to stay, but they are changing. They will never look like us but they are likely to incorporate many of our design features. Clive Loughlin
Call for papers AA 23:4 – Rapid Prototyping + Rapid Manufacturing (Deadline: 8 June 2003) New methods, materials and systems for rapid prototyping and rapid manufacturing. AA 24:1 – Precision Assembly Technologies for Mini and Micro Products (Deadline: 3 October 2003) AA 24:2 – Mechatronics (Deadline: 14 December 2003) The benefits of combining mechanical and electronic technologies. AA 24:3 – Internet + System Design + Modular Assembly Systems (Deadline: 25 March 2004) Internet enabled factories. Recent developments in modular assembly systems and planning and simulation software. AA 24:4 – Design for Assembly and Disassembly (Deadline: 8 June 2004) Assembly Automation Volume 23 . Number 2 . 2003 . p. 122 # MCB UP Limited . ISSN 0144-5154
122
Viewpoint Joining research and education: where must it go? How must it change? Robert W. Messler
The author Robert W. Messler is Professor and Director of Materials Joining in the Materials Science and Engineering Department at Rensselaer Polytechnic Institute, New York, USA, and Fellow of the American Welding Society and ASM International. He has written numerous articles and a couple of Viewpoint columns for the JAA in the past. Keywords Joining, New materials, MEMS, Adhesives, Nano technology Electronic access The research register for this journal is available at http://www.emeraldinsight.com/researchregister The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0144-5154.htm
Assembly Automation Volume 23 · Number 2 · 2003 · pp. 123– 124 q MCB UP Limited · ISSN 0144-5154 DOI 10.1108/01445150310471356
In the US and probably most other places in the world, only three things are certain: death, taxes, and change. For those involved with technology, not only the certainty but also the rapidity of change is most obvious – and, as long as we are not too old or too infirmed or too far indebt, death and taxes seem secondary! Many of us who are engineers learned to calculate on a slide rule and to produce engineering drawings with a T-square, triangle, and compass. So, the rate of change is painfully obvious to us. But, it is the change that has taken place – and continues to take place – in materials that impacts us the most; as consumers and as technologists. We are in the midst of not one, but several materials’ revolutions; the “Electronic Materials Revolution”, the “Composites Revolution”, the “Nanomaterials Revolution”, to name three. To those responsible for manufacturing the products the people of the world use, changes in materials mean changes in processing; with some of the most dramatic changes in processing involving processes used in assembly, i.e. joining processes. But, as has been expressed in Viewpoint columns of the past (e.g. see Vol. 20, No. 2), changes in joining have not kept pace with the changes in materials. This has been the case in the past and in the present. For the most part, automobiles are still assembled using the processes used for over six decades; i.e. by resistance spot welding. Commercial airliners are still assembled using the processes used since World War I; i.e. by riveting. Microelectronic devices are still assembled using the same processes since the technology first appeared; i.e. by soldering. Yet, in each case, the materials have changed or are trying to change. Aluminum and reinforced plastics are trying to replace steel in automobiles. Polymer-matrix composites are trying to replace aluminum in airliners. And, reductions in scale, together with everincreasing concern for and sensitivity to the environment are challenging soldering for electronic interconnect. So, where must joining go and how must it change? Or, more importantly, because processes cannot change until the people that create them and use them change, how must the research to advance joining and the education to prepare researchers and practitioners of joining change and where
123
Joining research and education
Assembly Automation
Robert W. Messler
Volume 23 · Number 2 · 2003 · 123–124
must each go? To quote Hamlet: That is the question! With the declining attraction of skilled trades like welding to young people threatening the future supply of skilled welders, the answer is automation. Just as computer – numerical controlled machining changed the needed skills in modern versus traditional machinists, automation must – and will – change the needed skills of welding practitioners; making them welding operators and technicians as opposed to welders. But, the possibilities and promises offered by automation can only occur with greatly expanded, enhanced, and accelerated research in automation; including active, embedded sensing, real-time monitoring, and intelligent adaptive control. And, control must become less complicated and more direct, not more complex and more indirect. Only what needs to be controlled should be controlled, and only desired outcomes, not every input, need to be sensed and monitored. Why sense and monitor the voltage, current, and travel speed during welding, if what is desired as an outcome is the nature and quality of the weld microstructure and associated properties? Why attempt to control all of the parameters that are involved in welding even though we do not understand all of the complex, inextricably linked mechanics, thermodynamics, kinetics, physics, and electronics to do so with 100 per cent certainty? With the changes occurring in the mix and the appearance of new materials in the
production of new products, new approaches to joining will be inevitable. Joining will, of necessity, increasingly become an integral part of integrated material synthesis and device or product synthesis. Joining would not occur after components are fabricated, more and more it will occur as components are created. MEMS – and then NEMS – will selfassemble, not be assembled. Interconnections in micro- and eventually nano-electronic devices will self form, not be formed. As a result, research is needed to explore, identify, develop, and optimize new methods for joining, and the education of engineers and technicians and other technologists must be expanded to include knowledge of joining, not just welding or riveting or adhesive bonding. The above are but two looming changes to which our research and education must not only respond but also presume will occur – if they have not already occurred. It is time to anticipate change, not react to it! It has been said that research is to education as sex is to sin; without the one, you cannot have the other. Research and education are inextricably linked. One cannot perform meaningful research without having been well educated in a fundamental body of knowledge. But, one cannot be well educated without research advancing the state of knowledge. Likewise, truly transforming breakthroughs in materials are only of value if there are equally transforming breakthroughs in joining. Will we be ready?
124
Company news
Assembly Automation Volume 23 . Number 2 . 2003 . 125–129
Company news SICK, OMRON, Rockwell and ODVA specify DeviceNet safety: SICK, Rockwell Automation, OMRON and ODVA collaborate on development of open, interoperable safety networking standard Keywords DeviceNet, Safety
Three of the industry’s leading global automation and machine safety companies – Rockwell Automation, OMRON Corporation and SICK AG – are collaborating on the development of an open protocol for safety communications. Based on the existing common, standard protocol found in DeviceNet, and Ethernet/IP, the newly defined safety protocol extensions for the safety protocol have received concept ¨ V Rheinland which represents approval by TU a major milestone in meeting the requirements of safety technology in an open, network environment. Based on customer requirements, the three companies, all members of the Open DeviceNet Vendor Association (ODVA), recognised the need for an open, safety protocol that was compatible with existing networks in the industry. Consequently, they began collaborating on a safety protocol to meet the emerging customer requirements. ODVA has decided having these companies continue this collaboration to implement the safety protocol. Consisting of an extension to the existing standard protocol, the safety protocol will allow both standard and safety devices to operate on the same network. It will also allow safety devices to seamlessly communicate across standard DeviceNet and EtherNet/IP networks, to other safety devices without any additional programming. This unique feature, which is not available today for safety communications, will be realised without the need for expensive safety specific hardware, such as gateways and bridges. Additional benefits of this new approach include network design flexibility, easier training and maintenance, and a common view of the both standard and safety networked devices. The safety protocol’s first implementation will be over DeviceNet, and will provide
fail-safe communication between nodes such as safety input/output blocks, safety interlock switches, safety light curtains and safety PLCs. As an open standard, the safety protocol is designed for usage in safety applications up to safety integrity level (SIL) 3 according to IEC 61508 standard. Use of the safety protocol will provide customers with a breadth of products from multiple vendors that work together. In addition, customers will have tremendous flexibility in network architecture design and installation cost savings. ¨ V is the first Concept approval from TU milestone in the development of the safety protocol. When completed and certified by ¨ V, this open specification for the safety TU protocol will be owned and managed by ODVA. The companies’ next step will be submission for approval of a system requirements specification, for implementing the safety protocol on DeviceNet. The companies expect to introduce the first DeviceNet Safety solutions in 2004. For more information contact: Ann White or Andrea Hornby, Erwin Sick Ltd, Waldkirch House, 39 Hedley Road, ST ALBANS, Hertfordshire AL1 5BN, UK. Tel: 01727 831121; Fax: 01727 856767; E-mail:
[email protected]; Web site: www.sick. co.uk
Dassault Systemes acquires KTI Keyword KBE
Dassault Systemes has announced the acquisition of Knowledge Technologies International (KTI) in an all-cash transaction. Through its consulting expertise and ICAD software, KTI has become the recognized leader in knowledge-based engineering (KBE) solutions for the capture and automation of proprietary customer design and manufacturing processes, particularly in the aerospace and automotive industries. Examples of KBE use include the automatic design of the thousands of standard parts in aircraft wings or electrical system components in automobiles. The acquisition of KTI complements Dassault Systemes’ existing V5 Knowledgeware solutions and places the company in a leadership position across the full range of KBE practices. KBE practices are divided into three domains that include:
125
Company news
Assembly Automation Volume 23 . Number 2 . 2003 . 125–129
generic applications, industry specific solutions, and customer proprietary process automation. KTI is the leader in customer proprietary process automation while Dassault Systemes is the leader in generic applications and industry specific solutions. Integrating the development, capture and reuse of a customer’s know how is an essential ingredient for successful PLM implementation. Beyond the ‘‘what’’ of products (the components), knowledge-based solutions capture and manage the ‘‘why’’ and ‘‘how’’ (the collective knowledge). Combining KTI’s KBE consulting competence with Dassault Systemes industry-processes expertise creates an unrivaled service offering in the PLM industry that makes KBE a reality for all customers. ‘‘This acquisition will deliver tremendous value to our existing ICAD customers’’, said Dr Prasanna Katragadda, Vice President, Research and Development, KTI. ‘‘It immediately secures their investments and provides a roadmap to further expand the management of their KBE activities beyond the design and manufacturing stages and into the rest of their product’s lifecycle management.’’ ‘‘Over the past 7 years, Airbus, with the support of KTI, has secured significant returns through KBE, enabling automation, integration and innovation of key processes’’, said S.E. Allwright, Head of KBE, Airbus. ‘‘With the acquisition of KTI by Dassault Systemes, Airbus expects to immediately benefit from synergies in the development of knowledge-based solutions and provision of services. This will enable Airbus to accelerate and broaden our implementation of KBE across our complete set of product lifecycle processes.’’ For more information visit our Web site: www.ktiworld.com
fieldbus technology during the previous year, with fieldbus device registrations and foundation membership also on the rise. The market study, which surveyed the Fieldbus Foundation’s global membership of process and manufacturing automation companies, showed that Foundation fieldbus is the ‘‘technology-of-choice’’ for a growing number of major end users. More than 4,000 Foundation fieldbus systems have been installed around the world, and over 205,000 fieldbus devices are now in service. Fieldbus installations were reported in diverse industries, including petrochemical, refining, chemical, oil and gas, metals/mining, water and waste, pulp and paper, pharmaceuticals, utilities and others. Fieldbus Foundation President and CEO, Richard J. Timoney said the study findings confirm that Foundation technology has gained market share and is now regarded as an industry-standard solution for mission-critical plant automation applications. ‘‘Our members have reported increased business activity involving Foundation fieldbus, with many leading end users now specifying fieldbus technology for their new control system projects’’, said Timoney. ‘‘The latest list of Foundation fieldbus installations around the world reads like a ‘who’s who’ of the industrial community.’’
Fieldbus Foundation growth indicated by recent market study: Installations, registered devices and membership on the rise Keyword Fieldbus
The Fieldbus FoundationTM has announced that a recent market study indicated a significant increase in end user installations of its open, non-proprietary FoundationTM
Mecelec purchase set to strengthen JFL Automation’s reputation and creates a strong platform for growing success Keyword Automation
The JFL Manufacturing Group is announcing the purchase of Mecelec – a long established name in the test and automation field. The company is linking up with JFL Automation to create JFL Mecelec in a move that is set to provide a broader service for both companies’ range of customers – from automotive and aerospace manufacturing to pharmaceutical and general engineering (Plate 1). It is a development which is combining the experience and engineering track-record of both organisations into a single, highly professional operation that will deliver at every level and, according to Ray Golden, Managing Director of JFL Manufacturing, is extremely logical – ‘‘For almost three
126
Company news
Assembly Automation Volume 23 . Number 2 . 2003 . 125–129
Plate 1 Confirming the purchase of Mecelec by the JFL Manufacturing Group – Managing Director Ray Golden with Mecelec Directors Bob Holmes (left) and John Strutt (right)
decades Mecelec has focused on the development and installation of automation equipment and has gained an enviable reputation, in particular, for its assembly and test station capability,’’ he says. ‘‘Because this mirrors the growth and development within JFL Automation, there is a synergy between the two companies which is so close that the resultant service is set to provide customers with access to a level of understanding and commitment which I believe is unmatched in its sector of industry.’’ For more information contact: William Bourn, JFL Mecelec Automation and Test Ltd, Llanthony Road, Gloucester, GL2 5QT. Tel: (01452) 413531; Fax: (01452) 307580; E-mail:
[email protected]; Web site: www.mecelec.co.uk
Huge future for low-cost miniature components Keywords Electroforming, Assembly
Designers and OEMs worldwide are set to benefit from the significant technological advancements recently achieved by photo-chemical machining specialist, Tecan. The company has breached traditional manufacturing obstacles to deliver low-cost micro metal parts, and larger parts with ultra-fine features, for a vast new applications
arena covering electronics, optical, medical and aerospace (Plate 2). The inherent advantages of the technology are expected to be exploited by a wide range of industries, for the demanding next-generation micro applications, where products will be made significantly smaller and finer, repeatedly, at lower cost and with fewer production processes. Typical micro-applications include sensors, actuators, hearing aids, medical devices, optical instruments, micro-lenses, meshes, masks, displays and micro-fluidic devices. Micro structures can be manufactured to extremely small scale, with features, such as apertures, fluid channels or raised lands, down to one or two microns with tolerances at sub-micron levels. They will also be employed in micro electro-mechanical systems (MEMS) and micro optical electro-mechanical structures (MOEMS). Similarly, the company can produce larger parts, up to 300 300 mm, with equally fine resolutions. Recently opened by His Royal Highness The Prince of Wales, the company’s dedicated new £2 million, 8,000 sq/ft facility, houses a state-of-the-art Class 1000 clean room, within which are a number of Class 100 areas. These highly clean areas ensure the provision of particle-free environments where components with sub-micron features can be repeatedly and consistently fabricated. The company has developed world-leading photo-electroforming production and in-house plating techniques to make such cost-effective accuracy possible. Previously in the manufacturing world, MEMS have predominantly been manufactured in silicon, using expensive technology from the semiconductor industry. Tecan currently produces such parts in nickel, at significantly lower cost, and has plans to move into other pure conductive materials such as copper and gold in the near future. Such parts offer significant advantages over brittle and expensive silicon options, such as mechanical strength, electrical conductivity, corrosion resistance and other unique benefits. The company describes the new technology as the hybrid application based on three established manufacturing technologies – silicon semiconductor, high-volume audio CD, and microembossing tools such as those used in micro-lens array and hologram manufacture. Ongoing market feedback from existing customer joint-ventures combined
127
Company news
Assembly Automation Volume 23 . Number 2 . 2003 . 125–129
Plate 2 Tecan has developed techniques to produce micro-miniature metal parts for next-generation applications where products can be repeatedly made significantly smaller and finer, at lower cost and with fewer production processes
Exceptional tolerance and accuracy are assured, across the maximum surface area of 300 300 mm, far greater than most existing silicon capabilities. Raised areas can be produced with a maximum aspect ratio of 5:1 and potentially greater. Tracks and channels can be fabricated as narrow as 2 micron wide, 4 micron pitch, with smooth walls and sub-micron tolerances. Surface smoothness is also exceptionally accurate, being sub-wavelength, at 600 nm across surface. (Lambda/4). For more information contact: Jill Steer, Tecan Components Limited, Tecan Way, Granby Ind. Est., Weymouth, Dorset, DT4 9TU. Tel: 01305 765432; Fax: 01305 780194; E-mail:
[email protected]; Web site: www.tecan.co.uk
Key2Automation Web site is invaluable tool in provision of solutions-led automation information Keywords Robotics, Automation, ABB
with 30 year’s experience in photo-chemical machining and photo-electroforming were also crucial elements in the development of the new processes. ‘‘One of the greatest benefits to our customers, is an increasing appreciation of photo-electroforming and the opportunities it offers their designers as they liaise with our own’’, said Noel Cherowbrier, Tecan’s Sales and Marketing Director. ‘‘As these relationships develop, designers increasingly realise that we can make micro-metal components with more exacting designs and tighter tolerances than they ever realised were achievable. Unique and extremely sophisticated three-dimensional parts are now being conceived as OEM engineers and designers explore the wider technology envelope now available to them – through the traditional ‘X/Y ’ axis, and now beyond – through exploitation of the ‘Z ’ axis too.’’
Key2Automation believes that its newly launched industrial automation Web site, www.key2automation.com, is set to become one of the most established ‘‘community’’ Web sites in industry by offering solutions-led information on robot-based automation (Plate 3). This industry initiative, which has been in development for 2 years with the backing of ABB, was launched this year and is already proving to be a popular business tool for those looking to streamline manufacturing operations through automation. It is both easily accessible and informative and is proving popular with systems integrators and product developers as it offers real solutions while also providing a discussion forum for open debate with others in the industry. Tinne Dierckx, Sales and Support Manager, Key2Automation, says: ‘‘We are fully expecting this site to become the marketplace for automation. Our aim is for system integrators, machine builders and manufacturing industry end-users to automatically turn to the site for detailed and quality information on solutions, products and services. We are not interested in developing this site purely as a directory of company names – it is far more substantial than that. First and foremost it is
128
Company news
Assembly Automation Volume 23 . Number 2 . 2003 . 125–129
Plate 3 Key2Automation Web site offers solutions-led information on robot-based automation
a solutions-led business tool which will generate an independent community that helps yield more and better business for all participants’’. David Marshall, Business Manager, Robotics, ABB, comments: ‘‘This site offers a real two-way communication tool for those looking for automated production and manufacturing solutions. For example, the ‘solutions’ page offers in-depth case studies on a myriad of applications and industry areas. It covers topics such as how, for one
company, rod line automation has improved output by 67 per cent with a saving of $13 m, or how a high-speed sorting system has been developed for a postal service which allows every hour over 4,000 tote bins to be unloaded automatically’’. For more information contact: David Marshall, ABB, Auriga House, Precedent Drive, Rooksley, Milton Keynes, MK13 8PQ, UK. Tel: +44 (0) 1908 350300; Fax: +44 (0) 1908 350301; E-mail: david.
[email protected] 129
Keynote Joining comes of age: from pragmatic process to enabling technology Robert W. Messler
The author Robert W. Messler is a Professor of Materials Science and Engineering/Director of Materials Joining at the Rensselaer Polytechnic Institute, Troy, NY, USA. Keywords Joints, Biotechnology, Information technology, Microelectronics Abstract From when we, as humans, first lashed a pointed stone to a split straight stick to make a more effective spear for hunting to now when we fasten and bond ablative ceramic tiles to the frail metal skin of the Space Shuttle to allow safe re-entry from manned excursions into space, joining has been a pragmatic, albeit critically important, fabrication process. As we move beyond the Industrial Age to the ages of Information Technology, Nanotechnology, and Biotechnology, joining must move from a secondary process for manufacturing objects or articles from pre-synthesized and pre-shaped materials to a primary process for combining materials into fundamental structures as these structures and even materials are being synthesized; where the boundary between the materials and the structure becomes blurred. This paper attempts to catch a glimpse of the future where joining comes of age to become an enabling technology practiced as much or more by technicians or physicians than as a trade practiced by helmeted welders or hard-hatted riveters. Electronic access The research register for this journal is available at http://www.emeraldinsight.com/researchregister The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0144-5154.htm Assembly Automation Volume 23 · Number 2 · 2003 · pp. 130– 143 q MCB UP Limited · ISSN 0144-5154 DOI 10.1108/01445150310471365
Where has joining been? From the dawn of humankind – perhaps even before, if Figure 1 represents anything more than artistic liberty – joining has been a pragmatic, albeit critically important, fabrication process for manufacturing physical objects or articles or edifices from materials. At first, the materials being joined were those found in their natural state, with little or no alteration; stones, logs and sticks, nuggets of gold or silver or copper. Perhaps the earliest examples of joining involved simply (but profoundly!) attaching a sharp stone to a split straight stick to produce a spear to make hunting more efficient, more productive, and safer for the hunter, or an ax for cutting down trees and hewing timbers to build a shelter or for hollowing out a log to create a dugout canoe. As knowledge of materials and processing evolved, more effective and sophisticated tools emerged; first soft shale, then hard flint, malleable bronze, strong iron made possible by first cleaving, chipping, smelting, refining using first wedging, lashing, soldering, and finally welding. As materials and processing (including joining) advanced, products advanced; first simple tools, then imposing structures, then imposing weapons. By medieval times, materials of construction included not only stone and wood joined using mortar and pegs, and even iron nails, but also glass and metals and cement joined using fusing and welding and bonding. Great cathedrals with magnificent leaded stained-glass windows, imposing siege machines, and treasure-hunting sailing ships reflected advances in materials and processing, including joining. More and more, natural materials were altered by humans as the making of glass and smelting of ore and alloying of metals provided improved Technical input and proof-reading by Professor James Napolitano (in the area of information technology), Professor Shyam Murarka (in the area of microelectronics and photonics), Professor Pulickel Ajayan (in the area of nanotechnology), Dr Tobias Winther (in the area of microsystems), and Professors Robert Spilker and Jonathan Dordick and Rena Bizios (in the areas of biotechnology and tissue engineering) is all gratefully acknowledged. Without the kind assistance of these colleagues at Rensselaer Polytechnic Institute expert in these new technologies, it would have been impossible to prepare this paper.
130
Joining comes of age: from pragmatic process to enabling technology
Assembly Automation
Robert W. Messler
Volume 23 · Number 2 · 2003 · 130–143
Figure 1 From the dawn of humankind, and maybe even before, joining has been a pragmatic, albeit critically important, fabrication process for manufacturing physical objects or articles or edifices from materials
on the threshold of understanding how we ourselves are made by decoding the human genome. Today, most of the materials we use are synthesized; with few having anything but the most primitive relatives in nature. From plastics to superplastic metals, extrinsic semiconductors to high-temperature superconductors, optical fibers to carbon nanotubes and “nanowires”, more and more of the materials we rely on for advancing designs have been conceived and produced by humans; almost to the exclusion of natural materials (Easterling, 1990; Forester, 1988). If this is where we have been, where are we going? And, where must joining go to let us get there?
Where is joining going? properties, and new and more intense sources of energy allowed more sophisticated and effective and productive metallurgy and joining; first with hot-set rivets, then by low-temperature soldering and brazing, followed by high-temperature fusion welding, first with gas flames, then with electric arcs and electrical resistance, and finally with high-energy beams of electrons or photons. As unfortunate a statement as it makes about human beings, one of the greatest driving forces for advancing materials and processing (as well as other technologies), including joining, has been some seemingly primal need to make war. From clubs to spears to long-bows to cross-bows, from cannons and muskets to automatic rifles and lasers, and from dugout canoes to military galley ships to Civil War Monitors and Cold War Tridents, and from World War II Spitfires and Gulf War Stealth Bombers to yet-to-bedeveloped Strategic Defense Initiatives and still only imagined Star Wars Starships, the need for more formidable weapons gave rise to more impressive materials and processing, and vice versa. The other, more enlightened and enlightening driving force is another seemingly primal need in humans to explore the unknown, to understand: Who we are?, How we got here?, and Where we are going? This inexplicable urge has led to better ways to move on or under the seas, ways to fly over the ground and beyond the Earth into space, and ways to call on a friend by traveling over interstates or chat with cyber-friends by traveling over the Internet. And now, we are
Joining has been as responsible as any process (and more responsible than most processes) for getting us to where we are as a species. The ability to join materials and devices, parts, or structural elements into assemblies has enabled, if not driven, the advancement of civilizations and the growth of economies. Welding can be attributed (and has been cited by the American Welding Society in its promotional literature) as being either directly or indirectly responsible for the rise of 60 percent of the gross domestic product (GDP) of industrialized countries (World Almanac, 2001); directly when it is used to actually create structures that themselves add to wealth, and indirectly when the structures created by welding are used to generate wealth (e.g. farm tractors and mining equipment). And welding is only one, albeit an important, option for joining. Figure 2 shows a pie chart of the estimated percentages of all joining accomplished by the various available options; mechanical joining (including fastening and integral attachment), welding, brazing, soldering, adhesive bonding, and various hybrids and allied processes of these (Messler, 1993, 1999). As surely as joining has brought us to where we are, it will bring us to where we want to go; wherever that turns out to be. Buildings and bridges will continue to be needed, and they will be built by joining materials like brick and stone and concrete and steel, and probably even composites. Automobiles (perhaps as hybrid- or all-electric vehicles) and aircraft (perhaps as supersonic or even hypersonic
131
Joining comes of age: from pragmatic process to enabling technology
Assembly Automation
Robert W. Messler
Volume 23 · Number 2 · 2003 · 130–143
Figure 2 A pie chart showing the estimated percentages (based on dollar or dollar-equivalent value) of all joining accomplished by the various options; mechanical joining (including fastening and integral attachment), welding, brazing, soldering, adhesive bonding, and various hybrids of these
vehicles) will continue to be needed, and they will be built by joining materials like aluminum and titanium and superalloys and intermetallics and composites with polymeric and metallic, and perhaps even ceramic, matrices. But, beyond these and many other survivors of our past (e.g. household appliances, office equipment, recreational boats and jet-skies, sports equipment, etc.), albeit in evolved forms, where will the world be going? Most see the immediate future, and even the future beyond that, as dominated by three technologies: Information Technology, Nanotechnology, and Biotechnology. Bill Joy, co-founder of Sun Microsystems and a computer scientist who, as much as anyone, revolutionized computer programming to the point that some programs are capable of generating other programs, writes in his thought-provoking article “Why the Future Doesn’t Need Us” ( Joy, 1999) that these three as well as Robotics[1] will dominate and shape our future; with the near-equal potential for either unprecedented progress or the annihilation of our species. And let there be no doubt, joining will be at the heart of each of these technologies. But, it is not difficult to see – or at least imagine – that how joining will have to be practiced in the future for these technologies will be dramatically different from how it has been practiced until now. In the past, joining was exclusively a “secondary” process; not in the sense of being less important than other processes, necessarily, but in the sense that it was often
one of the last processes to be performed during fabrication, manufacturing, and construction. Parts, pieces, and devices made by primary processes like casting, molding, deformation processing[2], or special processes (like lay-up, filament winding, braiding, weaving, etc. for composites) and by another important secondary process, machining, are almost always joined as a near-final, if not final, step (Lindberg, 1990). As a result, joining, while obvious with any foresight for designs requiring assembly of any kind (i.e. other than one-piece designs!), is often an afterthought that creates as many problems as it benefits (Messler, 1997). Joining has been a pragmatic process, sometimes a “necessary evil”, to get things done. In the future, at least for many applications in Information Technology, Nanotechnology, and Biotechnology, things will be different. We have already had a glimpse of the differences when we look at modern microelectronics, of what will become commonplace elsewhere. With the advent of solid-state semiconductor-based electronic devices like diodes and transistors, p- and n-type extrinsic semiconductor materials were joined with one another to create n-p, n-p-n, and p-n-p junctions that allow such devices to work. But, the “process” of joining was transparent. Pieces of pre-fabricated (or, more correctly, pre-synthesized) n-type material were not joined to pieces of pre-synthesized p-type material using joining as a secondary process. Instead, joining was (and is!) accomplished as an integral aspect of the process of synthesizing the n- or p-type materials themselves; with the “joint” or, more correctly, interface arising naturally (or automatically) as a by-product (actually, a principal goal!) of the material synthesis to lead to synthesis of the device. Here, in this application, joining is a primary process; as much a part of creating the materials comprising the device or structure as of creating the device or structure itself. In such applications, the boundary between the materials and the structure has become so blurred that the actual phenomenon – no less process – of joining is not even recognized or, often, acknowledged! What lies ahead – just ahead in many instances – is an entirely different way of thinking about and accomplishing joining. A new paradigm. Joining will “come of age” to
132
Joining comes of age: from pragmatic process to enabling technology
Assembly Automation
Robert W. Messler
Volume 23 · Number 2 · 2003 · 130–143
(2) altering the architecture of the storage device in some clever way (often using programming) to take advantage of shared memory, for example.
grow from a pragmatic process to an enabling technology. Let us look briefly at each of the three technologies that will not just drive but demand this change. Joining in information technology The Information Technology Association of America (ITAA) gives the following definition: “Information Technology is one of America’s fastest growing industries, encompassing computer, software, telecommunications products and services, Internet and on-line services, systems integration and professional services” (Hughes, 2001). The physical parts of this are, of course, the computing, telecommunications, and Internet hardware products. The physical foundation on which Information Technology is built is electromagnetically-based devices[3] and circuitry for: inputting information in the form of binary bits and bytes; operating on such electromagnetic data to add, subtract, multiply, divide, differentiate, or integrate; moving that data within the device and beyond the device; and storing that data in memory until it is no longer needed. Until now, electrons in the form of electrical currents and voltages have been used in the so-called microelectronic devices, or microelectronics for short. However, the future is almost surely going to include, if not be dominated by, optical versus electronic realizations using the technologies of optoelectronics and photonics. Certainly, optical switching is one key, and there are many others; many of which will depend upon joining functioning, no less properly. Advances in Information Technology rely on: (1) increasing the amount of electromagnetically-based data that can be handled (i.e. inputted, operated on, stored, and/or moved), or the sheer volume of data, and (2) increasing the speed with which the aforementioned events can be accomplished, or processing throughput. In microelectronic devices (typified by computers), the two basic ways in which data storage can be increased are: (1) decreasing the space needed for storing the basic elements of data (i.e. binary bits and bytes), and
The three ways in which the speed with which data can be handled are by: (1) decreasing the physical distance that the data has to travel (especially, but not only, during data processing), (2) increasing the speed with which the basic electromagnetic data-carrier (here, electrons) can move in the materials in which they must move[4], and (3) altering the architecture of the system to allow parallel operations to be performed, for example. While new semiconductor materials are being sought to increase the speed with which electrons can move in materials (e.g. III-V GaAs, where electrons move six times faster than in Si, and “strained silicon” created by depositing a very thin pure Si layer on top of a graded buffer layer of SiGe, with a larger lattice parameter due to the larger Ge atoms, on top of a conventional bulk Si substrate, where electron mobility approaches that found in III-V semiconductors, but with the ease of processing of Si), and new system architectures (as well as more efficient algorithms) offer gains, the greatest advances have come from decreasing the scale of devices more and more. Figure 3 shows how the capability (i.e. processing speed and capacity) of computers have increased over the period from 1970 to 2000+ as reflected in the increased number of transistors able to be employed in circuits (http://www.intel.com/research/silicon/ mooreslaw/html) following what has become known as “Moore’s Law (Moore, 1965). Figure 4 shows the trends in microelectronic packaging over the same period of time, with increased processing capability coming about as the direct result of reduced feature size (e.g. line width) (Tummala and Rymaszewski, 1989). As the scale of devices, in terms of the channel length (under the gate of the MOSFET) decreases, the challenge of producing joints between materials in devices (at the chip-level) and between devices (at the board-level) increases dramatically. But, there is a physical limit estimated to be at approximately 30 nm. This limit has been tentatively arrived at independently by both
133
Joining comes of age: from pragmatic process to enabling technology
Assembly Automation
Robert W. Messler
Volume 23 · Number 2 · 2003 · 130–143
Figure 3 A plot showing how a predictable increase in computing capability comes about as the direct result of an increase in the number of transitors employed, which forces down-scaling of feature sizes
Figure 4 A plot showing trends in electronic packaging over the period of time from 1970 to 2000+ as a result of decreased feature size, given here as “line width”
Carver Mead and James Miendl (Meindl, 1987) who examined the three most important limits imposed by thermodynamics, quantum mechanics, and electromagnetic theory. Some joining challenges facing the microelectronics community include the following. (1) To make electrical connections (or interconnects) between present- (and past-) generation chips to circuit boards, and, to a lesser extent, on chips, soldering has been the process of choice; with the 63 wt percent Sn-37 wt percent Pb eutectic alloy being the “workhorse” solder of choice. Two pressures are threatening this predomination; one environmental (and perhaps without much real merit!), and the other technical (and very real!). 134
From an environmental standpoint, there has been pressure to remove lead (Pb) from all solders; those used in microelectronic interconnects as well as those used in residential, commercial, and municipal water-supply systems[5]. This has led to a concerted effort, around the world, to find alternative, Pb-free solder compositions that offer good wetting characteristics, reasonable strength (including resistance to the thermomechanical fatigue that limits the life of most soldered microelectronic joints), and melting characteristics (most notably the 1838C melting point) of the Sn-Pb eutectic (NCMS, 1997). Finding viable alternatives is made difficult by the fact that many of the relatively small number of metallic elements that can produce low-melting alloys are also toxic (e.g. Cd, Sb, Hg, Ga, and, perhaps, Bi; leaving only In, Sn, and Zn). Beyond this is the problem that alloys which offer comparable (or superior) wetting characteristics and strength have melting points that exceed the temperaturetolerance of most circuit boards (most notably, FR-4) – which, in fact, were developed with eutectic Sn-Pb in mind. Hence, the search continues. Technically, there is a limit to how small the dimensions of solder joints can get and still have the soldering process work. Key to the formation of metallurgicallysound solder joints is wetting of the substrates by the molten solder alloy and flow of the solder to individual joints (as well as its ability to self-form and shape the joints through the molten alloy’s surface tension); both of which depend on capillary attraction, which arises from surface energy or tension. As dimensions get smaller, a point is reached (and may have already been reached at the level between 0.5 and 0.3 microns) at which the molten solder cannot differentiate one joint from another one adjacent to it. When this point is reached, “bridging” occurs between the joints and electrical shorts result; making the device ineffective/ inoperative. So, we may be looking at the beginning of the end for the predominance of soldered joints for on-chip and even chip-to-board interconnects.
Joining comes of age: from pragmatic process to enabling technology
Assembly Automation
Robert W. Messler
Volume 23 · Number 2 · 2003 · 130–143
(2) There may be an opportunity for intrinsically- (as opposed to metal-filled) conductive polymer adhesives to take over for the solders in certain applications, although the basis for sound adhesive bond formation is also wetting of the substrate by the adhesive and formation of individual joints by surface tension. However, in this case, flow of the adhesive by capillary attraction is not really involved, so the problem of “bridging” might be circumvented to a slightly smaller scale. As one example, small aromatic organic molecules, such as polyphenylene, have demonstrated the ability to conduct electricity at a current density approximately one million times as great as a 1 mm-diameter pure copper wire (Figure 5). By using such small aromatic organic molecules, or carbon nanotubes (that have shown high conductivity, if not superconductivity), or even biomolecules or semiconductor nanowires to enhance or supplant conventional bulk silicon, the potential to put as many as a trillion electronic switches in a square centimeter exists. This would permit, for example, a computing system that contains approximately ten billion switches to be fabricated on the top of a grain of salt. Equally important to intrinsic conductivity is the ability of such organic, polymeric molecules to switch small electric currents. Fortuitously, this ability has also been found to exist in certain molecules (Kwok and Ellenbogen, 2002). To understand what a giant step this would be beyond today’s electronic computers, one needs to appreciate that a conventional commercial microcomputer chip still contains only about 10-50 million switches in a much larger area, the size of a postage stamp. Beyond this scaling effect, however, it is envisioned that this million-fold increase in the density of computation using molecules will be changed by the fact that
computation will literally become a property of matter (Kwok and Ellenbogen, 2002). So, even the long-sought intrinsicallyelectrically conductive polymeric adhesive may not fill the void for solders at very small scales, but evolution to “moletronics” could revolutionize computing – and place new demands on joining. (3) There are many on-chip applications where it would be beneficial to be able to produce “self-forming bonds”. Getting pure copper (Cu) to adhere to the chemically-stable SiO2 substrate is very difficult, if not impossible. One attractive possibility is to employ Cu alloys with small amounts of Al or Mg as solute. When these Cu-Al or Cu-Mg alloys are sputtered onto the SiO2 substrate and heat treated, a bond-forming Al2O3 or MgO layer creates the needed joint. The joint forms itself by diffusion of the Al or Mg to the SiO2 to react, as part of the device synthesis, as opposed to as a secondary process. Early successes with this approach is shown in Figures 6 and 7 (Kirchner, 1996; Suwwan de Felipe, 1998). So, approaches for producing self-forming joints in situ, as part of device synthesis, are a worthwhile and achievable goal. (4) In devices where an SiO2 substrate is coated with a low electrical permeability polymer (e.g. k ¼ 2 or less, where k ¼ e /e o) before a metallization layer is deposited, there is a need to achieve Figure 6 A plot of the XPS-generated composition profile for 100 A˚ Al on SiO2 (as deposited) shows Al2O3 formation at the Al-SiO2 interface leading to a self-forming reaction barrier at the interface
Figure 5 Conductances of molecular wires expressed in appropriate nanoscale units compared with the conductance of a macroscopic copper wire expressed in the same units
135
Joining comes of age: from pragmatic process to enabling technology
Assembly Automation
Robert W. Messler
Volume 23 · Number 2 · 2003 · 130–143
Figure 7 High resolution TEM micrographs of the SiO2-Cu interface of capacitors A1 (a) and A2 (b) in which a thin layer of Al2O3 (a) and MgO (b) can be seen to have self-formed during annealing at 1508C under a bias of several millivolts
(5) A theoretical limit (or, at least, a limit beyond what is practical) of 0.03 microns or micrometers (or 30 nm) in the scale of features on chips will force evolution of 2-D to 3-D circuits. Even now, in fact, at the current sizes of wires and transistors in integrated circuits, 3-D circuits have much to offer. Such 3-D circuits will require that intrinsically and extrinsically semiconductive Si will have to be joined to insulative SiO2 or to a high dielectric constant polymer. So, the challenges posed in #3 and #4 above will again come into play; but with even greater challenges for even greater potential payoff. (Obviously, 3-D circuits pose entirely new and higher-level problems, as access to interior joints is impossible as an afterthought. Joining can only be accomplished as in integral step in the synthesis of materials and devices.)
good adhesion (ideally, through some “self-forming” process), but that is also “self-limiting” to prevent run-away reaction and unwanted and intolerable interaction (Murarka, 2002). The possible use of organics, organometallics, and self-assembling molecules or monolayers arise when the post-joining process and/or use temperatures are low. So, techniques for creating self-forming joints that are also self-limiting are further goals for joining in these applications.
Ultimately, the speed with which data will want/need to be processed for IT is at the speed of light, and the ability to store and otherwise handle data will want to approach having no limit! Far fetched? Not really! The answer is photons; and the technology is photonics. As often defined in the periodical Photonics, photonics is “The technology of generating and harnessing light and other forms of radiant energy whose quantum unit is the photon. The range of applications extends from energy generation to detection to communication and information processing”. By relying on photons as the data-carriers, the two limitations of electrons as data-carriers in materials are overcome. First, photons move at the speed of light, with only very minor decreases in speed (as manifested in an index of refraction) when they move in materials other than an absolute vacuum. Second, by being much smaller than electrons, photons require much less space for storage and other handling. Another great advantage of photons over electrons as data carriers is that they do not have charge. Thus, they do not interact with one another to impede one another’s movement. Unfortunately, while velocity at or very near the speed of light comes easily with photons, ability for storage does not! But, beyond this “challenge to be overcome”, there are some challenges associated with joining photonic devices. These include the following: . As stated earlier, as chip-borne circuits get thinner, and demands for
136
Joining comes of age: from pragmatic process to enabling technology
Assembly Automation
Robert W. Messler
Volume 23 · Number 2 · 2003 · 130–143
performance in speed and storage continue to rise, there will be an inevitable move from Cu to optoelectronic materials (i.e. photonics). This will occur, as shown in Figure 8, when feature dimensions fall much below about 1 mm, and especially below about 0.2-0.3 mm, due to inherent interconnect delays ( Jeng et al., 1994). Three key components in optoelectronic devices being considered are: (1) GaAs photoemitters, (2) Group III-V compound semiconductor modulators, and (3) SiO2/polymer waveguides. These must all be joined; with each joint posing its own set of complicated problems. So, the inevitable move from Cu-based microelectronics to optoelectronics as the scale of devices falls much further will create a host of challenging new joining problems. It should be clear; how fast and how far Information Technology can go depends in large measure on how clever we get at joining for microelectronic, optoelectronic, and photonic devices.
Joining in nanotechnology and microsystems The prefix “nano” refers to one-billionth (or 102 9). Hence, “nanoscale” refers to ˚ (or about 3-4 dimensions of 102 9 or 10 A atom diameters!). At such scales, many of the
Figure 8 A plot of the delay (which determines computing speed) arising from device and interconnect-related delays as a function of device size, showing the cross-over point much-below which optoelectronics and, at still smaller dimensions, photonics will have to supplant microelectronics
ways in which materials traditionally behave change. Examples of such changes – positive and negative – are: Positive effects . increased strength in nanograined metals, . increased ductility in nanophase ceramics, . increased fracture toughness from nanotube-reinforced ceramic-matrix composites, . enhanced magnetic properties (i.e. magnetoresistance), . enhanced electrical conduction and field emission properties in nanofibers (such as C nanotubes), including recently claimed room-temperature superconductivity (Zhao and Wang, 2002), . quantum behavior such as ballistic conduction in C nanotubes, . enhanced catalytic activity due to increased surface area, and . unique optical properties (such as changes in color at nanosizes). Negative effects . increased resistivity in metals, like Cu, when nanograined, . undesired viscosity effects, for example when nanostructures are used as fillers in composites. The area of Nanotechnology really encompasses two major sub-areas: (1) nanostructure (or, alternatively, nanodevices or nanosystems), and (2) nanophase materials. Both are important in the impact they can have in their own right (as indicated in the preceding list of effects), but they are each also important in the ways they can contribute to advances in Information Technology and Biotechnology; the former largely (but not solely) through the impact of nanomaterials, the latter fairly equally in both sub-areas. For the purposes of this brief overview, let us consider nanoscale materials (e.g. nanocrystalline materials, nanophase materials, nanoparticles, and nanocomposite materials) first, then nanostructures and nanodevices; and always from the perspective of the challenges each will pose to joining. Three joining opportunities seem to be possible using nanomaterials, each with its attendant joining challenges, as follows. (1) Use of nanomaterials, for example, discrete nanoparticles of ceramics or
137
Joining comes of age: from pragmatic process to enabling technology
Assembly Automation
Robert W. Messler
Volume 23 · Number 2 · 2003 · 130–143
that arise from degradation in service correct themselves by re-joining automatically (Chen et al., 2002). There are a number of different approaches, but one thing that would make this possible is the ability to employ what are known as “chemically functionalized nanotubes”. Here, the structure of the nanotube is altered to create chemically active (or “functionalized”) sites. When one such functionalized nanotube encounters another similarly functionalized nanotube, they would automatically chemically bond to effect healing. A similar approach using chemically functionalized nanotubes could be used for joining between newly synthesized, unflawed materials as well. Another entirely different approach would be to incorporate nanoparticle-encapsulated resins plus nanoparticle-encapsulated curing agents (catalysts) into crack-prone materials so that when propagating cracks rupture the encapsulated materials, they are released to react and cause “self-healing” in the form of bonding.
nanotubes of carbon (or, eventually, other elements), as fillers for 2-D and 3-D composites, either as: . filler to impart or improve special properties (e.g. electrical or thermal conductivity) other than strength, as well as to improve strength and/or toughness, . continuous nanotubes for unidirectional reinforcement (using laser CVD, for example) to improve strength and toughness (or damage tolerance), as well as impart some special properties. Research has already shown that the addition of nanofillers results in unique combinations of properties. Examples include: .
.
.
order-of-magnitude increases in scratch resistance in some ceramics, while maintaining optical clarity, improvements in wear resistance accompanied by decreases in the coefficient of friction in some composites (Ash et al., 2002a, b), and increases in strength and modulus of elasticity while maintaining high strain to fracture in some metals (Schadler-Feist, 2002).
For joining, there is the fundamental problem of getting good bonding/ adhesion between the nanophase reinforcement and the matrix phase; which, again, will have to be achieved by clever chemistry with joining occurring as an integral part of the synthesis of the composite. One opportunity that should exist is when nanotubes (or other nanoparticles) are added to filler employed to join composites (including those not comprised of nanophase materials) by any or a variety of traditional joining options (e.g. welding, brazing, soldering, adhesive bonding) as a secondary process (Ajayan et al., 2000; Schadler et al., 1998). So, nanomaterials could dramatically advance the properties obtained in joints between composites. Such an application might be designated “Nanophase Reinforcement Joining (NRJ).” (2) It has become possible to produce materials/structures that are “selfhealing”; in which any flaws that result during material/structure synthesis or
So, the opportunity exists to produce “self-healing” materials and structures in which joining occurs as needed – on demand – when flaws develop. Such an application might be designated “Functionalized Nanophase Joining” (FNJ). (3) It may be possible and worthwhile to develop integral nanoscale mechanical interlocks analogous to Velcro and the other so-called hook-and-look integral interlocking materials/structures. This could be done by growing what have been referred to as “pillars of nanotubes” on one substrate that could interact – either by mechanically interlocking or by chemically bonding (if the nanotubes have been functionalized) – with similar arrays of nanotubes grown on another substrate. By pressing the two together, a joint analogous to that formed between two hair brushes pressed together so their bristles enmesh would result. Figure 9 shows a dense “forest” of such nanotubes growing out of a substrate (Wei et al., 2002). It is not much of a stretch to see how such “forests” of nanotubes could be made to interlock like the bristles of mating hair brushes. Such integral nanoscale mechanical
138
TM
Joining comes of age: from pragmatic process to enabling technology
Assembly Automation
Robert W. Messler
Volume 23 · Number 2 · 2003 · 130–143
Figure 9 An extremely dense “forest” of carbon nanotubes growing out of a substrate to produce what could become the nanoscale analog of integral micro-mechanical interlocks already available commercially, like 3M’s Dual-LockTM fasteners (Wei et al., 2002). One can imagine how two such mating arrays could lock nanostructures together just as two interlocking brushes do
thereby dramatically increasing product yield and throughput, and reducing production costs accordingly. To do this, nanotube clusters could be deposited by a “nanowriter”, analogous to existing micro-writers. Again, joining is in situ and primary. In micromechanisms: (3) Joining would be more akin to present day secondary joining for assembling these tiny devices, but on an entirely new (nano- versus micro-) scale. Examples where such joining could be important are: . pressure sensors, . fluidic devices (like inkjet printers), . inertia sensors (actuators), and . photonic/optoelectronic devices. One approach that is being explored is “self-assembling” micro- (and, eventually, nano-) structures (Harsh et al., 1999; Tuantranont et al., 1999; Fan et al., 1997). In this approach, the mechanical elements comprising what will be the assembly are designed such that they self-interlock when they fall into the appropriate arrangement, and not any other way! The obvious challenges are: (1) the need for high precision; (2)joining dissimilar as well as similar materials; (3) avoiding locked-in stresses that could degrade or destroy the device; and (4) managing any thermal energy associated with the joining process (e.g. some form of micro- or even nano-welding).
interlocks would be small-scale cousins of integral micro-mechanical interlocks already appearing (Messler and Genc, 1998). So, nanoscale materials like nanotubes could create nanoscale integral mechanical interlocks. Such an application could be designated “Nanoscale Mechanical Interlock Joining” or NMIJ. Nanoscale structures, known as either nanostructures or nanodevices, are a logical application of nanotechnology. Such nanoscale devices are closely related to, but on a finer scale than, what are being called “microsystems”. Possibilities for such devices in microelectronics and micromechanisms abound, and pose the following challenges to joining. In microelectronics: (1) Producing seamless electrical interconnects using electricallyconductive nanoscale materials (e.g. carbon nanotubes) to create 3-D network architectures. To do this, topological defects would be generated in the nanomaterials by post-processing. Once again, the challenge to joining is to have joining occur in situ as an integral part of the synthesis of the 3-D device. (2) Producing conductive-path line repairs on chips to salvage otherwise lost product;
Nanotechnology is exciting because it represents both a challenge to joining (e.g. for assembling micro- or nanosystems) and offers new potential to joining (e.g. for producing nanoparticle- or nanotube-filled joints in composites). Joining in biotechnology A broad definition of biotechnology is simply the industrial use of living organisms (or parts of living organisms) to produce foods, drugs, or other products. The oldest examples of biotechnology include fermentation and plant and animal hybridization. The newest biotechnologies range from protein separation technologies to genomics and combinatorial chemistry. Like Information Technology, Biotechnology can encompass many areas, depending on one’s perspective and openmindedness, but a sampler of fields would
139
Joining comes of age: from pragmatic process to enabling technology
Assembly Automation
Robert W. Messler
Volume 23 · Number 2 · 2003 · 130–143
include: bacteriology, biochemical engineering, bioinformatics, bioprocessing, cell biology, chromatography computational and mathematical modeling, developmental and molecular genetics, DNA technologies, electrophoresis, embryology immunology, materials science, microbiology, nucleic acid chemistry, pharmacology, protein engineering, tissue engineering, and virology (Dordick, 2002). But for purposes of this brief overview focusing on joining, only the areas of medicine and tissue engineering will be considered. Interestingly enough, and somewhat paradoxically, medicine has been slow to adopt new and more innovative joining processes until fairly recently. This is paradoxical because, more often than not, medicine is one of the first application areas to adopt new technologies. Examples include, but are certainly not limited to: (1) early and aggressive adoption of nondestructive evaluation techniques for non-invasive diagnostics, including (a) x-radiography, (b) ultrasound, (c) radioisotope tracers, (d) computeraided tomography, (e) nuclear magnetic resonance, and (f) positron emission imaging; (2) early adoption of computers for (a) artificial intelligence systems for detecting potentially toxic combinations of drugs and assisting with diagnoses, (b) virtual reality for teaching anatomy, surgery, etc., and (c) signal processing to assist in the interpretation of non-invasive nondestructive evaluation techniques; and (3) early adoption of wireless communication and data telemetry for (a) remote monitoring of patients and (b) pick-up of diagnostic signals without physical connections or contact.
and in rejoining broken bones. In an analog of fusion welding, electric cauteries and lasers have been used to coagulate and fuse protein in tissue, and innovative techniques have been employed to speed mending of broken bones using electrical, mechanical, and chemical (growth factor) stimulation. The real opportunities for breakthroughs in biotechnology, however, will occur likely in tissue engineering, and joining will be there! The challenges relate to regenerating tissue of all kinds; soft, 2-D, vascularized skin; medium, 3-D nonvascularized cartilage; and hard, 3-D vascularized bone; and, ultimately, nerve tissue. The key in every case, but with critical differences in the details, is to find ways to stimulate “joining”, as the basis for tissue regeneration, by finding ways to create “scaffolds” on which anchorage-dependent cells can attach and proliferate. Such scaffolds need to be designed to be compatible (for adhesion), employ growth factors (for stimulating and controlling growth), take advantage of mixing cells (to suit needs), and matching of growth and resorption rates during healing. And, most important of all, and in every case, joining in tissue regeneration must always achieve the functional properties of the specific tissue; but this is nothing new, as the best joints must always achieve the functional properties of the materials being joined. Figure 10 shows one exciting recent development in which “self-assembling” peptide-amphiphile molecules create a cylindrical micelle, in a process that helps dramatically speed up the healing of broken
When it came to adoption of joining techniques, however, the situation was different; medicine was slow on the uptake. Joining to close wounds and surgical incisions has involved the use of stitches (or sutures) since the earliest practice of the profession, and only recently the inclusion of staples; both are mechanical fastening techniques. For mending (rejoining) broken bones, mechanical fasteners like screws, pins, and wires have also been used almost exclusively. Only quite recently surgeons have begun to use biodegradable or reabsorbable adhesives for closing wounds or incisions in soft tissues, 140
Figure 10 A schematic that shows the self-assembly of peptide-amphiphile molecules into a cylindrical micelle, a process that helps heal broken bones by mimicing the way collagen forms a scaffold on which new bone cells can form and grow
Joining comes of age: from pragmatic process to enabling technology
Assembly Automation
Robert W. Messler
Volume 23 · Number 2 · 2003 · 130–143
bones (ASEE PRISM, 2002). Perhaps the most interesting finding relating to “joining” for tissue regeneration, albeit not so different than that observed with certain other joining processes like adhesive bonding, soldering and brazing, where “wetting” is essential to joint formation, is that proper surface conditioning of the substrates is critical for anchoragedependent cells to attach, proliferate and function to produce new tissues. So, while it is an intellectual stretch for some, it is really quite clear that joining will be at the very core of major opportunities for advances in medicine and tissue engineering within Biotechnology.
Most notably, joining will move from a discrete “secondary” process performed as a last step in manufacturing – often as an afterthought – to a “primary” process that is fully integrated with the processes of synthesizing the materials involved and the device(s) or structure(s) being created. This is not to say that joining as a “secondary” process will totally disappear, or even diminish in its economic importance and impact, as it will not, but simply that joining will take on an additional, and much more powerful, role. What is especially interesting is that a great deal of the way in which joining will have to occur in the future in the new technologies of Information Technology, Nanotechnology, and Biotechnology implies it will have a “life” of its own. The concepts of “self-forming bonds” and “self-limiting bond formation” in micro- (and, eventually, nano-) electronics, “self-healing” in nanomaterials (especially composites), and “self-assembling structures” in microsystems and nanosystems and even tissue engineering make this clear. This poses profound new challenges to the chemistry, physics, and mechanics of this old “workhorse” process. With nanotechnology and microsystems, as with microelectronics (and what will inevitably become “nanoelectronics” or even “moletronics”) joining will be extended downward to a scale where the boundary between a device or structure and the materials comprising it will be blurred or lost. With biotechnology, joining will expand from life-enriching to life-enabling applications; which is a profound transition. And, perhaps most profound of all, the underlying chemistry and physics of joining will tend to dominate and pace advances more than – although obviously not supplant – engineering. The process will be practiced as much or more as a science by technicians and physicians than as a trade by helmeted welders or hard-hatted riveters. That is quite a change for what has been a pragmatic process!
So, what must change? For more than two millennia, joining has been a pragmatic, albeit critically important, fabrication process in which the end-result not only justifies, but usually drives, the means to that end. From its simplest beginnings when early humans (or perhaps humanoids) first lashed a sharp stone to a split straight stick to make a spear or an ax, to modern times where robots spot-weld trunk decks and hoods, quarter panels and door panels, and roofs into automobile bodies and to join bodies to chassis, joining (in many forms) has contributed significantly to the generation of the wealth of industrialized countries. When the ability to use materials advanced from stone to wood (in and before the Stone Age) to malleable copper and bronze (in the Bronze Age) to strong iron and steel (in the Iron Age and the Industrial Revolution), joining processes advanced in response to material advances. But, now, as the rate of material advances are occurring at an ever-increasing pace, joining processes have not kept pace (Messler, 1996). Few real changes – fundamental changes – have occurred in joining in almost a century (Messler, 1997). As we enter the 21st century and a new millennium, things are about to change – must change – with joining (Messler, 2000); the process is about to “come of age”. The major new technologies of Information Technology, Nanotechnology, and Biotechnology are forcing changes that are already showing themselves in microelectronics. The pragmatic fabrication process is rapidly becoming – and must continue to become – an enabling technology.
Conclusions The world is being changed by technology in ever-more rapid and dramatic ways. We are embarking on a path made possible by Information Technology, Nanotechnology, and Biotechnology that will determine the future of the entire species; for better or
141
Joining comes of age: from pragmatic process to enabling technology
Assembly Automation
Robert W. Messler
Volume 23 · Number 2 · 2003 · 130–143
worse, depending on how thoughtfully and appropriately we proceed ( Joy, 1999). Joining, as a process always at the heart of making possible new devices and structures and complex assemblies for accomplishing diverse and often formidable tasks, will, of necessity, become an enabling technology for progress in and full realization of the potential of these new technologies. Joining is about to “come of age”!
Notes 1 In fact, robotics, or the robots that result, rely on information technology and, seemingly, will rely on nanotechnology, at least for robots in the form of the so-called “microsystems” or MEMs, or, eventually, “nanosystems” or NEMs. It is even possible to envision reliance on biotechnology for robotic prosthetics (like in the “Six Million Dollar Man” of TV fame three decades ago) or even the human-like automatons of what was once science fiction. 2 Deformation processing includes a large number and variety of processes for shaping materials or parts by changing shape through plastic deformation. Metals, thermoplastics, and, to a lesser extent, glasses, are commonly deformation processed. Examples of specific processes are forging, rolling, extrusion, drawing, forming, spinning, etc. for metals; thermal molding and blow molding for thermoplastics and glasses. 3 Until now, virtually every piece of hardware used in Information Technology, most notably computers, has been based on electromagnetic devices, e.g. electrons in electronic devices, and photons in optical devices and, increasingly, in photonic devices. The possibility of the so-called “quantum computing” will also still be electromagnetically based. However, the possibility of molecular computing within what is being called “Moletronics” could easily be envisioned as going beyond electromagnetically-based approaches to biologically-based approaches; and, therefore, such devices could span both Information Technology and Biotechnology (Kwok and Ellenbogen, 2002). 4 The speed with which electrons can move in materials is referred to as their “electron mobility”; and that speed is always less than the speed with which electrons can move in the absolute vacuum of space, i.e. the speed of light. 5 In the latter case, there is genuine cause for concern and action, as the Roman’s learned the hard way what lead in water-supply systems can do to the public’s – and the Empire’s – health. In the former case, there is far less real data to show that lead from electronic solder joints is leaching into the ground water underlying landfills into which old computers and other electronic gear are being disposed!
References Ajayan, P.M., Schadler, L.S., Giannaris, S.C. and Rubio, A. (2000), “Mechanical response of singlewall carbon nanotubes in polymer composites”, Advanced Materials, Vol. 12, pp. 750-9. ASEE PRISM (2002), “Lucky break for bone fracture patients”, Briefings, March 2002, p. 14. Ash, B.J., Schadler, L.S. and Siegel, R.W. (2002a), “Glass transition behavior of alumina/ polymethylmethacrylate (PMMA) nanocomposites”, Materials Letters, Vol. 55, pp. 83-7. Ash, B.J., Rogers, D.F., Wiegland, C.J., Schadler, L.S., Siegel, R.W., Benicewicz, B.C. and Apple, T. (2002b), “Mechanical properties of Al2O3/ polymethylmethacrylate nanocomposites”, Polymer Composites, Vol. 23 No. 6, pp. 1014-25. Chen, X., Dam, M.A., Ono, K., Mai, A., Shen, H., Nutt, S.R., Sheran, K. and Wudi, F. (2002), “A thermally re-mendable cross-linked polymeric material”, Science, Vol. 295, pp. 1698-702. Dordick, J. (2002), in private communication. Easterling, K. (1990), Tomorrow’s Materials, 2nd ed., The Institute of Metals, London, England. Fan, L., Lee, S.S. and Wu, M.C. (1997), “Self-assembled micro-XYZ stages for optical scanning and alignment”, Proceedings of the IEEE LEOS Annual Meeting, 10-13 November 1997, San Francisco, CA. Forester, T. (ed.) (1988), The Materials Revolution, The MIT Press, Cambridge MA/London, England. Harsh, K.F., Bright, V.M. and Lee, Y.C. (1999), “Solder self-assembly for three-dimensional microelectromechanical systems”, Sensors and Actuators A, Vol. 77, pp. 237-44. Hughes, G. (2001), in private communication. Jeng, S-P., Havermann, R.H. and Chang, M-C. (1994), “Advanced metallization for devices and circuits – science, technology, and manufacturability”, Proceedings of the Materials Research Society Symposium, 4-8 April 1994, Pittsburgh, PA, Vol. 331, pp. 25-35. Joy, W. (1999), “Why the future doesn’t need us”, Wired, pp. 238-62. Kirchner, E. (1996), “Investigations of ultra thin aluminum as an adhesion promoter and electrically stable diffusion barrier for copper metallization on SiO2”, PhD thesis, Rensselaer Polytechnic Institute, Troy, NY. Kwok, K.S. and Ellenbogen, J.C. (2002), “Moletronics: future electronics”, Materials Today, pp. 28-37. Lindberg, R.A. (1990), Processes and Materials of Manufacture, 4th ed., Allyn and Bacon, Needham, MA. Meindl, J.D. (1987), “Opportunties for gigascale integration”, Solid State Technology, Vol. 30 No. 12, pp. 84-9. Messler, R.W. Jr (1993), Joining of Advanced Materials, Butterworth-Heinemann, Stoneham, MA. Messler, R.W. Jr (1996), “The challenges for joining to keep pace with advancing materials and design”, Materials and Design, Vol. 16 No. 5, pp. 261-9. Messler, R.W. Jr (1997), “Joining technologies for the next century: drivers and directions”, Journal of Assembly Automation, Vol. 17 No. 1, pp. 54-63. Messler, R.W. Jr (1999), Principles of Welding: Processes, Physics, Chemistry and Metallurgy, Wiley, NY, USA.
142
Joining comes of age: from pragmatic process to enabling technology
Assembly Automation
Robert W. Messler
Volume 23 · Number 2 · 2003 · 130–143
Messler, R.W. Jr (2000), “Trends in key joining technologies for the 21st century”, Journal of Assembly Automation, Vol. 20 No. 2, pp. 118-27. Messler, R.W. Jr and Genc, S. (1998), “Integral micro-mechanical interlock (IMMI) joints for composite structures”, Journal of Thermoplastic Composite Materials, Vol. 11 No. 5, pp. 200-15. Moore, G.E. (1965), “Cramming more components on to integrated circuits”, Electronics. Murarka, S. (2002), in private communications. NCMS (1997), Lead-Free Solder Report, National Center for Manufacturing Sciences, Ann Arbor, MI, NCMS Final Report 0401RE96. Schadler, L.S., Giannaris, S.C. and Ajayan, P.M. (1998), “Load transfer in carbon nanotube epoxy composites”, Applied Physics Letters, Vol. 73 No. 26, pp. 3842-4. Schadler-Feist, L.S. (2002), private communication. Suwwan de Felipe, T. (1998), “Electrical stability and microstructural evolution in thin films of
high-conductivity copper alloys”, PhD thesis, Rensselaer Polytechnic Institute, Troy, NY. Tuantranont, A., Bright, V.M., Zhang, W., Zhang, J. and Lee, Y.C. (1999), “Self-aligned assembly of microlens arrays with micromirrors”, Proceedings of the 1999 International Society for Optical Engineering, Micromachining and Microfabrication, Santa Clara, CA, Vol. 3878, pp. 90-100. Tummala, R.R. and Rymaszewski, E.J. (1989), Microelectronic Packaging Handbook, Van Nostrand Reinhold, NY, p. 53. Wei, B.Q., Zhang, Z.J., Ajayan, P.M. and Ramanath, G. (2002), “Growing pillars of densely packed carbon nanotubes on Ni-coated silicon”, Carbon, Vol. 40 No. 1, pp. 47-51. World Almanac (The World Almanac and Book of Facts) (2001), in McGeveran, W.A. Jr (Ed.), World Almanac Books, Mahwah, NJ, Vol. 151, pp. 131-2. Zhao, G.-M. and Wang, Y.S. (2002), “Research News”, Materials Today, p. 14.
143
Introduction
Feature Clinching with a superimposed movement – a method for force reduced joining Rolf Dieter Schraft Stefan Schmid and Achim Breckweg
The authors Rolf Dieter Schraft, Stefan Schmid and Achim Breckweg are all based at the Fraunhofer Institute for Manufacturing Engineering and Automation IPA, Stuttgart, Germany. Keywords Assembly, Joining, Metals Abstract Clinching is, due to its characteristics, a joining method with several advantages. The high joining forces, which require heavy process equipment are a major disadvantage. The Fraunhofer Institute has developed clinching methods which reduce the joining forces considerably to make clinching applicable for further developments and new application areas. Electronic access The research register for this journal is available at http://www.emeraldinsight.com/researchregister The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0144-5154.htm
Assembly Automation Volume 23 · Number 2 · 2003 · pp. 144– 146 q MCB UP Limited · ISSN 0144-5154 DOI 10.1108/01445150310471374
Clinching or press joining, gains more and more importance as a form and force fit joining method for sheet metal and metal profiles. The joining elements are directly shaped out of the sheet metal to be joined. Clinching can replace mechanical fasteners like bolts or rivets, requiring no pre-drilled holes. A spot joint, similar to a press button, is made by local plastic deformation with an adjusted punch and die. It does not matter whether the sheet metal is coated or surface treated, of dissimilar thickness, or of unlike materials, e.g. steel or aluminium. The joining element is made by a cold forming process. No heat is introduced into the sheet metal. No poisonous gases are discharged and there is no embrittlement of the material. Conventional clinching is characterised by a translatory movement of the male punch. Thereby forces up to 100 kN are generated. The high joining forces exert a negative impact on tool life and require heavy and cumbersome process equipment.
Developments To satisfy the need for flexible and applicable joining equipment, it was a prerequisite to reduce the joining forces dramatically. This was realised by new clinching processes with an orbital or radial movement of the punch. Clinching with a superimposed movement allows the orbital clinching tool to rotate on a circular path or a radial path (hypozycloide) (Figure 1). Here the punch-track always moves over the middle of the joining element without any rotation around its own axis. The schematic diagram in Figure 2 shows the rolling off movement of the joining punch. This causes a partial metal forming in the contact zone of the punch. A comparatively low axial force forms the two layers of sheet metal into a die. The joined material flows according to the shape of the die. The specific metal forming conditions promote the reduction of the axial joining forces as well as the generation of an undercut in the joining element. Using clinching with a superimposed movement, it is possible to achieve the following advantages: . 50-70 percent less joining forces, . larger throats for joining equipment, 144
. .
Clinching with a superimposed movement
Assembly Automation
Rolf Dieter Schraft, Stefan Schmid and Achim Breckweg
Volume 23 · Number 2 · 2003 · 144–146
joining of thicker materials, and less adhering of metal to the punch especially with washed aluminium.
Despite considerably lower joining forces (Figure 3) with the superimposed clinching method, the achievable process times and joint strengths are similar to conventionally produced joining elements. The orbital or radial punch movement noticeably improves the material flow.
Tooling The punch kinematic offers the usage of several shapes of punch tips. Punch with an undercut as shown in Figure 4 may be used for the orbital clinching. The radial clinching allows also a cylindrical or a conical shape. The results of this experiments with different punch tip shapes are shown in Figure 5. For shear load, the best results are
achieved with the conical punch. The reason for this result can be seen in the improved form for the stress flow in the joining element. Cylindrical or punches with an undercut offer advantages in vertically applied load due to the improved material flow during the clinching process. The results with a die (Eckold, 95010) show a comparable joint strength to conventionally produced spot clinching elements. These advantages were reason enough to promote the tooling development. The result of the development work at the Fraunhofer IPA are light-weight robot frames with a large throat. The orbital and accordingly radial clinching tools weigh 40-60 percent less than conventional robot tools. The light-weight robot tools facilitate shorter point-to-point times and shorter settling times. It is even possible to switch to lower robot classes and thus cheaper robots (Plate 1).
Figure 1 Orbital and radial punch kinematics
Figure 3 Plot of joining force versus joining time
Figure 2 Punch kinematics at the clinching with superimposed movement
Figure 4 Punch with undercut, cylindrical and conical tip geometry for radial clinching
145
Clinching with a superimposed movement
Assembly Automation
Rolf Dieter Schraft, Stefan Schmid and Achim Breckweg
Volume 23 · Number 2 · 2003 · 144–146
Figure 5 Shear strength for static load
Plate 1 Heavy duty radial clinching robot tool in modular design
vehicle production. Often these areas are reserved for spot welding. The orbital or radial clinching robot tools allow force reduced clinching, despite a more complex tooling system. The advantages for the user are a more economic and a more flexible process layout. The clinching equipment suppliers can expand their product range because of new fields of applications and the material manufacturer also gain advantages from the development due to an increased application potential. Especially the development of light-weight material requires economic joining methods. Using clinching with a superimposed movement, it will be possible to enlarge the field of applications for tasks using spot clinching.
Further reading
Conclusions Clinching with superimposed movement creates new areas of application in the automotive and white goods industry as well as in areas like commercial vehicle or rail
Schondelmaier, J. (1992), “Grundlagenuntersuchungen ueber das Taumelpressen”, Dissertation, University of Stuttgart. Spingler, J. and Woessner, J.F. (1999), “Taumelnd clincht es sich leichter”, Industrieanzeiger, Vol. 8. Thoms, V., Westkaemper, E., Kalich, J. and Breckweg, A. (2003), “Entwicklung von Verfahren und Einrichtungen zum Radialclinchen”, Project Report EFB/AiF 27 ZBG.
146
Feature Productivity and quality improvements through orbital forming Werner R. Stutz
The author Werner R. Stutz is Vice President at Taumel Assembly Systems, New York, USA. Keywords Assembly, Metal forming Abstract Orbital forming is an efficient and precise process to assemble component parts. It provides strength, an attractive finished appearance, and batch-to-batch uniformity. Orbital forming machines can produce hightorque assemblies and also freely swinging joints, and any degree of built-in resistance in between. These machines quietly flare and form all malleable materials, including many engineering thermoplastics, and work safely on delicate and brittle parts. The machine controls provide infinitely variable cycle times (speed), forming pressure and tool stroke on the micrometer dial with resolution to 0.001 in. Electronic access The research register for this journal is available at http://www.emeraldinsight.com/researchregister The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0144-5154.htm
Assembly Automation Volume 23 · Number 2 · 2003 · pp. 147– 152 q MCB UP Limited · ISSN 0144-5154 DOI 10.1108/01445150310471383
Introduction Orbital forming machines have been used to assemble component parts for 35 years. With the introduction of flexible manufacturing and advanced productivity concepts, including ISO, just-in-time (JIT) and measurable process control, the use of orbital forming in production is increasing. Orbital forming is a clean, silent, non-impact and vibration-free coldforming process. Often, it is used as an alternative to conventional staking, peening, crimping, pressing, swaging, spinning, rolling, riveting, welding, upsetting, and other fastening operations that may lack or vary in uniformity and workpiece quality. An orbital forming machine flares studs, pins, posts, hubs, tubing, spacers, rivets, and other fasteners quickly. If one or more shoulder pins must be secured to a blanked plate, a single spindle machine can be used to secure one pin at a time or a multi-spindle system is selected to flare any combination of pins, hubs or posts on the same workpiece (Plate 1). The process can also be used to assemble components without fasteners. For example, Figure 1 shows semi-pierced studs on a fine blanked part being flared over a mating plate. On an automotive part, two to four bent-up tabs on a stamped mounting plate may be designed as spacers that fit through matching square or rectangular holes of a mating cover plate. An orbital forming machine can be used to flare these tabs for a secure and correctly aligned assembly without fasteners (Plate 2). Likewise, an orbital forming machine will often eliminate the need for separate hardware: blanked ridges, bosses and integral projections of malleable material can be formed out to anchor components in position. For example, a leaf spring may be captured on a stamping by flaring out two blanked rib sections with a single form tool. High torque joints in solid or tubular form in stamping assemblies may be produced in many ways with orbital forming machines. Blanked holes in a plate may be designed in D or double-D configurations, often with the provisions for specific parts orientation. Shafts or tubes with the matching D or double-D shapes can be made to protrude through the stamping for assembly on the headforming machine.
147
Productivity and quality improvements through orbital forming
Assembly Automation
Werner R. Stutz
Volume 23 · Number 2 · 2003 · 147–152
Similarly, round shoulder pins may be flared in punched square holes of a mating plate, allowing the headforming machine to cause available material to flow into the corners of square holes. Further, one or more round posts or square tabs protruding through punched holes in stampings can be formed. Notches or serrations may be blanked into round holes to increase the contact area and improve torque resistance after the part assembly. Compared to any press method, the orbital process requires only 12-15 percent of the forming forces. The reason is the same as that used to explain how a 100-pound woman can gouge dents into hard flooring when she wears
spiked heels, while a much heavier person wearing flat-bottomed shoes does not. Orbital forming uses less force but concentrates it on a continuous radial line emanating from the center of the shaft, rather than over the entire area to be formed (Figure 2). This prevents damage to the opposed, threaded shaft end, facilitates the parts support, and allows for the use of simplified fixture tooling.
Plate 1 Multi-spindle machine simultaneously flares out six steel studs on stamped plate. Each (non-spinning) form tool orbits in its own forming path for superior joint strength and finish
The orbital forming process The orbital forming process can be used in place of most conventional fastening or assembly methods. With few exceptions, the results will be substantially better in terms of quality, uniformity, joint strength and appearance of the form produced. In orbital forming, a form tool, mounted off-center in a revolving spindle, much like a tool in a jig-borer, is inclined at a slight angle towards the center of the spindle. The tool axis of the form at the working end of the tool intersects with the true centerline of the spindle. The machine spindle rotates but the tool in the orbital head or chuck is free to rotate in its bearings. The drive spindle advances, bringing the tool into contact with the workpiece. At this point, a preset, constant pressure is applied for a pre-determined length of time. The line contact between the non-spinning tool and the work never varies. At each revolution of the spindle, the same line of contact is maintained to flare the material radially. Because the same point on the tool is always in contact with the same point on the workpiece, almost no friction occurs and no tearing of the work material results, regardless of the component shape (solid, tubular, triangular, square, hexagonal, oblong or semi-circled). The orbital movement can be combined with controlled pressure, tool stroke and cycle time (work speed). Depending on the application, two or more factors may be used – each of them infinitely
Figure 1 Semi-pierced studs can be flared over a mating plate
148
Productivity and quality improvements through orbital forming
Assembly Automation
Werner R. Stutz
Volume 23 · Number 2 · 2003 · 147–152
adjustable – in conjunction with the forming process. A combination of preset pressure and cycle time is sometimes used, usually when two or more parts to be joined by studs, pins or rivets are subjected to broad tolerance variations, or when a brittle base component such as ceramic, glass or phenolic is used. Orbital forming machines and modular units generally are stationary systems. Usually, parts and components must be brought to the machine. The use of portable headforming units in a gantry-type setup is rare and limited to the assembly of aircraft structures (fuselage and wings), railroad cars, truck bodies, and large turbine installations.
Although orbital headforming is used frequently for riveting work, typically on brittle material or for precision assemblies, conventional rivet-setting machines with automatic rivet feeds are sometimes more effective for less critical riveting applications.
Plate 2 Orbital multi-point unit flares out four semi-pierced posts on magnetic base to secure speaker basket stamping. The orbital forming process also eliminates breakage of the brittle magnet bodies that previously occurred with a press method
Types of orbital forming Multiple-point forming Multiple-point forming on a single workpiece or gang heading multiple parts is possible, with many variations (Figure 3). An orbital headforming machine can be configured to carry as many tools as needed and with centers as close as 3/16 in. to meet multi-task applications. Larger multi-spindle systems may have tooling plates up to 20 in. across, often with tools working at different heights or heading levels and on more than one part simultaneously. Changeable tooling sets allow simple conversion from one heading pattern (job) to the next. Standard multiple attachments – in-line and random pattern tooling and two, three, and four variable-center-distance spindles – fit most machines. Double-ended, opposed heading For double-ended, opposed heading, two modular heading units generally are mounted horizontally on a machine bed, facing one another. This configuration allows cut rods, shafts, or tubes to be used, as well as cutoff and feed devices to load the machine.
Figure 2 The orbital forming process
Swing-joint forming The precise stroke control on orbital headforming machines makes possible the production of pliers, scissors, pocket knife joints, surgical sutures, gear trains, bobbins, handcuffs, or any swing-joint assembly as “tight swing”, “loose swing”, or “floating”, as desired. Non-round headforming Square, rectangular, oval and single- or double-D solid shafts can also be orbitally formed. Three rectangular studs can be simultaneously headed with a single large, round tool in a single-spindle orbital head, for example, or two and more square or rectangular tabs may be flared with a multispindle head. 149
Productivity and quality improvements through orbital forming
Assembly Automation
Werner R. Stutz
Volume 23 · Number 2 · 2003 · 147–152
Figure 3 Examples of multiple-point forming. The pre-formed profile is on the right of each example, the post forming profile is on the left
Aspects of the process
Plate 3 Orbital bench top machine forms pins on precision bearing cage; parts are positioned on manual slider plate
Cycle time As a rule, shorter cycle times in manufacturing operations result in better production rates. In general, cycle times for orbital headforming run from 0.5 to about 2.0 s on solid steel studs of higher tensile strength. This includes tool approach, form dwell, and spindle return but not part loading. The softer and more malleable a material is, and the smaller its diameter, the shorter the work cycle is. However, even on steel pins of 1.0 in. in diameter, cycle times are about 2 s. When automatic slider plates or index-type fixturing are used, the time required to load parts manually or automatically has a larger impact on production rates than does cycle time (Plate 3). Forming or heading capacity Heading capacity for any size of machine is governed not so much by the diameter of the figuring as by the total surface area to be formed and the material’s tensile strength. For example, a machine that can flare a 5/16 in. diameter solid shoulder pin in mild steel can be used to swage over the shell of a Type D flashlight battery to crimp or seal its end. It can also flange a tube or a 3 in. diameter hollow aluminum body that has a wall thickness of 0.030 in. Fixturing Parts placed or fed into the fixture of an orbital forming machine usually can be left freestanding and require no clamps or holddowns. No spinning force is transferred from the work head to the parts, so they remain where they are placed through the cycle. For example, a two-plate assembly to be joined by one rivet usually requires a simple locating nest with a pocket to position the pre-formed rivet head. To orient or to secure springloaded parts before and during the assembly process, a hold-down device can be mounted
onto single tool or multi-point forming heads. A simple parts compression device can also be added directly to the form tool (Figure 4). Stud placement Hard-to-reach pins can be dealt with by orbital headforming. A shoulder stud may be secured close to a vertical wall or in a recess of a part. The reach of the tool is limited only by the clearance around it while it is orbiting. Changing the angle of the tool by as little as 28 can reduce the clearance requirement (Figure 5). Selectable pressure/time settings On orbital forming machines, cycle time and heading pressure can be set for specific tasks. Both are infinitely adjustable. Depth control
150
Productivity and quality improvements through orbital forming
Assembly Automation
Werner R. Stutz
Volume 23 · Number 2 · 2003 · 147–152
on the micrometer dial can be set in increments of 0.001 in. This makes it possible to form either firmly fixed joints (with selectable, increasing torque strength) Figure 4 A hold-down device can be used to secure spring-loaded parts
Figure 5 Hard-to-reach pins can also be formed
or smoothly moving swing joints (which can also have the chosen amount of required built-in resistance). Many manufacturers assemble thin or fragile materials, especially those that produce electrical or electronic parts. Common problems they encounter are breakage and loose assemblies. A typical example is a multi-contact thermostat assembly. The cumulative thickness tolerances on each stack of parts may be ^ 0.030-0.040 in. The selectable pressure and time settings on an orbital forming machine can help eliminate breakage and loose assemblies. For pressure to build up, there must be resistance to the force being applied. When the orbital tool is allowed to move as far as it wishes in its stroke, and its movement is stopped only after a certain pressure is built up in its chamber, the machine controls can compensate for any variations in the thickness of the stacked assembly. Pressure can be adjusted until it is sufficient to form a rivet, locking all the elements of the assembly in place without pressure building and cracking the ceramic insulators. In addition, pressure is maintained at a preset limit (set by a relief valve) long enough to prevent the material being formed from springing back after the pressure is relieved. Materials In general, any malleable material up to Rockwell 35C can be formed orbitally. This includes most ferrous and non-ferrous metals, stainless steel, zinc and aluminum die-cast material, some sintered metal, and many types of engineering thermoplastics (Plate 4). Plate 4 Fastening without fasteners: parts assembled by forming integral die-cast studs, pins, tabs or bosses to secure mating parts
151
Productivity and quality improvements through orbital forming
Assembly Automation
Werner R. Stutz
Volume 23 · Number 2 · 2003 · 147–152
In addition, thin case-hardened steels and plated, painted, or plastic-coated materials can usually be orbitally headformed, because material displacement is microscopic during each tool revolution. After orbital headforming, the coating surface is usually left in its original condition. The lustre of some plated surfaces actually improves. Microphotographs show that orbital headforming does not disrupt the molecular structure of metals. However, compressing the grain structure work hardens the material somewhat to make a stronger connection for rivets or a harder contact surface in the case of flared, flanged, or swaged parts (Figure 6).
Tool life Because the orbital headforming process is non-impact, tools sustain little wear. Instead, the process action causes the formed ends to become polished, work hardened, and nearly maintenance-free. A flat-faced tool to form mild steel studs may last for years without requiring any maintenance. For more complex tool shapes or those that are used on
certain aluminum or brass grades, periodic polishing may be required. Operator skill An orbital headforming machine usually requires a few minutes of work by a setup person, along with some trial and error to establish operational parameters. After it is setup and its controls are set to automatic mode, the machine can be operated by skilled or unskilled workers. Process control with orbital forming Orbital forming systems for cost-effective quality assurance with real time height measuring and pressure sensing for parts assembly are available with many process monitoring options for single work place, integrated rotary or inline multi-station assembly machines and flexible work cells.
Conclusion Orbital forming is an efficient and precise process to assemble component parts. It provides strength, an attractive finished appearance, and batch-to-batch uniformity.
Figure 6
152
History
Feature Robotic “layup” of composite materials David Groppe
The author David Groppe is based at Composite Systems Inc., Arnold, CA, USA. Keywords Composite materials, Robots, Automation Abstract This paper describes the history and current technology behind composite manufacturing and the development of a precision feed endeffector (PFE). The PFE is used on the end of a robot arm and performs many functions associated with the handling of prepreg and semipreg materials. The PFE helps to achieve higher levels of accuracy and productivity for automated layup systems. Electronic access The research register for this journal is available at http://www.emeraldinsight.com/researchregister The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0144-5154.htm
With the advent of “composite materials”, including cores, in its nearly limitless forms and characteristics that continues to evolve, new “application” technology has continued to be developed in concert with its use. With the improvements in plug and mold designs, engineers have been allowed to “assemble” complex parts. Of these many advances, lasers have assisted greatly in the verification of location and orientation of the materials used when layup is performed by hand. The use of water-jet systems has allowed for holes and cut outs to be performed after the “layup” has been completed and the part cured. Significant improvements in “XY cutting” systems and associated software has expedited the “profiling” or cutting of the materials prior to layup. In the “winding” regime, advances in fiber placement machines has provided increased throughput, although these systems are very “task specific”. With the use of automated tape laying (ATL) machines, the placement of “unidirectional” tapes has been employed in a very limited group of aerospace and marine applications due to their expense, limited flexibility and limitations imposed by the specific materials that can be used by the systems. Industries that are increasingly demanding both improved materials and methods for layup consists of the aerospace, automotive, marine, wind energy systems (blades), furniture, telecommunications, transportation (i.e. high speed rail, ship building, motor homes, sport boats, semi-trucks, trailers, shipping containers), residential homes, architectural applications, oil and gas exploration pipes and space systems, to name a few.
Industry dilemma
Assembly Automation Volume 23 · Number 2 · 2003 · pp. 153– 158 q MCB UP Limited · ISSN 0144-5154 DOI 10.1108/01445150310471392
A significant key to the growing use of composites, outside the benefits derived from the material characteristics themselves, is the ability to use such materials without, or at least limiting, the expense associated with the labor costs traditionally endured by the industry. To find qualified personnel, to perform the many and varied tasks associated with the “layup” or “placement” of such materials, can place limitations on which projects can be justified to use them. Herein 153
Robotic “layup” of composite materials
Assembly Automation
David Groppe
Volume 23 · Number 2 · 2003 · 153–158
lies the dilemma for manufacturing. The expression “pounds-per-hour” is fundamental to the justification of the use of these advanced materials. These issues affect all programs employing “wet layup”, “dry layup” (RTM applications) and prepreg or semipreg applications. Current technology, until now, has limited this “number” in concert with the fact that the systems currently available limit the “types” and “sizes” of materials that can be automatically “laid up” or placed and the speed or amount of material that can be dispensed per hour of production. Typical methodology dictates that more personnel are simply required to get the “pounds-per-hour” required to justify the program. At some point of time, however, you reach a paradox. The placement of core is affected by such issues as well. These materials, too, are varied in makeup and shape and must also be “placed” in the proper position during the layup regime. The handling of such materials can be awkward, at times, due to the sizes and shapes and the specific location for it to be placed, as required by engineering. The issue of labor, again, plagues the process.
Use of robotics The use of robots to perform a variety of tasks is well documented. However, their use in the field of composites has been limited, to date, as the “end-of-arm” equipment has been limited to, for the most part, water-jet, drilling/tapping, material handling (limited), assembly and fiber placement applications. These advents have significantly improved quality and speeds at which these varied aspects of composite use are performed. Because of the flexibility associated with robots and the use of additional “axes” of motion that are commercially available as “auxiliary axis packages” from most commercial robot manufactures, increased “cell size” or work envelopes can be realized. The use of floor or “wall mounted” tracks have given a single robot the ability to perform tasks in a variety of “work environments” along the track on which it rides. By using gantry systems, the robot(s) work envelope is substantially increased. These systems lend themselves to large work pieces where work is performed in all three “primary axes” (Figure 1).
Figure 1 Robot on traveling overhead gantry provides X, Y, Z layup capabilities
In concert with these systems, additional axes may be employed within the “cell envelope” to position or manipulate the work piece in a “coordinated” movement to the system, thus adding flexibility to the cell.
Precision feed endeffecter technology As mentioned, the requirement to precisely place or layup “pounds-per-hour” is critical to the justification of the use of composite materials, regardless of the process employed. To that end, precision feed endeffecter (PFE) – patents pending and applied for worldwide – technology for commercial robots has immerged. Commercially available material, whether dry or prepreg, is available in widths that range from 1.00 in. (25.4 mm) to excess of 60.00 in. (1,524 mm), with supply roll cores ranging from 3.00 in. (75 mm) to 12.00 in. (305 mm) in diameter, and outside diameter (OD) up to 26.00 in. (660 mm), of which the feed system has been designed to accommodate. With respect to prepreg or semipreg materials, they may be either unidirectional or woven in nature with varying amounts of resin impregnated to one or both sides of the material. Use of such materials is geared toward more control and uniformity in the layup regime, as both the resin type and volume and “fiber characteristics” lend themselves to be an “engineered” product. PFE technology was developed to utilize the flexibility and ease of programming associated with commercial robots. It also addresses “justification” concerns to implement such technology for a given program.
154
Robotic “layup” of composite materials
Assembly Automation
David Groppe
Volume 23 · Number 2 · 2003 · 153–158
Prepreg/semipreg The PFE device combines several aspects associated with the “handling” of prepreg and semipreg composite materials, namely, material feed, refrigeration of the material (Plate 1), the “peeling” of the protective film(s), “profiling” or cutting of the material on one or both edges simultaneously, discharge of the “waste” material, reactivation of the resin to the required temperature prior to placement (optional) and the ability to “absorb” the surface of the mold during layup, whether concave, convex or spline geometry, without requiring the programming of each “point” along a given “path”. Based upon a specific “process” selected, the roll of material is first loaded into the “feed” station on the PFE. The operator would perform this operation “offline” at a “tool crib” located adjacent to the robot cell. The tool crib may house several PFE devices, set up and configured for specific material types and widths. The tool crib would also house “end-of-arm” devices that may perform the placement of core material and imbeds
Plate 1 The PFE combines material feed, refrigeration, peeling and profiling
that may be part of a particular layup schedule. Further, such equipment may also include PFE devices configured to place “bagging” materials whether for “debulking” purposes during the layup regime or in preparation for curing. “Magazines” with core and imbed components may line the cells’ exterior from which the robot or robots may draw from to satisfy an “assembly”. Once the material has been loaded, the operator would then peel a “leader” of the protective film(s). In that event the two films were to be removed, the “bottom” film leader would be wrapped around the supply roll, thus placing both film leaders on top for feed to the “take-up” reel. There, the films would be attached to an empty “core” for take up by the system. The PFE has been designed for easy access and service and opens as a “clamshell” in the vertical orientation, with the hinge being at the top of the device. In the open position, the operator would pull the material down through a series of drive rollers that will both feed and provide “tension” in the “cutting or profiling” station. The drive system can be programmed to overcome specific “tackification” issues that are inherent to various prepreg and semipreg materials as they pass through the system. The profiling station has been designed to accommodate different methods of cutting that may include a “drag knife”, slitting and ultrasonic devices. The waste material exit just below this station by means of a series of pneumatic venturies that are “linked” to the profiling system and apply bursts of air in two directions, one for retaining the material that is to be laid up and forcing it to continue through the tool, and a second, simultaneous burst of air in the opposite direction that directs the waste material out of the device where it is captured for removal by the operator once the PFE device returns to the tool crib. One of the key elements to the PFE device is that of its “suspension system” (Plate 2). This allows the system to “absorb” the surface contours as the robot moves along the “path”. The suspension system provides 100 percent contact with the surface regardless of the supply roll width, and can be programmed to provide a “specific force” so as not to crush core material that may be employed in the layup. Further, because the device provides such controlled “contact criteria”, debulking,
155
Robotic “layup” of composite materials
Assembly Automation
David Groppe
Volume 23 · Number 2 · 2003 · 153–158
Plate 2 The PFE has a suspension system to absorb surface contours
Plate 3 The PFE can be inverted for layup of angled surfaces
to some extent, is provided as a consequence of the layup itself. The “path” described herein, consists of two elements to the robot system. This proprietary “link” between the robot and the PFE device allows for the coordinated “feed of material” with the travel of the robot. The path is combined with the “average offset” that the robot will maintain over the surface and the “nominal centerline” assigned to the direction in which it is moving. This “centerline” is given to the system offline from software capable of producing a “flat blank” or pattern, established at the engineering level and typically supplied to “XY cutting” systems. The robot, the PFE device and the roll of material supplied within it shares this same “centerline”. Laser systems monitor the “alignment” of all of these components to ensure that the “edge” of the material is placed where it is desired. The cutting system receives the “pattern” information in relationship to the path the robot is programmed to take. The system does not require complex algorithms to perform the layup. Verification of dimensional length is provided at the suspension system level of the PFE, just prior to the placement of the material. This is significant where “ply drop-offs” or “field build-ups” are critical to a given layup. By incorporating robot tracts or gantry systems, layup may be performed over or within large mold tooling. The PFE device is capable of being inverted (material/process specific – Plate 3) and may be “daisy-chained” or connected together, end-to-end, for large part layup such as wings, hulls of boats or ships, cylinders as well as flat
panels. Additional axes to the robot system can provide the ability to “manipulate” the mold so as to provide “in-position”, or within gravity, layups, depending upon the material tackification. In applications where two molds will be worked together and provide the “mirror” part, or the other half, gantry systems that employ at least two robots that share a common bridge can greatly reduce the amount of time involved in such layups, thereby, at least, doubling the throughput rate. This becomes evident when a large portion of the programming is shared by both robots, thus synchronizing speed down the long axis of the system. As to the speed of the ten axis PFE device, feed rates of material may exceed 1,20000 in. ipm (30.48 mpm), based upon material limitations. The gantry system or track system employed in the cell will perform these speeds, where the robot need to make only minimal adjustments within its speed range to maintain the path and relative perpendicularity to the surface of the mold (Figure 2). The PFE is manufactured in standard module widths, typically in 6.00 in. (152 mm) increments, assembled at the factory to 60.00 in. (1,524 mm) in width standard (Plate 4 – 24.00 in. /610 mm PFE shown). However, special length configurations and systems may be provided on request. Further, “combination” PFE systems can be provided affording “angled” layups, where, for example, a 6.00 in. PFE and a 24.00 in. PFE can be mounted at a right angle (908) to each other. This configuration would be applicable to where corners could be done in a single
156
Robotic “layup” of composite materials
Assembly Automation
David Groppe
Volume 23 · Number 2 · 2003 · 153–158
Figure 2 The gantry system tracks the robots at constant speeds over large surfaces
Plate 4 The PFE can be manufactured for different widths of layup materials
the PFE device performs a very similar set of functions with only a few additions, which are optional devices to the “base” PFE platform. The dry materials (fabric) are loaded and fed in the same manner as prepreg materials, with the exception of the take-up system, as no protective films need to be removed. The profiling and waste removal stations are the same. However, the additional attachments include resin supply and feed systems, which provide for temperature control as well as the mixing of materials. The feed system provides resin to both sides of the material as it is being fed, just prior to placement, with the ability to “meter” volume and viscosity on each side. The addition of a flexible “squeegee” attachment to the suspension system allows for the spreading and “bleeding” of the resin through the material as it is being placed. Excess resin is “recycled” and metered so that the “new resin supply” may be cut back accordingly, in “real time”, so as to “balance” the resin/material feed.
Control
pass. For example, when two planes or surfaces come together, creating a corner, as in the case of a box, where the ‘side’ meets the ‘botom’. The PFE will place both the ‘side wall’ and the ‘bottom’ at the same time. In addition, “cross configurations” may be employed where the layup of unidirectional material could be placed at a right angle to a linear path of woven material in a given layup, thus eliminating an additional pass by the system. Configurations, such as these examples, could be designed into the “process” at the engineering level. The PFE systems come fully integrated to the robot and system configuration is selected by the customer. These systems are designed to be expanded as the need arises.
Wet layup For “wet” layup applications, such as those associated with boat building, among others,
Control to the PFE system is PC/PLC based incorporating both digital and analog inputs and outputs to monitor speed/feed criteria in concert with the robot movement. This is done by means of a proprietary “chip” which “negotiates” with both of the devices in real time so that they work together seamlessly. This hardware “rides” with the PFE device so that when exchanges of tooling occur, each device carries what it needs to perform the tasks it is assigned. Patterns that the profiling system will perform are simply fed to the system offline and verified through “offline programming software” that is supplied with the robot system. The operator has the option to program on the floor or call predefined programs from archived files. These “turn-key” systems are typically equipped with an human machine interface (HMI) that includes a customized touch screen, incorporating icons that allow the operator to test and cycle individual aspects of the system. The HMI is located adjacent to the tool crib, as the operator will use it to communicate with individual PFE devices or other “end-of-arm” tooling located within for setup and testing prior to telling the “robot system” that a particular device is ready for use.
157
Robotic “layup” of composite materials
Assembly Automation
David Groppe
Volume 23 · Number 2 · 2003 · 153–158
The HMI also monitors what is happening during layup, which may include cameras that monitor aspects of the layup regime to ensure quality and provide documentation for various uses internally. The screen can be tailored to the customer’s requirements, based upon exiting processes. Training for the operator personnel is part of the packaged system.
the demand for faster, accurate, methods for the layup of composite materials, PFE technology fills these demands economically. Working in concert with track or gantry systems that may include “positioning” devices within the cell, rapid, accurate, automated layup can be realized. For further information on PFE technology and the layup systems please contact: David Groppe, Director, Composite Systems, Inc. PO Box 509 - 1653 Fourth Street, Arnold, 95223-0509, CA, USA. Tel: (+1) (209) 795-6977; Fax: (+1) (209) 795-3164; E-mail:
[email protected]; Web site: www.compositemfg.com
Conclusion Without question, the use of robots to perform arduous and even dangerous tasks, precisely, has been well established. With
158
Feature Advances in resistance welding for body-in-white Brian Rooks
The author Brian Rooks is an Associate Editor for Assembly Automation. Keywords Welding, Assembly Abstract This paper describes the developments in the control of spot welding from Bosch Rexroth with particular reference to body-in-white applications. The Bosch Rexroth MF system uses 1 kHz rather than conventional 50 Hz in the control of spot welding and DC current rather than AC at the weld gun. The several benefits of this arrangement are discussed including lower power and energy losses, lighter cabling and a more compact weld transformer. Also described is a new ultrasonic adaptive control system developed by Bosch Rexroth which enables the growth of the weld nugget to be monitored and recorded for traceability.
Electronic access The research register for this journal is available at http://www.emeraldinsight.com/researchregister The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0144-5154.htm
Assembly Automation Volume 23 · Number 2 · 2003 · pp. 159– 162 q MCB UP Limited · ISSN 0144-5154 DOI 10.1108/01445150310471400
Most car bodies today are a monocoque construction of mild steel sheet joined together by a multitude of spot welded joints. It is a process that has proved successful for the past 50 years or more and in which the automobile manufacturers have accumulated a vast deal of experience. This is as well because the joining process itself, which uses AC mains frequency (50 Hz) current to create the spot weld, is not particularly efficient. Any lack of precise fit up, the presence of foreign materials such as oil, the propensity for sparking and the lack of control of a 20 ms cycle mean that a percentage of welds are not fully made, so that the number of spots have to be over-specified to ensure the desired strength of the monocoque. Another problem for the conventional spot welding process is the trend towards more exotic materials including aluminium and high strength steels that are difficult to join by resistance welding. However, the resistance welding process is “fighting back” through developments by Rexroth Bosch in the control of spot welding using medium frequency (1,000 Hz) DC rather than the conventional 50 Hz AC. Not only does this process creates more consistent quality spot welds, it is also more power efficient and saves energy consumption and is able to join a variety of exotic materials. A further benefit of the MF system is the short 1 ms cycle that allows real-time adaptive control of the spot weld growth. To exploit this benefit, Rexroth Bosch has developed an unique ultrasonic sensing system that enables the growth of the weld nugget to be accurately controlled, ensuring all spots are of the correct size and strength. The Rexroth Bosch MF welding control system is a product of the Electric Drives and Controls Division of Bosch Rexroth AG, which was formed in May 2001 by the merger of Bosch Automationstechnik and Rexroth, and is a wholly-owned subsidiary of Robert Bosch GmbH – in UK, the headquarters of Bosch Rexroth Limited is at St Neots in Cambridgeshire. The company has been co-operating with most European car manufacturers for more than 50 years and is In preparing this article, the author appreciates the help and advice given by Andrew Davies, Branch Manager, Handling and Car Body, Bosch Rexroth Limited, Broadway Lane, South Cerney, Cirencester GL7 5UH, UK. Tel: +44(0) 1285 863 006; E-mail:
[email protected] 159
Advances in resistance welding for body-in-white
Assembly Automation
Brian Rooks
Volume 23 · Number 2 · 2003 · 159–162
currently installing around 3,500 resistance welding control systems per annum. It has used this experience to develop the PS 6000 series MF weld control system that is now in use by several European car manufacturers, including ChryslerDaimler on the new C-class, BMW on the three-series and Volvo. In UK, the new Mini and new Range Rover are both 100 per cent spot welded with the Rexroth Bosch MF system.
Table I Comparison of the power balance between AC and MF DC weld control systems Welding task Current 1 Power 1 Current 2 Power 2 Power balance
DC welding In a conventional system, single phase mains frequency is applied to the thyristor controller for input to the primary of the welding transformer whose secondary output current is 50 Hz AC. By contrast, the Rexroth MF system takes the three-phase mains supply as input to an inverter in the PS 6000, which generates a 1,000 Hz voltage input to the primary of the welding transformer (Figure 1). The output is rectified in the transformer secondary circuit to create a DC welding current. One of the most important advantages of the DC welding current is the zero reactive resistance, from which stems two practical benefits. First, the weld current is lower than with AC, which suffers inductive losses, resulting in lower power consumption (Table I) and lighter cables. Also, the higher 1 kHz frequency results in a smaller weld transformer. Consequently, installation costs for a body-in-white assembly line with several 100 robots each equipped with a welding transformer and cabling, are significantly lower than with AC welding equipment. A smaller, lighter transformer also reduces the load carried on the robot arm, raising Figure 1 The Rexroth Bosch MF weld control system uses medium frequency of 1 kHz rather than the 50 Hz of conventional AC weld systems
AC
MF
15.23 kA 67.5 kW 10.31 kA 28.9 kW
15.2 kA 54.09 kW 10.29 kA 25.38 kW
100 per cent
80-88 per cent
the possibility of a smaller capacity and therefore lower cost robot. Just as significant is the energy saving, which in Rexroth Bosch’s experience can be up to 30 per cent (Table II). This means a lower electricity bill and even a possible reduced investment in an electricity sub-station. Also important is the reduced peak current demand on the mains due to unity power factor rather than the 0.3-0.4 associated with AC weld systems and the need to balance mains loading. In some installations this has led to sharing a common power bus for weld and control systems. The second benefit of the zero reactive losses of DC is that different lengths of cables, different weld guns and the changing depth of electrode penetration have no influence on the welding current. This means that with DC, the parameters are standard for every gun, unlike AC, for which adjustments have to be made to compensate for the different gun-cable arrangements on the line.
Faster weld times The steady weld current and the 1 kHz regulation rate of the Rexroth MF system allow accurate metering of the energy and a steady growth of the weld nugget, compared to the fluctuating heat generation when Table II Comparison of the energy balance between the AC and MF DC weld control systems Welding task
AC
MF
Current 1 Weld time 1 Power 1
15.23 kA 200 ms 67.5 kW
15.2 kA 170 ms 54.09 kW
Energy per spot 1
3.77 £ 102 3 kV A h
2.55 £ 102 3 kV A h
Current 2 Weld time 2 Power 2 Energy per spot 2
10.31 kA 280 ms 28.9 kW 2.25 £ 102 3 kV A h
10.29 kA 240 ms 25.38 kW 1.69 £ 102 3 kV A h
100 per cent
68-75 per cent
Energy balance
160
Advances in resistance welding for body-in-white
Assembly Automation
Brian Rooks
Volume 23 · Number 2 · 2003 · 159–162
welding with the conventional AC system. The result is better control of the nugget growth giving a more consistent weld joint and a weld time approximately 15 per cent shorter, says Rexroth Bosch. Shorter weld times improve productivity with a higher rate of spot application per robot, which could reduce the number of robots needed on a line. In addition, the more uniform heat generation is “gentler” on the electrodes, and can reduce gun tip wear when welding some metals. Another problem of spot welding with an AC current is sparking. Although visually and photographically attractive, sparking does have a negative effect on the quality of the weld. It is caused when the spot becomes too hot too quickly and molten metal is ejected into the air, a risk that is greatest at the current peaks in the AC waveform. The result is that the weld joint is not fully made or even not made at all. There are no peaks with a DC current (Figure 2); thus sparking is much less likely and the chances of a poor weld nugget are drastically reduced. The growth rate of a spot weld nugget is dependent on three parameters, the pincer pressure of the gun calliper, the weld time and the weld current. Fixing the first two allows the nugget size to be controlled by regulating the weld current. But, a spot weld is completed in about 80 ms so that to control a 50 Hz frequency current, which has a 20 ms cycle time (Figure 2), would be impracticable. However, there is no problem with the 1 ms cycle of the MF system, which has allowed Rexroth Bosch to develop the PSQ ultrasonic weld quality assurance system.
Ultrasonic weld control In the PSQ system, an ultrasonic sender mounted on one of the gun electrodes transmits a longitudinal wave through the weld nugget for receipt by a receiver on the other electrode (Figure 3). As the pool of weld in the nugget grows, the received ultrasonic signal falls proportionately, and is used to close the loop in the MF weld control. The electronic card for the PSQ system sits in one slot of the PS 6000 MF system. The ultrasonic signal-to-time characteristic to achieve an optimum weld nugget growth, which is unique for each type of joint material combination, is determined during set-up, and then used as reference during each production weld. The PSQ monitors the received signal and uses the deviation to adjust the weld current so as to maintain the optimum characteristic. Only the weld nugget is being monitored so that any deviations due to fit-up or contaminants in the joint are irrelevant, ensuring that each and every weld is consistent and of the optimum quality. The Rexroth Bosch PSQ is uniquely a real-time adaptive control system for the resistance welding process. Also controlling the nugget growth reduces the potential for any sparking, and while it cannot prevent expulsion caused by bad fit-up or edge welding, the PSQ system does record the event for later visual verification of the weld. “One of the most notable things in walking along a body assembly line using the PSQ system is the total absence of flying sparks”, comments Andrew Davies, Manager – handling and car body at Bosch Rexroth Limited in Cirencester.
Figure 2 Comparative wave forms of the MF DC weld control system from Bosch Rexroth and the 50 Hz AC weld system Figure 3 Schematic of Rexroth Bosch PSQ weld quality assurance system
161
Advances in resistance welding for body-in-white
Assembly Automation
Brian Rooks
Volume 23 · Number 2 · 2003 · 159–162
Non-destructive testing The 100 per cent quality assurance of the PSQ systems has a number of key benefits. Perhaps the prime one is its non-destructive nature, eliminating the need to regularly tear bodies apart to check on the quality of welds. Not only does this waste materials and destroy added value in the car bodies, but also is labour intensive with an average of 60 people per plant employed just to chisel car bodies apart. Redeploying this number of people plus the time savings are valuable cost reduction benefits of the PSQ system. Because of the uncertainty of the nugget quality from conventional spot welding, body designers increase the number of spots by about 10 per cent to ensure the desired body strength. It adds time to the production cycle and to the number of robots on the line to apply the extra welds. With the PSQ system, only the minimum number of spots to achieve the optimum strength is required, with no need of additional spots “just in case”. Hence, it lessens the investment in robots and improves output. Another positive is the ability to collect and store the signal from every weld on every car body, and with approximately 1 MB data generated per robot per shift, it is perfectly manageable. This 100 per cent traceability feature has major implications for the future. It could, for instance, be used in a car crash investigation to assure the integrity of the body construction. And, with the growing demand to legislate for ever more safer motoring, 100 per cent assurance systems may well become legally imposed to ensure every body is built to the requisite standard. Other developments in the body assembly include the increasing use of materials more exotic than the ubiquitous mild steel. High strength steel is one such material because it can be used in thinner sections, saving in weight and therefore fuel consumption as well as being stronger than mild steel. But, it is not an easy material to weld with AC and because of the thinner section it is difficult to cut for inspecting the weld. It is, however, not a problem to weld using the Rexroth Bosch MF system because of the constant DC current, and with the PSQ system the quality of the weld is assured, so there is no need for destructive testing. Similarly, the MF system easily welds aluminium, a notoriously difficult material to
join with conventional resistance welding. Also, relatively new to body construction is the use of laminated materials, in which a carbon-impregnated polymer is sandwiched between steel layers, for use in areas such as the car dash to prevent “booming”. Even this material can be welded with the MF system, but impossible with AC, and monitoring the growth of the nugget with the ultrasonic sensing is particularly advantageous.
A-post welding Application of the Rexroth Bosch PSQ system is increasing rapidly in European car plants. One of the first applications “in vengeance”, to quote Mr Davies, was to weld an A-post made up of multiple sheets of thin material for a German car manufacturer. “The problem with this part was in ripping it apart to test the weld. Because of the multiple thin layers, it was tearing and the weld nuggets were not visible for examination. Eliminating the need to destructively test the A-post but still being assured of the weld quality, rapidly brought home to this manufacturer the benefits of our systems”, says Mr Davies. While the Rexroth Bosch PS 6000 MF weld control system has been available for several years and now has an installed base of over 10,000, the PSQ quality assurance system is new and is only now being introduced onto the broader market. However, its cost effectiveness has already been proven at some German car plants and over 8 million spot welds have been applied using the PSQ system. It is also retrofitable to any installed PS 6000 system. While most PS 6000 systems have been installed for spot welding car bodies, it is also applicable to other resistance welding processes including projection and seam welding. For instance, seam welding with the MF system supports line speeds approximately 30 per cent faster than with AC, and is proving to be particularly popular in USA. The majority of applications of the Rexroth Bosch MF weld control system are in automotive, but it is also used in aerospace, white goods and for welding electrical cabinets and has potential in any high volume production. While weld integrity may not be so critical outside automotive and aerospace, the energy saving potential of the MF system is valid in other industries and this alone could easily justify investment in the medium term.
162
Introduction
Feature IMTS leads with technological innovation Dick Bloss
The author Dick Bloss is an Associate Editor for Assembly Automation, Cleveland Ohio, USA. Keywords Assembly, Motion control, Systems Abstract Innovation drives suppliers even as demand is soft. Modular assembly stations, compact part machining/ turning systems, linear motor powered motion and smart end effectors lead the way as the industry drives ahead to provide added benefits to users and system integrators. Electronic access The research register for this journal is available at http://www.emeraldinsight.com/researchregister The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0144-5154.htm
Assembly Automation Volume 23 · Number 2 · 2003 · pp. 163– 165 q MCB UP Limited · ISSN 0144-5154 DOI 10.1108/01445150310471419
While trade show attendance may be off, in some cases way down, there is no shortage of new innovations from exhibitors. At the recent Chicago International Manufacturing Technology Show, IMTS, attendance was down about 25 percent from previous peak numbers but exhibitors were out in force displaying new technologies and products. One exciting concept exhibited by several machine tool builders is a rotary multi-station machining and turning system. Some are calling the new technology a rotary transfer manufacturing center. Multi-axis machining centers are positioned around an indexing multi-spindle center head. Manufacturers showed designs ranging from six position to 18 or more. One position is allocated to load and unload, frequently performed by a robot to eliminate any chance of lost time waiting for the operator. The machines are designed to manufacture components of about 10 cm (4 in.) cube or smaller which need extensive machining, drilling and turning. The component is mounted up on a fixture on a series of lathelike chucks, one for each position on the machine. At each work position, the spindle can rotate the part at the required RPM if turning is required. For the machining and drilling stations, the rotary fixture will position the component in the proper alignment for required operations. The rotary transfer concept is ideal for machining/turning components such as caliper bodies, valve blocks, cylinder bodies and similar parts widely used in subsequent automated assembly operations. Machines are offered with either the workpiece rotating in a horizontal plane or a vertical plane. Giuliani, a division of IGMI S.p.A (Italy) offers Proflex, vertical part rotation with a number of modular work holding spindles and turning, milling and drilling stations. Their flexible modular design permits stations to be changed or added as work requirements and workpiece needs change. In the past, to manufacture such components, very typical in automated assembly, a number of different machines might have been required. Each machine would have needed it’s own fixturing, there would be the need to load and unload different machines, possible delays in
163
IMTS leads with technological innovation
Assembly Automation
Dick Bloss
Volume 23 · Number 2 · 2003 · 163–165
logistics for moving work-in-process from machine to machine in the factory, all leading to extra cost and variations in parts made on the various machines. The rotary transfer manufacturing system provides in compact square-like floor space the functions previously needing a row of machines requiring much more factory floor space. Another innovation in machine design is the reconfigurable machine. The University of Michigan, the NSF Engineering Research Center for Reconfigurable Manufacturing Systems and several equipment builders exhibited concepts of reconfiguration in machining as well as assembly systems. ATS Automation Tooling Systems, Cambridge, Ontario offers the Supertrak programmable pallet conveyor. The Supertrak includes distributed optical position sensing systems, distributed object oriented servo controls and virtual pallets. When a pallet arrives at a section stop, it notifies us about which part is onboard; and when the cell is finished it gives instructions on the next stop with no supervisory latency. Operating systems have achieved pallet velocities higher than 2 m/s. A ten-part medical device can be assembled at one-half second cycle times. As work cells retire from a given program, they can be quickly reconfigured to assist in the ramp-up of other production runs. A new face in the automated assembly system integration market in North America is Meikle Automation of Kitchener, Ontario, Canada. Founded in 1994, the firm has grown through internal expansion and acquisition to now have a staff of 180 located in five facilities totally 7,500 m2 under roof. Meikle focuses on automated assembly systems for the medical, automotive, electrical, microelectronics, consumer products and photonics industries. A recent acquisition of ElectroMechanical Specialties of Sanford, North Carolina, launched the firm’s entry into a local presence in the United States. A complete printed circuit board assembly system in modular form is available from Siemens Dematic, formerly known as Mannesmann. Modules offered include conveyors, walk through gates, inverters, destackers, turntables, buffers and magazine loaders/unloaders. For the electro-mechanical assembly system integrator, Siemens Dematic offers a modular asynchronous conveyor system handling payloads to 1,000 pounds.
Branching out, Siemens Dematic has entered the systems integration business through the acquisition of the Birmingham, Alabama based Automation Technologies Industries (ATI). The objective is to utilize the former ATI as a focal point sole-source technology center for serving printed circuit board and other assembly automation markets. The new name for the operation is Siemens Dematic Assembly Systems. Bosch Rexroth introduced a new generation of rod-less cylinders, RexMovere (Plate 1) featuring an oval-shaped piston, three choices of air connection location and high-speed capability to 16 feet per second. For frame construction, Bosch Rexroth has added 11 new, high-strength profiles with 40 mm and 50 mm cross sections. The new profiles feature 10 mm t-slot design and a large center bore to handle heavy loads in tough applications. The extruded aluminum framing system and accessories provide users with new choices in modular framing designs with flexibility for rapid design, assembly and reconfiguration when applications require. For clean room applications, Bosch Rexroth introduced the PSK precision positioning modules. Certified for use in Class 10 applications, the units feature unique sealing strips, positioning repeatability to 0.005 mm, payloads to 44,760 N and speeds to 1.6 m/s. For compact linear motion applications, Bosch Rexroth has added a new toothed belt drive system option to its STAR compact module program. The CKR models feature speeds to 5 m/s with smooth running characteristics and lengths to 5.5 m. The drive motor may be front-mounted or side mounted. Plate 1 Bosch Rexroth RexMover rod-less pneumatic cylinder
164
IMTS leads with technological innovation
Assembly Automation
Dick Bloss
Volume 23 · Number 2 · 2003 · 163–165
Electric actuators continue to grow in popularity. THK America introduced the Type ELA electric actuator, combining ground ball screw, electric motor ball slides and angular bearings. With all rolling surfaces as opposed to sliding, the units are more energy efficient and better performing than hydraulic or pneumatic power actuators. An electric drive also eliminates oil leaks or air mist emissions that requires less maintenance. For applications needing linear motor actuators, THK introduced the RDM coreless linear motor design. Magnets are encapsulated in a stainless steel tube which moves inside drive coils encased in the movable carriage. Several variations of the RDM product line offer a choice of a long stroke, a rigid frame for beam and minimal speed ripple applications or compact highprecision features. For part gripping applications where axial and lateral misalignment can be a concern, Robohand Inc., a unit of DE-STA-CO Industries, offers a new CV and CH lines of end effectors. The compliance devices detect and adjust for misalignment. Compliance devices can improve repeatability and reduce binding issues from offset loads. For machine load/unload rotary actuator requirements, Robohand is the first company to offer a fully assembled gripper and sensors ready for machine installation by the customer. The rack and pinion gripper features zero backlash, low overall weight, internally routed pneumatic and sensor electrical lines and a direct connection mounting pattern eliminating adapter plates. Parker Hannifin has acquired a 40 percent interest in the Japanese hydraulic cylinder maker, Taiyo, Ltd. The partnership expands Parker’s presence in the Japanese market specifically and enhances both companies reach into industrial markets worldwide. The Parker modular structural framing system business, a unit of Automation Actuator Division, has been greatly enhanced
by the recent acquisition of the entire operations of IPS Industrial Profile Systems. IPS adds nine local service centers across North America where customers can obtain factory designed and assembled products quickly. Parker also introduced a new rod-less linear actuator, the ERV Series, compatible with the Parker modular structural framing system. External dimensions and t-slot design is compatible with the 56 mm and 80 mm modular structural aluminum profile products. An external carriage option is available. Picking up difficult to grasp components just got a bit easier with the new Schunk servo controller gripper, PT-AP70. Sensors coupled with DC servo drive motor and control electronics enables the PT-AP70 to almost duplicate the “light” touch of human fingers. The gripper can pickup difficult items such as thin glass, foam and sand molds without damage. The controller is available with communications options such as CANbus, RS232 or Profibus-DP protocols for integration into overall system operations. A new three finger gripper from Schunk offers longer fingers and greater load capacity. Grip forces to 18,000 N and strokes to 50 mm are available. For more information visit the following Web sites: . THK – www.thk.com . Bosch Rexroth – www.boschrexrothus.com . Siemens Dematic – www.siemensdematic.de . Giuliani – www.giulianico.com . Meikle Automation – www.meikleautomation.com . Robohand – www.robohand.com . Parker Hannifin – www.parker.com . ATS Automation Tooling Systems – www.atsautomation.com . Schunk – www.schunk-usa.com
165
Feature Robotics and assembly automation at TEAM Brian Rooks
The author Brian Rooks is an Associate Editor for Assembly Automation. Keywords Robots, Machine vision, Packaging, Consumer goods, Automotive Abstract Ten events made up the new total engineering and manufacturing (TEAM) exhibition, one of which, factory automation (FA) is the focus of this article. A main sponsor of FA was Bosch Rexroth, and a report is given on the state of the company since the acquisition of Rexroth by Bosch. Descriptions are given of a new range of Rexroth Bosch articulated arm robots and an application of the company’s Turboscara robots assembling lambda exhaust control probes. Other exhibitors featured are ABB with a new heavy duty robot, Orwin Automation who demonstrated an automatic bagger, part of its flexible intelligent packaging systems (FIPS) range and DT Assembly and Test who promoted its project and value engineering expertise. Finally, reference is made to GE Panametrics hand held ultrasonic weld testing system. Electronic access The research register for this journal is available at http://www.emeraldinsight.com/researchregister The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0144-5154.htm
Factory automation (FA) was one of ten events or sectors that made up the inaugural total engineering and manufacturing (TEAM) exhibition held at the Birmingham NEC in November 2002. FA was primarily concerned with the hardware and the dedicated software of automation on the shop floor while the IT and computer integration of factory systems was covered in computers in manufacturing (CIM), another TEAM event. In the main, it was the components of automation, such as bearings, conveyors, drives and motors and linear actuators that dominated FA, but there was a smattering of advanced technology in the form of assembly systems and robotics to wet the appetite. A company with activities in many of the events of TEAM is Bosch Rexroth and a main sponsor of the exhibition. Over the past 12 months the group, which employs 26,000 people worldwide and revenues of approximately e3.86 billion in 2001, has continued its rationalisation and re-branding following the acquisition of Rexroth from Mannesman in 2001 by Bosch. Rexroth is now the brand name of all its products in hydraulics, pneumatics, electric drives and motors, linear motion and assembly technology and robots, replacing such well known names such as Indumat, Mecman, Bosch AT and Star. Speaking at TEAM, Paul Cooke, Managing Director of Bosch Rexroth Limited, said, “The sales forces of the two companies (Rexroth and Bosch AT) have been combined and it has been a very successful integration. We are now achieving rewarding results and are on track for “one company to be better than two”. In UK, the headquarters of the group is at St Neots in Cambridgeshire, which is also the centre for industrial and mobile hydraulics. The sales headquarters and engineering activities for electric drives and controls, linear motion and assembly technology, including robotics, and pneumatics is at Cirencester, a facility set up just 2 years ago.
German robot leader
Assembly Automation Volume 23 · Number 2 · 2003 · pp. 166– 171 q MCB UP Limited · ISSN 0144-5154 DOI 10.1108/01445150310471428
Bosch’s involvement with SCARA type robots goes back 30 years when it introduced a hydraulic-drive device for press feeding. The first Bosch SCARA to be produced in quantities, the electric drive SR800, was 166
Robotics and assembly automation at TEAM
Assembly Automation
Brian Rooks
Volume 23 · Number 2 · 2003 · 166–171
launched in 1984 but was mainly used in its own plants. There then followed the SR60/SR80 in 1991 and in 1998 came the “Turboscara” S6/S8 with its PC-based control, which established Bosch as the German market leader in swivel arm robots. Each development has seen an increase in cycle speed so that the latest open control technology S series robots are capable of 490 ms cycles compared to 2 s of the SR800. Now the company has added to its range the Rexroth AR6/AR8 articulated arm robots (Plate 1), which have identical control technology to the S4/S6/S8 Scaras. The AR6 with a 650 mm working radius and AR8 with a 850 mm range, both handle loads up to 5 kg at high speeds and position to within ^ 0.02 mm. Other features include an absolute encoder system and the option of multiple communications and table or ceiling (inverted) mounting. The Rexroth Scara robots are often at the core of systems built by Rexroth’s handling systems and assembly technology group. An example of their application for precision assembly is in the manufacture of Bosch lambda probes used for exhaust gas control in vehicles. These probes consist of a flat rod-shaped sensor around which the exhaust gas and reference air flow. The gas and air are separated by seals and Rexroth SR6 Turboscara robots equipped with vision perform the assembly of these into the correct precise position (Plate 2). A seal pack consists of one sealing disc sandwiched between two discs made of steatite (a soft porous ceramic material).
The outer dimensions of these discs correspond to the cylinder-shaped housing of the lambda probe plus a central opening shaped to accommodate the sensor. The discs are a precise fit over the sensor so while a significant force is needed to push the pack into place, the nature of the soft ceramic material limits the force that can be applied without damaging the discs. The precision of the robots ensures that this assembly is performed safely.
Image processing The discs are fed by clock-pulsed conveyor belts through a simple escapement and their position and orientation identified using an image processing system that performs a 2D centre of gravity analysis while a second vision system determines the position of the sensor in the workpiece nest. Evaluation of the image data using gray scale analysis occurs in parallel with the robot’s motions. The positional data is used by the robot to pick up, in sequence, a steatite disc, a sealing disc and another steatite disc in the specially developed gripper (Plate 3).
Plate 1 The new range of Rexroth Bosch AR6 and AR8 articulated arm robots
167
Plate 2 Assembly stations for Bosch lambda exhaust gas control probes employ two Rexroth Bosch SR6 Scara robots and vision systems
Robotics and assembly automation at TEAM
Assembly Automation
Brian Rooks
Volume 23 · Number 2 · 2003 · 166–171
Plate 3 The gripper specially developed for handling the disc packs for the Bosch lambda probes
Plate 4 RTS uses the ABB IRB 140 six-axis robot for handling loads up to 5 kg and for increased dexterity in its PixCell concept
It then forces the pack over the sensor element and into the required location. The only principal robot manufacturer exhibiting at TEAM was ABB, albeit jointly with RTS Flexible Systems International with whom it has a strategic partnership. RTS provides robots and vision systems in, to quote, “the most demanding of applications, no more so than for the fast-paced, ever changing consumer goods industry in which it (RTS) is currently investing heavily.” RTS has recently incorporated both ABB’s IRB 340 Flexpicker and IRB 140 robots into its PixCell concept that offers high speed vision guidance picking for a range of applications in food, plastics, medical devices and pharmaceutical. In picking and sorting operations for a major sweet manufacturer, a RTS vision system plus IRB 340 working at 150 pieces per minute was able to match the best manual, sorting and quality inspection rates. For higher payloads RTS will employ the six-axis IRB 140 (Plate 4). David Bradford, RTS Business Development Manager, comments, “The six-axis robot is an effective alternative to the FlexPicker in that, it gives just 70 picks per minute but its 5 kg payload
often allows the picking of more than one product at a time and its six axes gives increased dexterity. We recently specified the IRB 140 for the multiple picking of a product that had to be turned through 908 before being placed into a cardboard box.” A new fast version of the IRB 140, the IRB 140T (turbo), was demonstrated at the show by ABB. It is approximately 10 per cent faster than the standard 140 having been clocked at 0.77 s for the 25-300-25 mm “goalpost” work cycle. It can be floor, wall or ceiling mounted and is available also in foundry (IP67) and clean room (class 10/ PI67) version.
Heavy duty robot At the other end of the scale ABB exhibited at TEAM its new IRB 7600, the first in a series of heavy duty six-axis robots (Plate 5). There are two versions, one with a reach of 2.3 m and 500 kg payload and the other with 2.55 m and 400 kg, respectively. A wrist torque of 3,010 Nm makes the IRB 7600 ideal for handling heavy components such as car body sides, framing fixtures and complete engine assemblies. With the capability to handle up to 500 kg, safety is a key issue and ABB has addressed this by developing “Active Safety” for the IRB 7600. This collection of software features designed to protect both the robot and its operators, includes active collision detection (reducing collision forces by up to 30 per cent), self-tuning performance (reducing cycle times by adapting to the true dynamic payload), a service information system (supervising the machine’s motion and load while optimising
168
Robotics and assembly automation at TEAM
Assembly Automation
Brian Rooks
Volume 23 · Number 2 · 2003 · 166–171
Plate 5 ABB has introduced the new IRB 7600 range of heavy duty robots able to handle up to 500 kg
its service requirements) and ABS brakes (simultaneously controlling braking and maintaining the programmed path as well as allowing rapid recovery). A regular exhibitor at automation shows also to be found at TEAM was Orwin Automation, a precision manufacturer of special purpose automation machinery with a complete in-house design and build facility. Formed in 1983, the company is now part of the Anglia Autoflow Group, employing 80 staffs and occupying a 3,000 m2 factory in the north east of England.
The bagger has a conventional vertical bagging arrangement for bags up to 200 mm wide and 255 mm long. Using electro servo drives eliminates the need for plant air but to quote, “is still competitively priced”, and enables the unit to bag at rates up to 50 parts per minute. A touch screen message display allows the user to optimise the unit for controlling the bag length, film feed rate, acceleration and time delays. It makes for easy changeovers from one bagging specification to another. Built-in to the design is provision for a number of peripheral units including laser and ink marking units for printing logos, product ID and other relevant textual information onto the bags. A range of feeders can also be fitted, including hopper and bowl with escapements, elevating hopper and step feeders. The bowl and step feeders can also be equipped with vision systems for on-the-fly inspection for “foreign bodies” as well counting for control of bag quantities, at up to 100 pieces per second. Another use of vision is to separate products off the feed conveyors into different delivery positions for distribution to a maximum of six bagging units. Adding these standard peripherals is claimed to be “fast and simple” and when connected the whole system operates as a single machine. The Orwin bagger may be used as a stand-alone unit for bagging a wide variety of hardware parts in combinations or as singles. The company is also supplying it as part of an automation package with the bagger integrated “end of line” to assembly machinery. Being manufactured in stainless steel allows its application in a range of industries including food and pharmaceutical.
Project capabilities
Auto bagging unit Until recently Orwin’s business was solely the design and manufacture of customised automation assembly and test solutions for industries ranging from automotive through food and cosmetics to electronics. Now it also offers a standard product range branded as flexible intelligent packaging systems (FIPS), at the centre of which is an auto bagging unit, which was demonstrated at the show.
Food and pharmaceuticals are areas that the company DT Assembly and Test – Europe is expanding into as a means to smooth out its essentially cyclic business in the automotive industry. The company used the exhibition to promote its project capabilities under the heading “Build to Print”. Started in 1934 as Hartridge and with an international reputation for fuel and exhaust systems test equipment, Assembly and Test is now a division of DT Industries that operates
169
Robotics and assembly automation at TEAM
Assembly Automation
Brian Rooks
Volume 23 · Number 2 · 2003 · 166–171
on 14 sites in USA. Still operating in Buckingham, where it now has 2,500 m2 of manufacturing space and employs 150 people, automotive test equipment is still an important part of the business, accounting for 40 per cent of sales, and responsible for the testing of 110 million fuel injection systems per annum. The remainder of the business is concerned with both product and project assignments from sub-contract fabrication and finishing through subassembly work to full turnkey ranging in size from £2,000 to £20 million, and delivered to anywhere in the world. Recent multi-million projects include one for Delphi Systems installed in China and two lines to Siemens in South Carolina, USA. Assembly and Test is capable of taking a project from the earliest concept stage to installation and commissioning. Machinery can be built, tested and debugged with EMC testing and provision of a CE mark as part of the service. As well as project management the company has a depth of value engineering expertise, which it demonstrates by achieving significant cost reductions for its customers. It claims regular savings of 25 per cent and often more. It believes it can offer this level of saving and improved competitiveness to a variety of industries and not just automotive, including converting, pharmaceutical and consumer goods.
control instrumentation, was acquired by GE Power Systems in the summer of 2002 (Figure 1). The Panametrics Epoch IV exhibited at TEAM is a handheld probe that, with trained personnel, enables a spot weld to be inspected in a matter of seconds, and a claimed 95-97 per cent correlation with destructive testing is not uncommon. The probe incorporates the ultrasonic emitter and receiver and is of a size chosen to match the dimensions of the weld joint, a cast nugget spread through the two forged sheets being joined. The signal reflected from the joint differentiates amongst several conditions including a perfect oval-shaped nugget, an undersized nugget, a stick weld that has made limited penetration and a non-weld (Figure 2). For smaller components that can be bench top tested, Panametrics has developed the FreeScan portable spatial hand scanner. This 5-axis jointed arm device incorporates a holder for the Epoch
Figure 1 Ultrasonic testing of a good spot weld. (a) Transducer schematic and (b) signal waveform (source: GE Panametrics)
Ultrasonic weld testing One of the most common production processes in automotive is spot welding and yet is one of the most difficult to test for quality. Traditionally, welding quality assurance has involved destructively tearing car bodies apart and visually examining the weld joint, which if perfect is a nugget that completely encompasses both sheets. However, this expensive and labour intensive method is not ideal as all it does is, test that the equipment is functioning correctly. Individual welds on a production body, and therefore untested, may be faulty due to “local” conditions for that weld such as poor fit up or contaminants in the joint. The solution to the problem is non-destructive testing using ultrasonics, in which both Bosch Rexroth and GE Panametrics – Panametrics, a leader in high-technology ultrasonic testing equipment and process 170
Robotics and assembly automation at TEAM
Assembly Automation
Brian Rooks
Volume 23 · Number 2 · 2003 · 166–171
Figure 2 Ultrasonic test of an undersized weld. (a) Transducer schematic and (b) signal waveform (source: GE Panametrics)
probe that enables scanning of a component within a 3D workspace of 1.27 m. As the probe is moved over the object under test, the ultrasonic signal is registered while the arm provides the spatial data through the FreeScan’s serial port, to an accuracy of 0.23 mm.
Exhibitor details Bosch Rexroth – contact: Stuart Forrester. Tel: 01480 223200; E-mail: stuart.forrester@ boschrexroth.co.uk; Web site: www. boschrexroth.co.uk ABB – contact: David Marshall. Tel: 01908 350396; E-mail: david.marshall@gb. abb.com; Web site: www.abb.com RTS – contact: Susan Jones. Tel: 0161 777 2000; E-mail:
[email protected]; Web site: www.rtsflexible.com Orwin Automation – contact: Ted Chatt. Tel: 0191 4177092; E-mail: t.chatt@ autoflowpackaging.com; Web site: www. orwin.co.uk DT Assembly and Test – Europe – contact: Mark Richardson. Tel: 01280 828550; E-mail:
[email protected]; Web site: www.dtindustries.com GE Panametrics – contact: John Skidmore. Tel: 01709 836115; E-mail: “John”,pan-ndt@ btclick.com; Web site: www.panametrics.com
171
Feature A novel indexing mechanism for paper cutting machines Rajanish K. Kamat
The author Rajanish K. Kamat is a Lecturer in Electronics, Department of Physics, Goa University, Goa, India.
1. Introduction The paper industry very often requires cutting machines for various purposes. To achieve high productivity in these machines, the cutting table is moved by using motors with the help of lead screws. However, the cutting action is always manual for safety purposes. The author has observed that, while processing a job, more than 70 per cent of time is required to set the machine. Only 30 per cent or less time is utilised for cutting. In addition to the under utilisation of the machine, there is considerable wear and tear of the lead screw. This is because of the heavy load and the small portion of the screw which is used again and again leading to its fast wearing out.
Keywords Paper industry, Cutting
2. Inaccuracies
Abstract
The above mentioned wear and tear of the lead screw directly affects the time required to process the job on the machine. More time is required for adjusting the exact cutting position. The backlash worsens the situation when the job is to be taken back slightly. The state of the art machines use only a single motor for adjustments. The usage of the single motor leads to speed-accuracy trade-off. High speed is possible at low accuracy and vice-versa. The revolutions plus angle measurement of the lead screw does not give the requisite accuracy. The resulting error in this process is of the order of 0.5-1 mm. The errors that results in reversing of the job is of the order of millimeters and is attributed to backlash. The attachment of “optical follower and position indexing” module to the paper cutting machine eliminates most of the drawbacks mentioned above.
This article describes a novel indexing mechanism to improve the productivity of the paper cutting machines. The mechanism is based on dual light sources and an optical detector. Electronic access The research register for this journal is available at http://www.emeraldinsight.com/researchregister The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0144-5154.htm
3. Optical follower and position indexing
Assembly Automation Volume 23 · Number 2 · 2003 · pp. 172– 173 q MCB UP Limited · ISSN 0144-5154 DOI 10.1108/01445150310471437
The optical follower and position indexing module consists of a calibrated lead screw which is used as an index. A high frequency pulsed IR LED source is used as a reference and mounted on the index lead screw. A dual sensor null detector is mounted on the main lead screw. The absolute position of the main and index lead screws is sensed by using resistance wire (potentiometric) sensors. The schematic of the entire system is shown in Figure 1. 172
A novel indexing mechanism for paper cutting machines
Assembly Automation
Rajanish K. Kamat
Volume 23 · Number 2 · 2003 · 172–173
Figure 1 Schematic of the improved indexing mechanism for paper cutting machines
At the beginning of the paper cutting operation, both the lead screws are brought to a reference position. Then the index screw is rotated by desired angle to index the first position of the cutting. A dual speed arrangement with two separate motors actuate the main lead screw. The main motor moves the lead screw by an appropriate distance to bring the job within the limit of the optical sensors (photo-transistors). The main motor is then stopped and a low torque auxiliary motor is energised for fine movement of the lead screw. The movement continues till the null point is reached.
4. Electronic controller The electronic controller comprises of two main parts: motor controller hardware and job program sequencer. 4.1 Motor controller hardware The motor controller hardware takes the decision regarding whether the main or auxiliary motor is to be energised. This decision is based on the error signal derived through a difference amplifier fed by the absolute position sensors. In order to avoid the excessive ringing of the motors, the difference amplifier is followed by a Schmitt Trigger circuit which provides a dead-band within which the signal variation is ignored. If the amplitude of the error signal is too large (i.e. when the optical sensor signal is not available), the main motor is selected and the auxiliary motor is disabled. When
the optical sensor signal is available, the main motor is stopped and the auxiliary motor is energised. The energisation and de-energisation of the motors is achieved by using a solenoid valve. The small motor runs until the desired position is achieved. Once the desired cutting position is reached a “ready to cut” indication is provided. This optical indicator and follower mechanism is not allowed to operate further until the cutting process is over and the operator presses the “next” button. 4.2 Job program sequencer The job program sequencer has been implemented by using microcontroller 8031. The data of cutting positions of each client is stored on separate EPROM’s. The job number and cutting position is indicated by using an LCD display. Keeping in view the cost factor, CMOS version of 8031 microcontroller has been used in the present application.
5. Conclusion This article describes a novel indexing mechanism to improve the productivity of the paper cutting machines. The mechanism is based on dual light sources and an optical detector. With this new arrangement, indexing accuracy of the order of 30 mm has been achieved. The new mechanism has drastically reduced the setting time of the job. Thus, the productivity of the machine has been found to increase manifold, which well justifies its high cost.
173
1. Introduction
Research article The automated filling of bonded joints – Part 2 three dimensional joints Ken Young and Ian Pearson
The authors Ken Young and Ian Pearson are based at Warwick Manufacturing Group, University of Warwick, UK. Keywords Adhesives, Bonding, Joints Abstract Building automobile bodies from lightweight materials using space-frame construction techniques is increasingly popular because of exhaust emission legislation. One proposed method of achieving this is by using plug and socket joints, which are injected with adhesive after assembly. A method for controlling this process, irrespective of component tolerances, is proposed here. A test rig representing a plug and socket joint was injected with the adhesive and a method for successfully filling the butt-jointed end of the joint found. The addition of a restriction to the joint’s open end gave a method of filling the cavity without creating any air gaps. The use of neoprene O-ring seals for creating the restriction was investigated. The pressure of the adhesive at the joint inlet (gate) was recorded (data logger), and an analysis of this has been used to determine the point when adhesive injection can be arrested and the joint correctly filled.
Electronic access The research register for this journal is available at http://www.emeraldinsight.com/researchregister The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0144-5154.htm
Assembly Automation Volume 23 · Number 2 · 2003 · pp. 174– 180 q MCB UP Limited · ISSN 0144-5154 DOI 10.1108/01445150310471446
The automotive industry is increasingly looking to adhesives to add strength and rigidity to car bodies. It is also using them to allow new methods of construction and new materials to be used (Mortimer, 2001a). While most uses of adhesives are based on the the adhesive being applied to one surface before the parts are assembled, this work looks at a process whereby assembly is completed before the adhesive is injected into the joint. Early work in this area at Warwick concentrated on two-dimensional joints where flat sided aluminium extrusions were inserted into apertures machined in other flat sided extrusions (Young et al., 2002). This is a construction technique which was a development of the technology used for vehicles such as the Lotus Elise and the Aston Martin Vanquish (Kochan, 2001; Mortimer, 2001b) and was used for the Light-weight Concept Vehicle Programme (LCV2 and LCV2/3) by Land Rover (Taylor, 2002). This construction is limiting in terms of vehicle design and tends to give a very “boxy” vehicle, ideal for a Land Rover but not for many other vehicles. Other vehicle producers such as Audi have adopted an approach where large castings are introduced and are used to link more complex shaped extrusion profiles. This leads to the joints being more of a three-dimensional nature, requiring adhesive to fully surround the joint and not appear as two distinct patches on either side of the joint. The use of these castings allows very complex nodes to be produced and massively reduces the parts count on vehicles such as the A2 and A8 (Lewin, 2000). It should be noted, however, that these joints are currently welded and not adhesive bonded. In many ways bonding joints of this configuration offers huge opportunities to simplify the manufacturing process, as in theory only a single point of injection is now required per joint. It also poses problems, however, since the adhesive control as used by Young et al. (2002) is no longer applicable due to the joint gap changing continually around the joint. Being plug and socket joints it is unwise to apply the adhesive prior to assembly as it will be removed on assembly and therefore a different method must be found. A number of issues must be overcome. . Manufacturing tolerances on the components cause a variable volume of
174
.
.
The automated filling of bonded joints
Assembly Automation
Ken Young and Ian Pearson
Volume 23 · Number 2 · 2003 · 174–180
adhesive to be required to fill the joint cavity. A method of finding the correct volume and applying the correct dose of adhesive must be found. The thickness of the adhesive must be controlled between minimum and maximum values in order to ensure a sound joint with adhesive all around it. A method of sealing the joint is required throughout the range of manufacturing tolerances for the components so that adhesive does not spill out and cause contamination. A small riser to take small amounts of excess within the joint is acceptable.
2. Flow of adhesive
adhesives that are appropriate for space-frame construction are available and they have a wide range of curing regimes, but this injection technique should be suitable for them. Adhesive could be injected into the joint through any of the three injection holes at the top of the rig (Figure 1). When a defined volume of the adhesive is injected into any of the injection sites, at constant pressure, it flows in a circular pattern until it comes into contact with the end plate (far left of photograph, Figure 1) or flows around the edge of the plug. The circular arc of the advancing edge can be seen to the right of the photograph below (Figure 2). When Figure 1 (a) Test rig and (b) Cross-sectional view of the test rig
It is envisaged that production joints will be assembled by sliding an extruded socket over a cast plug. This will form a butt joint at one end and will require some method of containment for the adhesive at the other end. Alternatively, because of tolerancing problems it may be advantageous to slide the plug completely inside the socket and produce a joint where both ends require containment for the adhesive. 2.1 Filling the butted end A test rig was manufactured with a transparent outer which allowed the flow path of the adhesive around the joint to be observed. The joint inner (plug) was made of Nylon with its outside dimensions machined to be 7 £ 40 mm: The joint outer (socket) was machined from Perspex and provided a 1.2 mm bond-line thickness to all four sides. The joint was closed at one end with an aluminium plate (shown at the left side of the joint in Figure 1), this plate also held the plug centrally in the socket. The plug was further supported by spigots (grub screws) at the open end, projecting 1.2 mm from the top and bottom surfaces. For the tests, the injection rate was maintained at 4 cm3/s. The adhesive used for this study was 3289 Y 5000, supplied by PPG. It is a structural synthetic rubber adhesive, with a non-volatile content of approximately 98 per cent. This adhesive, like many others manufactured for the automotive industry, will cure when subjected to the paint oven environment (approximately 1808C for approximately 40 min). However, many other
Figure 2 Top view, injecting at Site1
175
The automated filling of bonded joints
Assembly Automation
Ken Young and Ian Pearson
Volume 23 · Number 2 · 2003 · 174–180
a defined volume of the adhesive is injected at Site3 the leading edge of the adhesive advances further to the right and less adhesive reaches the far side of the plug. When the adhesive is injected at Site1 the edge does not advance as far as the previous example because more adhesive flows to the reverse side. The reverse side shows a developing circular pattern, centred on the injection site. When the adhesive is injected at Site3, little adhesive reaches the far side of the plug and a large gap is left between the advancing faces. When the adhesive is injected at Site1, closer to the end plate, more adhesive enters the far side and the gap is closed to a greater extent than before (Figure 3). These experiments show that the closed end of the joint can be satisfactorily filled simply by moving the Injection Site (gate) closer to that end plate. The advancing edge of the adhesive at the open end will continue to flow in a circular form (centred on the Injection Site) until the injection gun was emptied or a restriction is encountered.
2.2 Filling the constrained end The rig was modified to restrict the flow of the adhesive along the joint away from the end plate. This was accomplished using a band of masking tape that fills the space between plug and socket and incorporates a gap (the riser) and is shown from the underside in Plate 1. Adhesive injected at Site1 (nearest to the end plate, Figure 1) flows around the plug, completely filling the cavity and then flows along the joint and through the riser. 2.2.1 Restricted volume with off-set riser For the previous experiment the riser had been conveniently placed on the centre line of the plug, which, because of the symmetry of the rig would be the last place for the adhesive to flow. As it cannot be guaranteed that this will always be possible in practice a series of experiments was carried out to determine if successful filling (no air gaps) was dependant on the riser being placed at the point where the adhesive would flow last. The sealing ring (masking tape) was modified to contain an off-centre riser. The adhesive was again injected at Site1 (near the end plate) and the result is shown in Plate 2 where the cavity has again been filled successfully.
3. Air gaps (disbonds) For the formation of an air gap to be averted, the air would have to escape through a route other than the riser. Work carried out by Hart-Smith (1981) shows that the incorporation of air gaps or disbonds in an adhesively bonded joint need not have a great effect on the mechanical performance of the joint. When the adhesive is injected into the above rig, even with an offset riser the cavity is successfully filled and no air gaps are evident. This is because the band of masking tape does not make an effective seal onto the Perspex Figure 3 Reverse view, injecting at Site1
Plate 1 Reverse view, restricted, injecting at Site1
Plate 2 Reverse view, restricted, off-set, injecting at Site 1
176
The automated filling of bonded joints
Assembly Automation
Ken Young and Ian Pearson
Volume 23 · Number 2 · 2003 · 174–180
outer. However, a more practical seal material would be required for production and it would be necessary for this seal to allow air to escape whilst retaining the adhesive.
Consequently, this phenomenon was investigated as a possible means of controlling the adhesive injection gun and so programmes were written in MATLAB Software to identify this point. The data (pressure and time) had been digitally collected at 1,000 Hz. Consequently, if the signal (pressure in this case) has not changed sufficiently before the next sample is taken (after 0.001 s for 1,000 Hz sampling rate) it will be recorded as the same value as the previous recording. Therefore, if the time between recordings is short and the signal increases only a small amount, the first differential will appear erratic, since the signal may have increased and produced a small gradient in reality, but the digitally recorded signal will show no increase and hence a zero gradient. There are also some spurious peaks in the recorded signal as a result of electrical interference etc. and these will produce a large difference and hence appear as a large gradient. For the purposes of investigating the effectiveness of the change in gradient as a means of determining the “shut off” point for the injection gun, a Matlab macro file was written to: (1) remove unnecessary data (the data removed from the pressure history shown in Figure 4, anything before approximately 0.4 s and after 6.2 s), (2) reduce the data to 20 Hz, (3) calculate the gradient for every five data points along the pressure history, and (4) ask for an estimate of the minimum filling time (largest plug in the smallest socket).
4. Pressure history of injection event An SCA AK 3100 adhesive injection gun was used for this feasibility study. It had been modified and a nozzle added to allow a pressure sensor to be incorporated and determine the injection pressure as close as possible to the injection site (Young et al., 2002). The pressure measured from this modified gun whilst filling the rig (with restricted volume and off-set riser) is shown in Figure 4; the joint was completely filled at both ends. A large increase in the rate of applied pressure can be seen at the right hand side of the pressure history. This change of rate was further investigated as a means of determining when the cavity had been filled and hence when the adhesive injection gun should be stopped.
5. Analysis The point at which the gun is switched off will inevitably be some multiple of a value determined early in the injection cycle. For the process to be robust, the point at which the gun is switched off should be a large multiple. The change in gradient was believed to be an indication of the joint being almost full. Figure 4 Pressure history
177
The automated filling of bonded joints
Assembly Automation
Ken Young and Ian Pearson
Volume 23 · Number 2 · 2003 · 174–180
The gradient for every five data points from the start (the unnecessary data has been removed) is calculated. A running total is maintained and when the time reaches half of the “estimate of the minimum filling time”, it is averaged. This is deemed to be the “initial gradient”. The programme then continues to process further data until a point is reached where the newly calculated gradient is more than eight times greater than the “initial gradient” and this is where the injection gun should be switched off.
are filled (using the left-hand and right-hand O-rings). For this feasibility study, a simplified approach to determining the point at which the gun should be switched off was made. A programme was written for the PLC which determined the minimum applied inlet pressure during a time period shortly after the initial peak and a simple multiple of this was used as the “switch-off ” point. When the rig was filled with the adhesive with injection time controlled by the PLC, both cavities were correctly filled, producing the pressure histories shown in Figure 6. For the large cavity, the unsupported length of the plug is greater and consequently, it is subjected to greater bending and the pressure history shows a pressure drop after approximately 0.75 s. An aluminium plug would be stiffer and so the displacement for the same pressure would be smaller, but this deflection can be controlled by a spigot or boss being included in the design of the plug, opposite the injection site.
The programme, working on post-test data, was made to draw a vertical line through the point where the gradient was found to exceed eight times the “initial gradient” so that the results from several tests could be compared. These results are shown in Figure 5.
6. Test rig incorporating O-rings A further rig, incorporating neoprene O-ring seals was designed and manufactured and as before comprised a Perspex outer and a Nylon inner; Plate 3 shows the upper (injection) surface of the rig. For this feasibility study, the O-rings were manufactured by cutting the required length from a continuous length of O-ring material and bonding the ends to form a circle. For the production environment, the rings would need to be moulded to produce an oblong seal (with rounded corners) that would conform more closely to the geometry of the cast plug and be easier to assemble, but neither the manufacturing nor the assembly of the O-rings has been addressed in this study. This rig has no closure plate to hold the plug centrally in the socket since this and the tape band have been replaced by O-rings. However, it was found that the O-ring material was not sufficiently stiff to withstand the force generated by the injection pressure without significant displacement and this resulted in an O-ring at the upper surface being displaced from its groove. To restrain displacement of the plug away from the injection site, spigots (grub screws) were incorporated to the bottom face of the plug, similar to the previous rig. This rig was designed to incorporate two cavities and either one or both could be filled by the adhesive. Plate 3 shows that with O-rings in all three grooves but in practice, either the left hand cavity is filled (using the left-hand and middle O-rings) or both cavities
Figure 5 Determining switch-off point from gradient
Plate 3 O-ring seal test rig
178
The automated filling of bonded joints
Assembly Automation
Ken Young and Ian Pearson
Volume 23 · Number 2 · 2003 · 174–180
Figure 6 Controlled filling of large and small cavities
7. Joint design When designing an adhesively bonded plug and socket joint that will have the adhesive injected into it, the following points should be considered. (1) Restrain the deflection of the plug by including spigots or bosses in the design of the cast plug. Bosses or spigots outside the bonded area will allow “beam bending” of the plug, caused by large pressures acting over large areas. A boss opposite the adhesive injection hole will greatly reduce this. (2) Restrain deflection of the socket. This can only be achieved by increasing the stiffness and therefore either the material properties or the second moment of area of the extrusion walls has to be modified. The material has been chosen because of its extrusion capabilities and so the only possibility is the wall geometry. (3) O-rings should be used to seal the joint. The manufacturers recommend a minimum of 20 per cent compression of the diameter and the design should therefore achieve 20 per cent when assembled into the largest bond-line (smallest possible plug with largest possible socket). The material outside the bonded area should be as large as possible (the maximum possible material allowing some clearance to the smallest possible extruded socket). (4) Internal corner radii of the extrusion should be large as they are required to seal
against an O-ring. They are also related to the corner radii of the grooves of the plug. (5) For production purposes, it is believed that the O-ring diameter should be large (about 6 mm, in the feasibility study only 3 mm was used), as this will aid assembly. However, the minimum bend radius of the O-ring is dependant on the O-ring radius. These seals would be better moulded as a circular section oblong seal.
8. Issues remaining Numerous algorithms for switching off the injection gun can be generated. Those investigated in the feasibility study should now be developed into robust programmes. The switch-off signal is determined by the ratio of “initial pressure” to current pressure. By modifying the shape of the plug in the area under the injection site (gate) geometry may be developed that produces a lower “initial pressure” and therefore a larger ratio will be generated and a more robust method will be created. The adhesive’s viscosity is affected by temperature and this should be investigated as a means of lowering the forces induced in the joint and therefore the displacement. Lowering the viscosity will also lower the amount of energy stored in the injected adhesive and therefore the volume of the adhesive forced out of the injection site after the injection process will be decreased.
179
The automated filling of bonded joints
Assembly Automation
Ken Young and Ian Pearson
Volume 23 · Number 2 · 2003 · 174–180
The adhesive may be subject to detrimental aging effects and this may have an unfavourable influence on the injection parameters. This requires further study. For the purpose of creating a robust process for production, the following physical properties of the adhesive need to be investigated: (1) shear thinning, (2) temperature/viscosity curves, and (3) compressibility.
geometries used in the test cases and therefore further work is required to prove the technique with “real” samples.
Assembly of the O-ring to the plug has not been addressed in this feasibility study but methods of assembly need to be developed. This must be done using aluminium components.
9. Conclusion This study shows that the adhesive injection is a promising solution for automotive production. Many of the parameters employed are specific to the materials and
References Hart-Smith, L.J. (1981), “Effects of flaws and porosity on the strength of adhesively-bonded joints”, McDonnel Douglas, MDC J4699. Kochan, A. (2001), “Weighing up the options”, Automotive Manufacturing Solutions, June 2001 pp. 32-49. Lewin, T. (2000), “A space odyssey”, Automotive World, May 2000, pp. 56-60. Mortimer, J. (2001a), “Coherent solutions for vehicle construction – the ultimate three-day cars”, Automotive Manufacturing Solutions, June 2001, pp. 152-9. Mortimer, J. (2001b), “Space frame vehicles – the ultimate three-day cars”, Automotive Manufacturing Solutions, June 2001, pp. 145-50. Taylor, J. (2002), Lightweight Chic Land Rover Enthusiast, December 2002. Young, K., Pearson, I. and Bull, R. (2002), “The automated filling of bonded joints – Part 1: two-dimensional joints”, The International Journal of Assembly Technology and Management, Vol. 22 No. 4.
180
1. Introduction
Research article Optimum assembly design utilizing a behavioral modeling concept Y.J. Lin and R. Farahati
The authors Y.J. Lin and R. Farahati are at the Department of Mechanical Engineering, The University of Akron, OH, USA. Keywords Assembly, Modelling, Computer aided design Abstract This paper presents a versatile and economical knowledge-based assembly design of blade and shell assemblies by employing behavioral modeling concepts. Behavioral modeling is a new generation CAD concept aimed at achieving ultimately optimum results with the efforts made in the early stage of the product development cycle. As a result, the assembly process of any oddconfigured parts such as torque converter blades, can be accurately planned, and made adaptable to all potential in-process alterations due to either changes of components design or that of the assembly kinematics. Optimum assembly design is achieved when the volumetric interference meets a desired value based on an expert’s determination. Experimental verification of the proposed optimum assembly design conducted in Luk, Inc. with two different blades’ assemblies demonstrates satisfactory results. Electronic access The research register for this journal is available at http://www.emeraldinsight.com/researchregister The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0144-5154.htm
Assembly Automation Volume 23 · Number 2 · 2003 · pp. 181– 191 q MCB UP Limited · ISSN 0144-5154 DOI 10.1108/01445150310471455
A behavioral modeling design system develops the CAD system one step closer towards an intelligent modeling design system. That means it is using the design intent and design constraint for generating all possible geometric shapes for a design. An intelligent CAD system is a target for automating a manufacturing design system. It contains all the specification and process information needed to adapt to the environment. In addition, it applies to all kinds of analyzing and optimizing methods to automatically model and generate all possible designs that fulfill the design intents. In other words, in designing a torque converter there will be no need to design, analyze, and optimize the components individually without considering the inter-relational effect of the components in a whole mechanism. However, in order to create such an intelligent CAD system, we must quantify all inter-relational effects of the components of the torque converter during manufacturing and operation. In order to embed and trade specific functions efficiently in a CAD system, it is necessary to understand and quantify the semantics of the technological objects. This is concerned with determining the mapping relation between technology and geometry and to clarify the logical process of specification of these technological objects (Dominique, 1999). Therefore, this paper is focused on benchmarking the interrelationship effect of the torque converter components during the assembly process. The new generation of CAD system, equipped with a behavioral modeling engine, addresses these needs and promotes the creation of well-designed products through the synthesis of requirements, desired functional behavior, design context, and geometry in an open, extensible environment (PTC, 2000). The torque converter is a very complex turbo machine. Its geometry is highly The project was supported financially in part by Luk, Inc. and University of Akron. The unrestricted research grant funded for this project by Luk, Inc. was gratefully acknowledged. The company’s proprietary blades and turbine shell prototypes provided by Luk, Inc. for the experimental verification of the proposed design and suggested assembly paths were also very much appreciated.
181
Optimum assembly design utilizing a behavioral modeling concept
Assembly Automation
Y.J. Lin and R. Farahati
Volume 23 · Number 2 · 2003 · 181–191
three-dimensional. Design technology in torque converters has advanced very little during the past few decades. The design procedure has been based greatly on the use of a trial and error approach (Robert and Mahoney, 1988). In order to quantify the entire process of the assembly of the torque converter we need a suitable method. Assembly has traditionally been one of the most important stages for product development, which generally accounts for 50 percent or more of manufacturing costs, and also affects the product quality (Lotter, 1993). Assembly is more than just putting parts together. Assembly is the activity in which all the upstream processes of design, engineering, manufacturing, and logistics are brought together to create an object that performs a function. Assembly, which actually creates the product, is by comparison much less studied and by far one of the least understood processes in manufacturing (Whitney et al., 1999). In the torque converter industry, the blade and shell components possess irregular geometric shapes with freeform surfaces resulting from complicated computational fluid dynamics (CFD) design constraints (PTC, 2000). Therefore, they cannot be assembled based on conventional methods of design for assembly of which the mating, aligning or offsetting of the regular counter faces are being used for assembly. Also, due to the constrained characteristics of the blade in torque converters, the shape of the blade cannot be altered for the sake of simplifying the assembly process because the blade geometry is highly sensitive and critical to the torque converter performance. This implies that the original design configuration of the blade component must be obeyed in the assembly to achieve the desired power transmission. Therefore, it poses many challenges on the practical assembly work of blade insertions into the pump shells’ slots, which is the main reason for this type of assembly to have been accomplished manually up to the present time. As blades are usually in odd geometric shapes from the original specification, the conventional design guidelines for part assembly cannot be applied to the blade assembly design directly. This is because the orientations of the tabs on a blade vary between them. Hence, the trajectory path of a blade during the kinematical assembly
process cannot be obtained directly based on the tab orientations. In this situation, the tabs on the blade cannot be assembled in a direction normal to the slot, and each tab has a different three-dimensional path during the assembly, which complicates the process design. Due to the requirement of the different orientations in space for blade tabs, a blade must have six degrees of freedom in order to be properly inserted into the shell slots during the assembly process. Therefore, up to the present time in the torque converter manufacturing community, it is impossible to predict individually the trajectory of each tab on a blade relative to the slot before making a real prototype. Currently, in the torque converter industry, the tab and slot are designed by a trial and error method. Its feasibility is checked by making real prototypes. Apparently, this method is too costly, time consuming and not accurate. The recent work of the authors proposed a method to create a CAD-based virtual blade assembly prototype for the elimination of the real prototyping requirement (Farahati, 2001; Lin and Farahati, 2003). In this paper, the previously mentioned work (Farahati, 2001; Lin and Farahati, 2003) is naturally extended by integrating a behavioral modeling program to the virtual prototyping results. By doing so, the optimum trajectory of the assembly paths can be obtained effortlessly using the developed methodology for meeting the desired behaviors. In an experimental research, the proposed optimum assembly design model is applied to two existing virtual turbine designs in LUK Incorporated in Wooster, Ohio. The first turbine design has a three-tab blade design and the second one has a four-tab blade design. The three-tab blade turbine was in the first stage of design and there was no real prototype built at the time. But the four-tab blade design had already a third real prototype ready. The virtual blade assembly prototype was applied to the three-tab blade design to visualize the blade motion during the assembly process and to predict the feasibility of the blade assembly. Also by applying the behavioral modeling to the blade assembly model, the optimum trajectory path was predicted. In this paper, the results of the experiment is illustrated in detail. Second, an experiment was applied to the four-tab blade virtual model in order to compare the result
182
Optimum assembly design utilizing a behavioral modeling concept
Assembly Automation
Y.J. Lin and R. Farahati
Volume 23 · Number 2 · 2003 · 181–191
Plate 1 Sequence 1 position of the three-tab blade assembly in a turbine shell
with the real prototype and to verify the model’s effectiveness.
2. A three-tab blade form model development As mentioned earlier, our approach proposed in this work requires the establishment of two essential aspects of assembly design, namely, the form (CAD component models) and behavior (assembly simulation models) of blade and shell assembly. The objective of this experiment is to predict the feasibility of the blade assembly process in a three-tab blade form development by applying the virtual blade assembly model. The virtual form model is then used as a feedback mechanism for obtaining potential blade trajectory paths and the maximum blade volumetric interferences with the turbine shell during the assembly. 2.1 Form object development using visualization of the blade assembly process After defining all constraints, drivers, and trajectories of the kinematic motion in the blade virtual assembly model, the process of the blade assembly can be visualized in the developed virtual assembly prototype for a three-tabs blade. In this virtual assembly animation, the entire mechanism of assembly and the concept of the blade assembly process can be actually seen on the computer screen. The virtual assembly animation can be defined for various paths and generate different trajectories in a series of different experiments. The form model of the virtual prototypes can emulate the kinematic motion of the blade performed by an automatic assembly machine. This form model enables us to check the trajectory against the desired behavioral object – the optimum volumetric interference – before applying the model to an automatic assembly machine. Plates 1 and 2 show the three positions in the sequence of a three-tab blade assembly in a turbine shell. 2.2 Behavior object development with an analysis of the blade interference The assembly process can be analyzed with different kinds of measurements such as tab length, which penetrates the shell body through the slot. In other words, it is the length of the tab offset from the slot. Another measurement is the volumetric interference
between the two parts engaged in any instant during the assembly process. The volumetric interference of the assembly can be compared against the assembly time. The entire volumetric interference of tab number 3 in a three-tab blade assembly process is shown in Plate 3. In the figure, the interfered volume is represented by the meshed web. Similarly, it is required to develop a behavioral object regarding the interference of tab 2 with the turbine shell. Plate 4 shows the virtual form model of the entire volumetric interference for tab number 2 during the assembly process. The volumetric interference measured in cubic millimeters versus assembly time expressed in seconds for the three-tab blade is then readily plotted in Figure 1. This simulation diagram is utilized as one of our behavioral objects for feedback checking against the functional requirements. According to this diagram, there is 10.4 mm3 volumetric interference for
183
Optimum assembly design utilizing a behavioral modeling concept
Assembly Automation
Y.J. Lin and R. Farahati
Volume 23 · Number 2 · 2003 · 181–191
Plate 2 Sequence 3 position of the three-tab blade assembly in a turbine shell
Plate 3 Volumetric interference of tab #3 with the shell in the three-tab blade assembly
Plate 4 Volumetric interference of tab #2 with the shell in a three-tab blade assembly
the three-tab blade assembly, which is clearly exceeding the required/minimum value according to the expert’s determination for a normal blade assembly. Also the diagram indicates that the maximum volumetric interference occurs at 0.75 s of assembly duration. Based on this simulation, it reveals that the blade assembly experiment in this set of form and behavior modeling is not feasible with the existing tab-slot design in the threetab blade turbine. It is also noted that the blade disassembly process is established instead of its assembly. In fact, this is an identical process in a reverse sequence for the assembly process. Apparently, in the diagram shown, it can be seen that a disassembly time started from zero seconds when the blade was inside the shell, and ended at 0.95 seconds when the blade was completely out of the turbine shell with no interference.
3. A four-tab blade form model development Again, to obtain the optimum assembly design of the four-tab blade assembly process, we need to establish the corresponding form and behavior objects for the evaluation. Therefore, verification of the virtual model is the objective of this experiment.
184
Optimum assembly design utilizing a behavioral modeling concept
Assembly Automation
Y.J. Lin and R. Farahati
Volume 23 · Number 2 · 2003 · 181–191
Figure 1 Diagram of interfered volume versus assembly process time for a three-tab blade
The resulting blade assembly design is obtained by applying the behavioral modeling approach to the four-tab blade electronic model, and the result is then compared with the physical prototype for manual assembly. 3.1 Obtaining the optimum blade trajectory in assembly kinematics The behavioral modeling program is applied to the virtual blade assembly prototype in order to identify the optimum trajectory in the assembly process automatically. The behavioral modeling system can be employed on an assembly design only if the design intent and design parameters are defined for accomplishing the design. In other words, the behavioral modeling program captures the design intent and design parameters as input data to drive the process for achieving an optimal outcome. Then, it changes the critical parameters step-by-step and in each step the designer runs the virtual test for potential design outcomes. All possible changes in the parameters are tested and the results recorded. At the end, the program compares
the entire results with the design intent as thresholds and suggests the best results corresponding to the given parameters in order to satisfy the design intent based on certain anticipated functionalities. Sometimes, the final result can be more than one. Since the blade assembly process is quantified already in the previous paper (Lin and Farahati, 2003), the design intent and design parameters can be captured from the model easily and they can be applied to the behavioral modeling program for obtaining the optimum trajectory for the underlying blade assembly process. The following subsections serves as a guideline as to how to define the design parameters and design intent for a virtual blade assembly prototype with optimum volumetric assembly interference as the ultimate behavior objective. 3.2 Identification of the design parameters for a blade assembly The general design parameters in the virtual blade assembly process can be considered
185
Optimum assembly design utilizing a behavioral modeling concept
Assembly Automation
Y.J. Lin and R. Farahati
Volume 23 · Number 2 · 2003 · 181–191
including, but are not limited to the following variables, namely: (1) blade pitching angle – this is the 1st of the three degrees of motion for rotation of the blade, (2) blade yawing angle – this is the 2nd of the three degrees of motion for rotation of the blade, (3) blade rolling angle – this is the 3rd of the three degrees of motion for rotation of the blade, (4) three drivers at the pivot point for keeping the tip of the blade constant during the blade rotation, (5) three drivers on the blade, which drive the blade out of the shell in three different directions for disassembly kinematics, (6) tab and slot locations, orientations, and their geometries.
The experiment was run for various yawing angles and the volumetric interference was recorded in the volumetric interference versus time diagrams. In the second step, the rolling angle is changed to a new value to represent a continuous kinematic motion of the expert’s assembly practices and remains constant during the test. Again, the same experiment is repeated. The volumetric interference is recorded and can be compared as to which angle leads to the minimum volumetric interference designated as the desirable behavior. The graph reports the volumetric interferences between the blade and shell versus assembly time in motion. These diagrams are used as feedback mechanisms to assist the decision-making in reaching an ultimate assembly design. The goal is to obtain optimum assembly motion information for establishing an automatic assembly plan with optimum blade motion loci during the assembly process.
On the other hand, we consider the design intent for the blade assembly to possibly include the following quality variable, namely, “The volumetric interference of the blade with the shell should be equal to a pre-specified value.” In fact, this quality variable is the principal design objective in this investigation. The value is determined from an expert’s knowledge base, usually based on an experienced assembler’s good experimental and hands-on knowledge. The pre-specified value for volumetric interference should be a positive quantity to ensure that there is sufficient but not excessive engagement of the blade and turbine shell which provides the necessary friction between mated components after the assembly is completed. After capturing these selected parameters as the inputs, the behavioral modeling engine is fired up and runs the test for generating all-possible combinations of the blade motion angles (i.e. the pitching, yawing, and rolling angles). As mentioned earlier, in a general blade rotation, there are three degrees of freedom of movement for the spatial rotational movement. The first tab slot is chosen to be the motion guiding reference since the tab of a blade has to go inside the slot first. As a reference, it is found that when the first tab surface on the blade is parallel to the first slot surface on the turbine shell, the yawing angle should be zero. In the first step of the blade-shell assembly, the yawing angle and rolling angle should both stay constant. At this instant of time, the pitching angle is varied to drive the blade assembly process.
4. The development of behavioral objects As described in the previous section, the ultimate goal of this study is to obtain an optimum blade and turbine shell assembly process in terms of a motion locus. In view of the previous derivation, it is seen that the optimum blade trajectory can be obtained for the given tab and slot design from the coordination and counter-balance of the combined information of both form and behavior objects. In other words, if the shortest blade path does not satisfy the desired volumetric interference (i.e. design intent in a pre-specified value), the tabs will be adjusted and the test will be run again. However, if the desired volumetric interference could not be achieved simply by tab adjustments, the turbine shell slots will be redesigned and the test will be run again. This iterative sequence is described in the following subsections for the four-tab blade experiment, to simply demonstrate one practical example of our proposed approach. 4.1 First simulation to establish a reference path This experiment starts with a reference path when all angles (pitch, yaw, and roll) are set to
186
Optimum assembly design utilizing a behavioral modeling concept
Assembly Automation
Y.J. Lin and R. Farahati
Volume 23 · Number 2 · 2003 · 181–191
zero. Figure 2 shows the result of this simulation. For this experiment, a 12 mm3 maximum volumetric interference can be found from the diagram, which is almost in the middle of the assembly process with respect to the assembly time, at around 2.4 s under our preset testing period and motion speed.
at the final time (simulation ending point), the blade is disassembled or completely disengaged from the turbine shell. This is why a very small volumetric interference is seen at the starting time, but no interference appeared at the end of the assembly time.
4.2 Second simulation with yawing angle 5 0.18 In this test, we increase the yawing angle of the blade tab to a non-zero value, 0.18. Figure 3 shows the result of this simulation. Based on the readings in the diagram, a maximum volumetric interference of 6 mm3 can be seen, which is nearly one-half of that of the first experiment. Also revealed in the results is a shift of the maximum peak to the left of the diagram. This means that the maximum volumetric interference occurs in the later stage of the assembly relative to the first experiment. We must point out again that the diagrams we obtain always show the disassembly processes. This means that at time zero (simulation start point), the blade is fully assembled into the turbine shell. However,
4.3 Third simulation with yawing angle 5 0.28 This experiment increases the yawing angle of the blade tab to 0.28. Figure 4 shows the result. In the chart, the maximum volumetric interference is approximately 13 mm3, which occurred around the end of the assembly motion. This implies that in order to twist the blade to an angle of 0.28 when it is inside the shell, 13 mm3 volumetric interference, or a corresponding equivalent torque for deflecting the blade, is needed. 4.4 Other simulations for obtaining behavioral objects Several more simulations were run with different parametric yawing angles for the blade assembly processes. The results are summarized and shown in Figure 5.
Figure 3 Interference diagram for yawing angle ¼ 0.18, in a four-tab blade with the maximum interference of 6.2 mm3 Figure 2 Interference diagram for yawing angle ¼ 0.08, in a four-tab blade with the maximum interference of 12 mm3
187
Optimum assembly design utilizing a behavioral modeling concept
Assembly Automation
Y.J. Lin and R. Farahati
Volume 23 · Number 2 · 2003 · 181–191
Figure 4 Interference diagram for yawing angle ¼ 0.28, in a four-tab blade with the maximum interference of 13 mm3
Figure 5 Volumetric interference diagrams for various yawing angles, in a four-tab blade assembly
Figure 5 clearly shows that the results of the volumetric interference were apparently too high for yawing angles greater than 0.28 and less than 0.08. Therefore, blade tabs having these angles for assembly are not acceptable. It also indicates that the scope of the path is narrowed to the range of angles between 0.08 and 0.28. In Figure 5, six results between 0.08 and 0.28 can be seen. The diagram depicts this by increasing the yawing angle where the maximum peak is shifted to the left. For an angle greater than 0.18, there are two maximum peaks, one near the left and another one close to the right. According to these results, the minimum volumetric interference occurred with the twisting angle equal to 0.158, which yields an interference volume of about 4.4 mm3. By comparing the first three simulations in Figure 5, a simulation with a combination of these three can be run. This means that at each moment, the trajectory follows the optimum path in the diagram so that it starts with the twisting angle of 0.18 and rotates for 1/3 of the total time in disassembly. Next, the angle is changed to 0.158 and rotates until one half of the disassembly time.
Finally, the yawing angle is changed to 0.28 for the rest of the disassembly. The combination result for the three paths is shown in Figure 6. The maximum interference is 3.8 mm3. According to the expert’s knowledge verified by manufacturing department’s statistics, this experimental result is quite satisfactory.
5. Optimum path with minimum volumetric interference In the developed behavioral modeling program, all simulation results are put together to serve as referenced behavioral objects. By adjusting the form objects (blade tab angles and kinematics) based on the behavioral objects (constraints and references), similar to a closed-loop feedback control process, the optimum assembly path is obtained from the combination of the simulation results. Figure 6 demonstrates the optimum disassembly path and the minimum volumetric interference for a four-tab blade during the assembly process. The minimum volumetric interference is found to be about 3.6 mm3. This interference implies that some snapping force is needed to assemble a blade to the shell turbine as the two
188
Optimum assembly design utilizing a behavioral modeling concept
Assembly Automation
Y.J. Lin and R. Farahati
Volume 23 · Number 2 · 2003 · 181–191
Figure 6 Volumetric interference graph for combined twist angles, in a four-tab blade with the maximum interference of 3.8 mm3
components are engaging each other. However, the volumetric interference done in this experiment is not as high as that of the three-tab blade assembly, which was found to be about 11 mm3. The analysis described in the previous sections enables us to accomplish our goal of successfully developing an optimum assembly design of blade and shell assemblages by implementing behavioral modeling concepts in the design process. Towards the end, in Figure 7, the optimum blade trajectory which also yields minimum volumetric interference at the assembly engagement can be obtained. The arrows in Figure 7 shows the optimum blade trajectory for the four-tab blade assembly. Detailed interpretations of the graphs are given as follows. The duration of the blade assembly is 28 units or 2.8 s and the rolling angle is set to zero. However, the pitching angle is varying from 0.008 to 40.008 in all assembly paths. The following facts are shown in the graphs of Figure 7: . in the first 8/28 portion of the total time, the blade yawing angle was 0.18.
.
.
.
next, the angle changed to 0.158 until it reached 10/28 of the total time, then, the yawing angle changed to 0.168 until it reached 13/28 of the total time, and finally, the yawing angle increased to 0.28 until the end of the assembly period.
The simulation results were checked and verified with the physical four-tab blades of a torque converter prototype in Luk, Inc. In the lab testing of realistic blade and turbine shell assembly, we verified that the assembly trajectory of the four-tab blade was exactly the same as the predicted trajectory of the blade in the assembly simulation. In addition, it is found that the location of engaging interference and volume in the simulation agreed very well with those measured in the physical prototype testing as well. Therefore, the new approach proposed in this research, namely, developing optimum assembly design by employing behavioral modeling concept has proved to be effective, reliable and accurate for the mechanical assembly of blade and turbine shell of a torque converter.
189
Optimum assembly design utilizing a behavioral modeling concept
Assembly Automation
Y.J. Lin and R. Farahati
Volume 23 · Number 2 · 2003 · 181–191
Figure 7 Volumetric interference graph for optimum path for the blade assembly in a four-tab blade with the maximum interference of 3.6 mm3
6. Conclusion In this paper, behavioral modeling concepts were introduced in an implementation of optimum assembly design for freeform blade and turbine shell assemblies in an automobile torque converter. Behavioral modeling is superior to conventional CAD for assembly design in that it incorporates design intents of the engineers and required functionalities of the underlying product/assembly into the geometric form model which turns the design into a much more manageable task for the designers, similar to the design of a feedback control system. Specifically, two aspects of assembly design, namely, the form (CAD component models) and behavior (assembly simulation models) of blade and shell assembly were established in the first place. Then, by manipulating these aspects, designers were able to fine tune the identified parameters controlling the optimum design outcomes through a virtual prototype simulation of the assembled system being designed. This virtual prototype simulation provided instant feedback regarding design decisions/best choices by continuously checking whether all pre-specified
functionalities were met. In this case, the pitch, yaw, and roll angles of the freeform configuration of a blade component object in an assembly kinematics motion into a turbine shell’s slots were selected. These parameters were identified to establish parametric relations between volumetric interferences in the assembly between the engaged blade and the turbine shell. Sufficient number of experimental simulations were conducted using a behavioral modeling engine in a CAD system. These simulation results provided enough feedback data for the fine tuning of design parameters. Optimum assembly design was achieved when the volumetric interference met a desired value obtained by the experts’ determination. Experimental verification of the proposed optimum assembly design conducted in Luk, Inc. with two different blade assemblies demonstrated satisfactory results. All in all, the new approach proposed in this research, namely, developing optimum assembly design by employing behavioral modeling concepts has been proved to be effective, reliable and accurate for the mechanical assembly of the blades and turbine shells in a torque converter. Therefore, the work presented in
190
Optimum assembly design utilizing a behavioral modeling concept
Assembly Automation
Y.J. Lin and R. Farahati
Volume 23 · Number 2 · 2003 · 181–191
this paper can contribute to serving as an assembly design guideline for any mechanical assembly.
Lin, Y.J. and Farahati, R. (2003), “CAD-based virtual assembly prototyping – a case study”, I. J. Adv. Manufacturing Tech. (in press). Lotter, B. (1993), Manufacturing Assembly Handbook, Butterworths, London. Parametric Technology Corporation (PTC) (2000), “Consistent innovation with behavioral modeling”, Parametric Technology Corporation White Paper. Robert, R. and Mahoney, J.E. (1988), “Technology needs for the automotive torque converter – Part 1: internal flow, blade design, and performance”, SAE Paper No. 12482. Whitney, D.E., Mantripragada, R., Adams, J.D. and Rhee, S.J. (1999), “Designing assemblies”, Research in Engineering Design, Vol. 14, pp. 441-7.
References Dominique, D. (1999), “Introduction to assembly features: an illustrated synthesis methodology”, Journal of Intelligent Manufacturing, Vol. 21, pp. 178-84. Farahati, R. (2001), “CAD-based virtual prototyping of blade assembly process in vehicle torque converters utilizing behavioral modeling technique”, PhD dissertation, University of Akron, Akron, Ohio, Fall.
191
Introduction
Research article A simulation method and distributed server balancing results of networked industrial robots for automotive welding and assembly lines Paul G. Ranky
The author Paul G. Ranky is a Professor at the Department of Industrial and Manufacturing Engineering, New Jersey Institute of Technology (NJIT/MERC), Multi-lifecycle Engineering Research Center, Newark, NJ, USA. Keywords Automotive, Software Abstract The fundamental purpose of building simulation models is because it is often impossible to experiment with real-world systems. In terms of our networked robots in the automotive welding and assembly lines, simulating possible scenarios, both for design as well as for operation control purposes, is very important for the design team, as well as for management, due to the per-minute-cost of every failed robotic operation. In order to support, both the design as well as the management community of such systems, in this article, we discuss a generic simulation methodology, using the IT-Guru OPNET simulation program, as well as show practical results, that demonstrate the benefits of simulation methods. Electronic access The research register for this journal is available at http://www.emeraldinsight.com/researchregister The current issue and full text archive of this journal is available at http://www.emeraldinsight.com/0144-5154.htm Assembly Automation Volume 23 · Number 2 · 2003 · pp. 192– 197 q MCB UP Limited · ISSN 0144-5154 DOI 10.1108/01445150310471464
Simulation, as a method, like most system analysis and design methods, involves systems and models of them. A system for our purposes can be defined as a robot welding or assembly facility in an automotive environment, or robotized welding, or assembly process, either actual, or planned, with all important automation and human resources, computer communication systems, sensors, tools and materials, and fail-safe systems (Banks and Carson, 1984; Carr, 2002; Dornan, 2002; Ranky, 2002c). A real-world network of robots is composed of physical sites that can send, route, or receive information, between a network of distributed robot controller PCs, server workstations and various routers and switches. In terms of networking, these sites are referred to as nodes. In our study, we employ a simulation software package, called IT-Guru, by OPNET Technologies Inc., USA. This modern, object-oriented simulation system uses communication node objects to represent the physical sites of our networked robots. Real-world nodes are connected by and communicate across links, represented in IT-Guru by communication link objects. Links carry information over distances and can represent items such as electrical, orfiber optic cables, or even wireless communication links between networked devices. As communication systems of distributed networked robots become more complex, it is often necessary to view groups of nodes within a network as a single entity. Such a grouping of nodes, along with the links connecting them, is called a sub-network (IT-Guru, OPNET Technologies, 2002; Ranky, 2002a, b). As a practical example for such an architecture imagine that the modern robot controller PC not only controls the arm’s movements, but also the smart sensor equipped welding guns, or active sensory controlled assembly heads, a variety of vision and other non-contact and contact sensors, safety devices, communication between the controller and the rest of the network, and We hereby acknowledge the generous support offered by OPNET Technologies Incorporated, Bethesda, MD, USA, for offering the author full access and support, and a class and research license for his NJIT undergraduate and graduate students of the simulation system IT-Guru.
192
Networked industrial robots
Assembly Automation
Paul G. Ranky
Volume 23 · Number 2 · 2003 · 192–197
many others. Therefore, a robot in such a modern, networked system has both an inside-distributed network, run by the realtime side of the robot controller, and the higher level robot controller PC, being integrated into a distributed network, such as the robot line. The term simulation applies to a wide range of methods and applications designed to mimic the behavior of real-world, or imagined, virtual world systems. Simulation for our networked robot engineering purposes is usually performed on a computer with powerful software; nevertheless “mental simulation”, typically preceding any computerized work is a valuable and very powerful engineering problem solving method too, because it enables the system analysts and engineers to think of real-world, or imagined experiences and reason over which parameters and values are truly important, and which are ignorable. In terms of supporting this stage of reasoning with an analytical, quantitative and computational approach, we recommend our componentoriented requirements analysis (CORA), and our process failure risk analysis (PFRA) software tools; both subject of other publications (Ranky, 2002a, b, c).
A distributed server network model of robots
.
.
The networking requirements have to be decided. The capabilities of different hardware suppliers should be evaluated. The existing suppliers and their support for the selected network must be evaluated.
.
. .
.
.
The readiness of existing plant and systems for network compatibility must be evaluated. Back-up, safety and security has to be considered. The potential risk of failing processes, exposure to downtime if activities are focused on a single communications network must be discussed. Costs have to be determined. The reliability of the service must be evaluated. A detailed investment appraisal must be conducted. A time-phased implementation plan must be developed.
Based on the above reasoning, for this case, we have decided to follow a simple, nevertheless powerful network architecture, in which we have: .
.
.
.
For the purpose of our analysis, let us assume that we will build a distributed robot network, specifically focusing on automotive welding and assembly applications, in which safety, quality, communication speed and reliability are of utmost importance, due to the cost of downtime and danger to human life in such systems. As discussed by Ranky (2002c), in such developments, staff involved in the networking project has to be educated in the scope of the selected network(s) and the related connectivity issues, implying the following. .
.
three servers (called server_1, server_2 and server_3 in our models), that are linked to our robot lines, as well as to our factory IT system; then linked to a load balancer (called load_balancer), to assure robust and failsafe operation 24 h a day, 7 days a week; that is then linked via a server family gateway (called server_fam_gateway in our simulation model) to the networked robot line, run over an internet/intranet protocol, including a client gateway to the individual robot cell controllers (in networking terms, called “clients”), as well as linking to management computers for remote data analysis, programming, quality control, and other purposes via the company intranet/internet link.
Following our organization-wide engineering network architecture, including the internet, our network is integrated with the robot cell/ line network via sub-networks. Before running this simulation model, and discussing the results, we introduce some basic network design and simulation methods, principles and terminology (following the IT-Guru notation as in IT-Guru, OPNET Technologies, 2002).
Sub-networks In terms of the way IT-Guru handles sub-networks, a sub-network contains other 193
Networked industrial robots
Assembly Automation
Paul G. Ranky
Volume 23 · Number 2 · 2003 · 192–197
network objects and abstracts them into a single object. A sub-network can encompass a set of nodes and links to represent a physical grouping of objects (this can be a local area network (LAN) of robot PC controllers) or it can contain other sub-networks (for example, including the material handling system control of the line). Sub-networks within other sub-networks form the hierarchy of the network model. This hierarchy can be extended as required to model the structure of the network. A sub-network is considered as the parent of the objects inside it, and the objects considered as the children of the subnetwork. The highest level sub-network in the network hierarchy does not have a parent, therefore it is the top sub-network, or global sub-network. Sub-networks can be created and interconnected within this top level, or within other sub-networks (Frazer and Ranky, 2002; IT-Guru, OPNET Technologies, 2002). Sub-networks provide a powerful mechanism for manipulating complex networks by breaking down the system’s complexity through abstraction. As an example, a large network of hundreds of robots with many components can be segmented based on the proximity, connectivity, or other architectural considerations of its elements. For example, sub-network objects can represent networked robot cells, or lines, or design/quality control/robot programming office LANs within an automotive factory. Following this model, within each robot line, sub-network objects can represent each robot cell or design, or programming department sub-network. In IT-Guru, other than the objects it contains, the primary attributes of a sub-network are its geographical position and size. However, these attributes can be ignored when a sub-network is used strictly to abstract other network level objects or when sub-networks exist independently from each other.
Attributes IT-Guru allows the networking engineer to manipulate attributes in many ways, including the following: . setting attribute values for a single object or a group of objects,
.
. . .
promoting attributes to specify their values when a simulation is run, creating new attributes, defining attribute properties, and saving attribute properties for reuse with other attributes.
Three types of attributes may comprise an object. These types can be as follows: . . .
built-in attributes, model attributes, and extended attributes.
Built-in attributes provide basic information about an object. They specify information such as the object’s name, location in the work space, and graphic representation. Built-in attributes are characteristic of each object type. When an object is created in IT-Guru, the program automatically specifies initial values for these attributes. (These values can naturally be edited.) Model attributes add information to a node or link object. They determine how the object behaves in a network model by specifying protocols, data rates, and similar information. Model attributes automatically become part of any object using that model. The model developer defines model attributes and sets initial values for them (Banks and Carson, 1984; Tanenbaum, 1999). Extended attributes are optional model attributes that the model developing engineer can add to an object, thus further customizing its behavior in a network model. Support for extended attributes must be provided in a model by the model developer. In IT-Guru, all attributes share a set of basic properties. Attribute properties define the attribute, specifying information such as its name, data type, and allowed values. These properties are generally fixed, but can be given different values in some cases (IT-Guru, OPNET Technologies, 2002).
Communication nodes Communication nodes exist within sub-networks and represent network devices that transmit and receive information. Node behavior is determined by its node model. The node model specifies the node’s internal structure and can be specified via its model attribute node object, that is said to be an instance of its node model.
194
Networked industrial robots
Assembly Automation
Paul G. Ranky
Volume 23 · Number 2 · 2003 · 192–197
Distinct instances of the pieces of equipment (such as a robot controller, called “client” in our model) are independent of each other in a real network, although they may have identical capabilities. A network model can contain any number of communication nodes, using the same, or different node models.
values. Formatted packets can be read by corresponding communication protocols only. Unformatted packets have no pre-defined fields (IT-Guru, OPNET Technologies, 2002; Kelton et al., 2002; Kirby, 2002; Ranky, 2002c, d; Tanenbaum, 1999). In IT-Guru, packet formats are pre-defined and typically named according to the model in which they are intended to be used.
Communication links and channels Links in IT-Guru allow nodes to communicate with each other using structured messages called “packets”. When a packet is given to a transmitter in a source node, it is conveyed over a link to a receiver in a destination node. A communication channel can be thought of as a pipe, where packets are placed in one end by a transmitter channel and retrieved at the other end by a receiver channel. If a link has multiple communication channels, it can be thought of as a bundle of pipes, each one conveying packets from the source node to the destination node (IT-Guru, OPNET Technologies, 2002). IT-Guru supports two types of links: point-to-point, and bus. Each link type provides fundamentally different types of connectivity. Point-to-point links connect a single source node to a single destination node, whereas bus links connect each node of a fixed set to all others in the particular set.
Packet formats Packets carry information and can be sent between transmitters and receivers. In our example, packets can carry robot programs, when uploaded from the design/programming office servers to the robot lines, and then to the individual robots, or parts of them, if there is a need for an update, or edit, or quality control, production control, maintenance and other data. (Packets can include mission critical, “panic” related real-time data between the robot controller PCs and the line servers.) Packets are data structures consisting of storage areas called fields and can either be formatted, or unformatted. Formatted packets have fields designed according to a packet format, which specifies the packets’ field names, data types, sizes, and default
Simulation runs Having introduced some basic concepts, in order to evaluate critical design cases, as design, and/or operation control and management alternatives, by simulation, we demonstrate the following scenarios. (1) Run our networked robot system without server load balancing, as seen in Figure 1, and therefore consider not to spend the extra resources on the “load-balancer” computer. (2) Next, as an alternative, run loadbalancing in such a way, that the load balancer computer can choose randomly between the available application servers. This model has been configured in such a way, that it utilizes server_3 twice as much as the other two servers. (As we can see from the results in Figure 2, by following this rule, this server will still be utilized less than if it were bearing the entire load.) Althogh there are several other alternatives, for the purpose of illustrating, and proving the benefits of robot (and other) network simulation studies, before the actual systems are committed to, and built, in this paper, we will run IT-Guru for evaluating only these two different scenarios.
Evaluation of simulation results The goal of most simulation scenarios is to evaluate some aspect of a system’s behavior or performance, and to quantify, typically in terms of statistics, the results, and then use the results for decisions. This requires a simulation environment with software tools that provide insight into a model’s dynamic operation.
195
Networked industrial robots
Assembly Automation
Paul G. Ranky
Volume 23 · Number 2 · 2003 · 192–197
Figure 1 Networked system simulation screen in IT-GURU (by OPNET Technologies, USA) running the model without server load balancing. The results clearly indicate that CPU_1 is taking the entire load, and the other two server CPUs are wasted. Considering that, dynamic capacity need alterations, this can lead to a disaster in the given line and even lead to a shut down
Figure 2 Networked system simulation screen in IT-GURU (by OPNET Technologies, USA) running the model with server load-balancing in such a way, that the load balancer computer can choose randomly between the available application servers. This model has been configured in such a way, that it utilizes server_3 twice as much as the other two servers
Note: As we can see from the results, by following the above rule, the server will still be utilized less than if it were bearing the entire load, therefore representing a very safe solution, even if server capacity requirements change dramatically.
196
Networked industrial robots
Assembly Automation
Paul G. Ranky
Volume 23 · Number 2 · 2003 · 192–197
References
Based on IT-Guru’s in-depth analysis, the network engineering analyst can collect object, scenario-wide and global statistics as follows. . Object statistics are collected from individual objects. They allow the network engineering analyst to evaluate the performance of specific network nodes or links (a single hub’s Ethernet delay, or a server balancing change, as in our example). . Scenario-wide object statistics are collected from all relevant objects in a network (for example, Ethernet delay for every node). They allow the network engineering analyst to easily monitor the performance of all objects of a specific type. . Global statistics are collected from the entire network. They represent results that apply to the network as a whole (such as global end-to-end delay) and let the designers and management analyze aspects of the network’s overall performance. More specifically, IT-Guru offers the following types of statistics, when analyzing networks: . . . . . . . . . . .
queue size, available space, overflow occurrences, delay, inter-arrival times, packet sizes, throughput, utilization, error rates, collisions, and application-specific statistics defined by a model developer.
Because there are many possible statistics to collect, the data files would quickly grow past practical use if the simulation program recorded them all. Therefore, the analyst must specifically select the statistics that is valuable for the particular study to collect before running a simulation (IT-Guru, OPNET Technologies, 2002; Ranky, 2002a, c). As in our simplified example, as shown in Figures 1 and 2, based on the plotted graphs and screens, management can easily evaluate the need of the balanced server architecture, and even investigate “what if” scenarios further, without committing themselves to major upfront investments.
Banks, J. and Carson, J.S. (1984), Discrete Event System Simulation, Prentice-Hall, Englewood Cliffs, NJ, USA, p. 514. Carr, J. (2002), “Blueprints for building a network test lab”, Network Magazine, Vol. 17 No. 4, pp. 54-8. Dornan, A. (2002), “The next wave of distributed processing?”, Network Magazine, Vol. 17 No. 4, pp. 38-42. Frazer, A. and Ranky, P.G. (2002), A Case-based Introduction to the National Electronics Manufacturing Initiative (NEMI) Plug and Play Factory Project, An interactive multimedia publication with 3D objects, text and videos in a browser readable format on CD-ROM/intranet by http://www.cimwareukandusa.com, CIMware USA Inc. and CIMware Ltd, UK, ISBN 1-872631-41-x, Ver. 3.0, 2000-2002. IT-Guru, OPNET Technologies, USA (2002), System and User Manuals, Simulation package (used for all simulation studies in this paper), OPNET Technologies Bethesda, MD, USA. Kelton, W.D., Sadowski, R.P. and Sadowski, D.A. (2002), Simulation with Arena, McGraw-Hill, p. 630. Kirby, R. (2002), Of Wired and Wireless, Remember the Checklist, The BugNet Report, Network Magazine, Vol. 17 No. 4, pp. 34-6. Ranky, P.G. (2002a), CORA: Component Oriented (Disassembly and) User Requirements Analysis, An interactive multimedia publication with 3D objects, text and videos in a browser readable format on CD-ROM/intranet by http://www.cimwareukandusa. com, CIMware USA Inc. and CIMware Ltd, UK, ISBN 1-872631-50-9, 2001-2002. Ranky, P.G. (2002b), PFRA: Component Oriented Disassembly (Process) Failure Risk Analysis, An interactive multimedia publication with 3D objects, text and videos in a browser readable format on CD-ROM/intranet by http://www. cimwareukandusa.com, CIMware USA Inc. and CIMware Ltd, UK, ISBN 1-872631-47-9, 2001-2002. Ranky, P.G. (2002c), ”A method for planning industrial robot networks for automotive welding and assembly lines“, Industrial Robot (in print). Ranky, P.G. (2002d), An introduction to computer networking and the internet with engineering examples, A 3D Interactive Multimedia Presentation on CD-ROM with off-line Internet support. Ver. 3.0, published by CIMware, 1998-2002. Tanenbaum, A.S. (1999), Computer Networks, Prentice-Hall, Englewood Cliffs, NJ, USA, p. 658.
Further reading Ranky, P.G. (2001), An Object Oriented Model and Cases of Design, Manufacturing, and IT Knowledge Management Over the 3D-enabled Web and Intranets, July 2001 (in the Proceedings), Multi-lifecycle Engineering Research Center, NJ, USA, INFORMS International Meeting, Maui.
197
Mini features
Assembly Automation Volume 23 . Number 2 . 2003 . 198–206
Mini features Welding of body scanner vessels improved using robot vision system Keywords Welding, Machine vision, Robotics, Seam tracking unit
At the UK factory of Oxford Magnet Technology, 750 liquid helium-filled vessels are produced each year for housing the superconducting magnet that forms the heart of magnetic resonance imaging systems for body and brain scanners. It takes for an experienced operator 14 h to weld each stainless steel vessel by hand. However, following the installation of a turnkey robot welding cell fitted with a seam tracking system from Meta Vision in summer (2001), the cycle has been reduced to just 2 h. ‘‘The system has been coded by Tu¨V in Germany as meeting its pressure vessel regulations. It is the first time that any vessel that has been robotically welded has met a relevant European Union standard’’, said Mark Tullett, Production Engineer in charge of the project at OMT. The Meta Vision weld seam tracking unit has been fully integrated with the robot welding system. It uses laser light to check the position of the torch and adjust it automatically to within ±0.5 mm, a level of accuracy and repeatability that is impossible to achieve by hand (Plate 1). The system comprises the company’s standard MTR tracker linked to a PC, Plate 1 Meta weld seam tracking is fully integrated with the robotic welding system at OMT. It uses laser light to adjust the position of the torch automatically to within ±0.5 mm
enabling continuous compensation for any deviation between the programmed weld path and the detected position of the butt joint. A real-time graphical representation of the joint is shown together with the wire position. The deviation of the wire as well as the joint gap to be filled are displayed in two dimensions in millimetres. Not only is the robotic welding process much faster than by hand, average welding speed being 60 cm/min, but the result is also more consistent. Each new welder hired to operate the three manual TIG cells at Eynsham, which are now devoted mainly to special vessel production, has to be qualified to meet the Tu¨V code, whereas the robotic MIG cell only needs to be qualified once. The automated welding system has eight degrees of freedom, six on a Motoman articulated-arm UP-series robot and a further two CNC axes on a bespoke, 10 tonne manipulator that orientates the 2 m diameter vessels vertically, horizontally or upside down. Orchestrating the computer controlled movements of the entire cell is a Motoman XRC controller that interfaces with the weld tracking system’s PC. One of the problems associated with welding the product by hand is that the code stipulates full penetration of the 6 mm thick 304LN stainless steel, an OMT specification that combines austenitic (non-magnetic) properties with impact resistance at 4 K. As there is a relatively wide gap – typically 6 mm – to fill with metal, six or seven TIG passes are needed to complete the butt weld. These take a long time and introduce variation in the weld seams. In contrast, the robot equipped with the laser sensor is able to MIG weld the joints quickly and consistently in just one pass. When it came to invest in the automated welding system, Mark Tullett and his team drew on experience gained with another successful robotic cell installed at the factory in 1999 for high definition plasma cutting of aluminium and stainless steel parts for body scanner magnets. ‘‘Meta was selected for the project as it is one of the few companies in the country capable of undertaking this type of work. Furthermore, it was prepared to put the effort into jointly developing this one-off application’’, commented Mr Tullett. The success of the installation has prompted OMT to investigate a further cell
198
Mini features
Assembly Automation Volume 23 . Number 2 . 2003 . 198–206
to automate welding of the outer vacuum vessel for MR1 scanners and feasibility studies are under way. For more information contact: Dr Andrew Pryce, Sales and Marketing Director, Meta Vision Systems Ltd, Oakfield House, Oakfield Industrial Estate, Eynsham, Oxfordshire OX8 1TH. Tel: 01865 887900; Fax: 01865 887901; E-mail: sales@ meta-mvs.com
produce melting, rather than unnecessary heating of the metal that can lead to thermal damage and distortion. Other figures of merit are the physical extent of the heat affected zone or the fusion zone size tolerance to a changing base metal temperature. Desktop computer models to quantify these and other figures of merit for the numerous welding processes in use today present a formidable task that has only recently been undertaken[1-3]. For laser beam welding, a dimensionless parameter model[4] has been shown to be effective in relating melting to power, speed, and the material thermophysical properties. By combining this thermodynamic based relationship with additional correlations for penetration depth, weld shape, spot size, and energy transfer efficiency, a computer model of the continuous wave CO2 laser welding process has been developed. The authors say continued development will yield similar models for other welding processes. As the models evolve, better optimization strategies may result in even more robust weld procedure parameters. The authors expect that users of this type of software will increase as the advantages of model based weld procedure selection are realized. Someday, they hope, weld procedure development for automated processes utilizing software such as OSLW will be common.
Weld procedure development with OSLW Keywords Welding, Sandia
Phillip W. Fuerschbach and G. Richard Eisler of Sandia National Laboratories, and Robert J. Steele of the Naval Air Warfare Center have developed computer models for CO2 laser beam welding based on dimensionless parameter correlations derived from solutions to moving heat source equations. They expect their optimization software for laser welding (OSLW) to be extended to many different materials. Finding the best automated welding parameters to achieve a specific weld size on a new material is usually an expensive and time consuming task. To determine a weld procedure, engineers must consider many competing factors including productivity, thermal input, defect formation, and process robustness. The tradeoffs between these factors can be substantial as well as hard to quantify. For example, process robustness might be expected to be inversely proportional to productivity, but in fact, the result depends on the defect being considered. Humping and undercut are defects that occur primarily at high feedrates, however, thermal damage and base metal distortion are deficiencies that tend to occur at lower feedrates. Choosing among numerous weld procedures can be hastened with computer models that find parameters to meet selected weld dimensional requirements while simultaneously optimizing important figures of merit. Two fundamental figures of merit for fusion welding processes are the energy transfer efficiency and the melting efficiency. Energy transfer efficiency indicates what fraction of the energy incident on the workpiece is actually absorbed by the metal. Melting efficiency quantifies the fraction of net heat input to the workpiece that is used to
Notes 1 Reutzel, E.W., Einerson, C.J., Johnson, J.A., Smartt, H.B., Harmore, T. and Moore, K.L. (1995), ‘‘Derivation and calibration of gas metal arc welding dynamic droplet model’’, Trends in Welding Research, ASM, Gatlinburg, Tennessee, pp. 377-84. 2 Sudnik, V. (1997), Modelling of the MAG Process for Pre- Welding Planning, pp. 791-816; Mathematical Modelling of Weld Phenomena 3, Institute of Materials. 3 Eisler, G.R. and Fuerschbach, P.W. (1997), ‘‘SOAR: an extensible suite of codes for weld analysis and optimal weld schedules’’, Seventh International Conference on Computer Technology in Welding, NIST, San Francisco, California, pp. 257-68. 4 Fuerschbach, P.W. (1996), Welding Journal, Vol. 75 No. 1, pp. 24s-34s.
Simultaneous engineering the key to success Keywords Simultaneous, Engineering
When Philips Speaker Systems were awarded the contract to manufacture a new
199
Mini features
Assembly Automation Volume 23 . Number 2 . 2003 . 198–206
micro-loudspeaker for one of the world’s leading mobile telephone suppliers, only 5 months remained for the development of a high-speed assembly system. Only one manufacturer, Mikron Assembly Technology had the experience of simultaneous engineering to make it happen (Plate 2). After detailed evaluation of several system designs and suppliers, Philips selected Mikron’s Flexifactor, high performance cam driven linear assembly system to replace the less efficient rotary machines previously employed. The project was completed on time and the mobile phone was launched successfully. The greatest challenge that faced the Mikron engineers was the degree of accuracy required in the assembly of the tiny loud-speaker. With an overall diameter of only 13 mm and tolerances as small as 0.01 mm the ability of the assembly system to consistently deliver absolute precision while maintaining optimum cycle speed was vital. In particular, the development of a process to transfer, position and solder a 28 micron copper wire with pin-point accuracy required considerable ingenuity. Certain assembly tasks required specific processes only available to Philips, so it was essential to keep those in-house. Thanks to the modular design of the Flexifactor system, the first cell of the assembly line in which most of these operations were to be carried out was delivered to the Vienna factory in only 3 months. This enabled the Philips specialists to undertake the fine-tuning of the processing stations while the rest of the system was being produced in parallel by
Mikron’s engineers in Switzerland. This policy of simultaneous engineering was a major factor in delivering the system on time. The standardised interfaces ensured that the subsequent assembly and commissioning of the completed system at the Philips site was accomplished without difficulty. Throughout the project regular scheduled meetings took place between Mikron’s project management team and the customer, as well as ad hoc discussions when fundamental changes needed to be made. Video conferencing was employed to keep travelling time to a minimum and e-mail was used for the prompt transfer of drawings and technical information. Mikron used a system of actual-to-target comparisons to monitor progress, quality and project costs at every stage of the systems development. This management tool enabled Mikron to keep the processes under control and guaranteed compliance with the customers performance specifications and delivery requirements. For more technical information contact: Richard Krusts, Regional Sales Manager, Mikron Assembly Technology, 74 Newland lane, Ash Green, Coventry, CV7 9BA, UK. Tel: 02476 366071; Fax: 02476 366084; E-mail:
[email protected] Plate 2 Mikron has developed a new manufacturing system for Philips Speaker Systems in Vienna. The system took just 5 months to develop from initial concept to delivery
New automation system keeps manufacturing in the UK at Medic-aid Keywords Automation, Medical
Modular Automation of Birmingham has developed a new plastic insert loading system for Medic-aid. The new system is faster and more reliable than the previous method and has improved efficiency sufficiently to allow the company to keep production in the UK (Plate 3). The system loads plastic inserts into a Negri Bossi 80 tonne press to form part of a moulding for an asthma inhaler. The inserts aid the smooth delivery of the drug through the appliance. According to Ed Waters, Head of Engineering for Medic-aid, the system has made a tremendous difference to efficiency. ‘‘We work in a very competitive environment and seriously considered transferring the entire assembly to China to reduce costs’’, he explained. ‘‘However, this new automated system, together with some significant tooling
200
Mini features
Assembly Automation Volume 23 . Number 2 . 2003 . 198–206
single article is represented visually, showing the aisle, shelf, and floor in which it is located and how far away that position is from the issuing point of the relevant shipping location. At the same time, a red marker indicates which articles have become fast-moving items without having been noticed – or are about to be in short supply. The opportunity is thereby given to those responsible for logistics and inventory to make necessary inventory movements early enough and to prevent bottlenecks at shipping locations. The basis on which this display is calculated includes forecast data, mathematical calculations, and sales figures from the past. Calculations can also be performed using sales figures from the present.
Plate 3 This new assembly system from Modular Automation has improved efficiency sufficiently to enable medical device manufacturer Medic-aid to keep production in the UK
changes, has allowed us to improve efficiency by up to 40 per cent and keep production in the UK.’’ The system first bowl feeds the inserts then collects them by vacuum, eight at a time, using the latest design of Pressflow SNC robot. The robot loads the inserts into the machine’s moulding head immediately before injection. Sensors check that the correct inserts are in place before the mould shuts. Completed parts are unloaded automatically. Previously the operation was a semi-automatic system. This was much slower and was less reliable as it was not possible to check that all the inserts were in place before the mould was shut. In turn this meant that faults had to be picked up during subsequent operations further down the production line. For more information contact: William Bourn, Modular Automation, Talbot Way, Small Heath Business Park, Birmingham, B10 0HS, UK. Tel: 0121 766 7979; Fax: 0121 766 6385; E-mail: bourn@ modular.co.uk
Axxom visualizes warehouse locations graphically Keywords Production planning, Distribution
Axxom Software AG has expanded its distribution tool ORion-Pl1 Shipping Line Balance through the addition of a function that is thus far unique: the entire distribution center is depicted in a two-dimensional graphical display and the position of every
No more confusing shipping floor designations The tracking down of an individual article in the distribution center and the checking of the relevant shipping floor with respect to the question of whether it is still adequate to the current level of order activity can – depending on the size of the distribution center – take a great deal of time. Classical warehouse management systems list inventory and order figures in table form – shipping floors are shown in the form of numeric codes. A better and faster overview is made possible by the new version of ORion-Pl1 Shipping Line Balance: in a graphical display, the articles are shown in their current warehouse locations, and previously unrecognized ‘‘hot items’’ are marked in red. At the same time, the forecast figures for an article are placed in relation to its warehouse location. Using drag-and-drop, the planner can then simulate the appropriate, necessary warehouse movements directly on the PC, i.e. try shifting the articles to other shipping floors. The number of products to be marked in the graphicallydisplayed distribution center plan as potential fast-moving items can be defined by the user. Along with fast-moving items, the degree of utilization of the individual aisles of shelves is also displayed in the graphical distribution center plan. Three different colors are used to mark overutilized, underutilized, and optimally utilized pick stations. Taken into account in the calculation and display are factors such as the time that specific movements and actions by the shipping agents take as well as the number of articles
201
Mini features
Assembly Automation Volume 23 . Number 2 . 2003 . 198–206
to be packed and the length of the path to a particular shipping floor. For more information contact: Axxom Software AG, Cathrin Lessner, Paul-Gerhardt-Allee 46, 81245 Munich, Germany. Tel: +49 (0)89/568 23-377; Fax: +49 (0)89/568 23-399; E-mail: cathrin.
[email protected] likely to increase take up in the future, though it is less than half the price of its predecessor and offers quadruple the speed capacity.
Practicality rules when choosing a servo bus system for machine building Keyword Motion control
The increased take-up of servo systems in all areas of manufacturing and process industry has inevitably resulted in a rise in the use of field buses to synchronise multi-axis motion control. There is a wide range of bus options, but the amplifier and system programming complexity can be a hindrance to the machine builder, the user and the party responsible for maintenance according to Vic Harris, Business Development Director for Industrial Electronics at Wyko (Plate 4). Vic Harris’s selection criteria is governed practically and as well as by performance when it comes to recommending a servo bus and holds/his store out for SERCOS; historically the bus is most familiar to control engineers. ‘‘It was introduced in the early 1990s and was probably ahead of its time. Initial comments on the system were that it was quite costly and not that easy to implement.’’ As with so many emerging technologies, it has taken a while to catch on, but 14 years later, the one millionth SERCOS system chip has just been produced. The introduction of the new SERCON816 chip is Plate 4 Practicality rules when choosing a servo bus system for machine building
Open the door. . . As an open system, SERCOS in theory allows machine builders to fit SERCOS compatible devices from different manufacturers straight from the box. The optical link provides a plug-and-play function that allows devices to be simply attached in series, reducing the cost and complexity of wiring and allowing power electronics to be fitted closer to motors, reducing EMC problems. The reduction in wiring also means, ‘‘Faster build times and reduced panel sizes. Machine builders can then spend more time developing and run-testing a machine rather than the programming taking up the majority of design and development time, as is so often the case.’’ The drive towards standardisation in the servo control layer through SERCOS has been shadowed with the need for integrated machine logic and motion control, reducing application effort and cost. This has been achieved with the IEC1131 standard, which has been adopted by all major equipment manufacturers. The simplicity of a single programming language to solve an entire application means that IEC1131 ladder logic is a powerful tool for machine control programming. It incorporates a complete motion instruction set ranging from cam profiling to interpolated motion and allows the on-line editing of a motion control program, reducing development time significantly. The use of standard IEC1131 ladder logic programming with standard motion control function blocks represents a significant benefit for Vic Harris, ‘‘Not only does this allow fast set-up times because engineers are often more able to use this language rather than more advanced languages such as C++ and motion basic, but it can reduce the cost of programming and maintenance because the skill base is more common. The use of ladder also means there are common building blocks of code available for all types of products and applications making programming even easier.’’ Deskilling the programming aspect of the machine building process also benefits the end user with faster and lower cost maintenance. As Wyko have recently become
202
Mini features
Assembly Automation Volume 23 . Number 2 . 2003 . 198–206
the UK’s largest supplier of industrial electronic asset management services, Vic Harris is interested in keeping maintenance fast and simple. ‘‘Maintenance contracts are often operated on a cost-down basis to limit repairs and downtime, so fast servicing benefits the customer as well as the service provider.’’ ‘‘For example, if a component in the SERCOS control circuit needs replacing, a spare unit can simply be fitted at the same place and plugged in. Because the bus is also a carrier for the programme, all the information is there for the system to recognise the replaced item, and it will automatically reload the programme. Alternative systems would require individual programming of the replaced device, and/or the operating programme to accommodate it, both are time consuming and expensive.’’ Wyko offer repairs on any type of industrial electronic device from drives and machine controllers to servo motors and encoders through over 140 local sales branches. Wyko also provide contract services to several large food and automotive manufacturers where all industrial electronic repair, stock maintenance and monitoring are carried out. For more information contact: Sarah Evans, Marketing Department, Wyko Industrial Services, Amber Way, Halesowen, West Midlands, B62 8WG, UK. Tel: 0121 508 6341; Fax: 0121 508 6333; E-mail:
[email protected] A big linear step forward. . . for small applications Keyword Linear motion
Hepco’s new profile driven unit (PDU) plugs a significant gap in linear motion technology. It is a robust, high performance system designed specifically for applications requiring a cost effective solution to fit into a compact space envelope. Mechanically simple but highly innovative, PDU is very attractively priced and can satisfy the light-load, slow moving application to the more dynamically arduous high speed requirement (Plate 5). PDU is the product of a marriage between proven belt technology, advanced engineering materials and Hepco’s expertise in aluminium profile design. It comprises a slotted profile into which fits a drive belt and carriage. This arrangement effectively seals the unit. But perhaps the most significant element of
Plate 5 The latest innovation from Hepco Slide systems is the new PDU which was introduced in the UK in November at the TEAM exhibition
1
the PDU system is Hepco’s new Herculane wheel technology. These wheels run virtually friction-free on the inside surface of the profile providing stable support for the carriage plate. The unit offers extremely low friction and due to the wheel rolling motion it will run without the need for any lubrication. Hepco guide wheel systems have long been associated with high-speed operation unmatched by ball-based systems. PDU is no exception to this rule. It is rated for operating at speeds up to 6 m/s. At 600 N, the PDU’s load capacity is equally remarkable for a unit that measures just 60 51 mm in cross section. The carriage plate of the PDU is designed to accommodate the mounting of a second PDU unit. In this way a compact, high-performance X-Y system can be quickly and easily created. Integral end of stroke protection is provided as standard. A choice of motors is provided with PDU, offering driving forces of up to 280 N. A typical AC inverter motor will deliver a linear force of 145 N and speeds of over 5 m/s. Optional tubular flanges ensure easy connection to most makes of motors and gearboxes. Drive shafts can be configured left, right or double and other available options include holding brakes, positioning encoders and limit switches. Another time-saving feature of the PDU is the built-in switch cam on the carriage plate that can be rotated to change sides. Naturally, PDU is fully compatible with Hepco’s MCS aluminium profile system and its competitors making it an ideal replacement for failing ball-based systems.
203
Mini features
Assembly Automation Volume 23 . Number 2 . 2003 . 198–206
Typical applications for the new PDU would be the movement of print heads and similar automation needs in food and pharmaceutical packaging. For further information contact: Hepco Slide Systems Limited, Lower Moor Business Park, Tiverton Way, Tiverton, Devon EX16 6TG. Tel: +44 (0)1884 243400; Fax: +44 (0) 1884 243399; E-mail:
[email protected]; Web site: www.hepco.co.uk
of 1.5 m/s and acceleration of 19.6 m/s2. Mounted on this stepper motor is a piezo-electric friction motor which performs the subsequent fine positioning. This motor moves payloads at 100 mm/s over a 10 10 mm area, and offers an optical encoder feedback resolution of 10 nm and a repeatable accuracy of < 30 nm. This combination of speed and precision is believed to be unmatched on the commercial market today. By employing this combination of stepper and piezo motor technology, Baldor believes that OEMs can both boost the productivity of automation, and open up new markets for equipment in areas served to date by application-specific engineering programs. In addition to productivity enhancements, the NanoStepper stage provides its X-Y positioning capability within an exceptionally small profile of just 31 mm. This is around half the height (or less) of a combination of a conventional air bearing stage and a piezo motor for example – a technology currently used by some OEMs producing ultra-fine positioning – and smaller still than any system based on rotary components such as belts or ballscrews and servo motors. A datasheet on the new linear stage is available by E-mail: nanostepper@ baldor.co.uk Downloadable video clips are available on the Web site: www.baldorlinear.com/ video.htm For further information contact: Mark Crocker, Baldor, Mint Motion Centre, 6 Bristol Distribution Park, Hawkley Drive, Bristol, BS32 OBF, UK. Tel: +44 (0) 1454 850000; Fax: +44 (0) 1454 859001; E-mail:
[email protected] Linear motor driven stage brings ‘‘nano-positioning’’ capability to large-travel applications Keywords Nano technology, Linear motor
Baldor has released a unique new type of X-Y positioning stage, which provides rapid sub-micron positioning over very large working areas (Plate 6). Based on a combination of stepper and piezo-electric linear motors for coarse and fine positioning, the stage is available with travels of up to 31.5 m, for payloads of up to 1.8 kg. These features provide scope for OEMs to improve productivity on a wide range of test, inspection and handling equipment involving ultra-fine positioning, such as wafer probing, fibre optic assembly, and biomedical sample manipulation. Called NanoStepper, the stage provides two axes of movement implemented using a novel combination of linear stepper and linear piezo-electric friction motors. The linear stepper motor with its frictionless air bearing provides very fast movement for coarse positioning of the payload, with a velocity Plate 6 Baldor has released NanoStepper a unique new type of X-Y positioning stage, which provides rapid sub-micron positioning over very large working areas
Worldwide adhesive coating manufacturer uses online measurement system to improve production efficiency and quality control Keywords Infrared, Adhesives, Measurement
NDC Infrared Engineering’s 5203TC-4 system (Plate 7) is being used by Northern Ireland based Perfecseal Ltd to accurately measure adhesive coating thickness and has significantly reduced coatweight variations. Perfecseal Ltd are worldwide experts in adhesive coating for various substrates, including films and papers, providing flexible packaging for medical device and healthcare 204
Mini features
Assembly Automation Volume 23 . Number 2 . 2003 . 198–206
Plate 7 The NDC Infrared Engineering system helps Perfecseal achieve more efficient production processes, and meet customer demand for improved quality control
related markets. The system is helping Perfecseal to achieve more efficient production processes, and meet customer demand for improved quality control. Perfecseal adopted a six sigma business model to help the company evolve product quality and productivity in response to customer demand. Six sigma is a disciplined statistical problem-solving methodology focusing on improving processes toward near-perfect quality. The model identifies where wasteful costs exist and helps identify precise steps to take for improvement. Perfecseal developed a project to evaluate the areas where enhancements in the process would result in overall product quality improvements. NDC Infrared’s 5203TC-4 system helped Perfecseal to achieve this quality improvement and provide accurate and real time measurements of its adhesive coatings. The system provides closed loop control so as coating thickness can be accurately and quickly adjusted to maintain product quality and consistency. This heightening of process control brings benefits through eliminating waste and also reducing the overall product variability. ‘‘Previously we had no gauging meaning coatweight checking was done manually. Adding NDC Infrared Engineering’s 5203TC-4 system has made an immediate impact on the quality of our products’’, commented John Muir, ‘‘Black Belt’’ member of Perfecseal’s six sigma team. ‘‘This has been evident through the reduction in scrap, internal product rejection, as well as customer returns. The system has performed well, with
dramatic improvement in coating quality which has seen coatweight variations significantly reduced.’’ The NDC Infrared Engineering 5203TC-4 system is designed for continuous measurement and display of coating weight. The system has four sensors, which are mounted either down or cross-web to calculate coating weights. The model also features a high-resolution colour CRT and touch screen interface. The CRT continuously displays base, total and net coat weight in real time whilst allowing the operator to alternate between display screens. For more information contact: James Millard, NDC Infrared Engineering, Bates Road, Maldon, Essex, CM9 5FA. Tel: +44 (0) 1621852244; Fax: +44 (0) 1621 856180; E-mail:
[email protected] Araldite SMC bonds Scania truck grills Keywords Adhesives, Automotive
Araldite’s SMC bonding range has once again proved its versatility and ability to cope with a wide range of extreme environmental conditions, in an application for Scania Trucks. Polynorm Plastics (UK) Ltd in Merseyside, who specialise in the production of composite parts for the automotive industry, chose Araldite 4910 to bond a front radiator grill and engine hood assembly on a range of Scania Trucks (Plate 8). The bonded joint had to be able to withstand 800 C during the paint cycle and varied weather conditions ( 20 to +80 C, up to 100 per cent humidity) during the Plate 8 A Scania truck
205
Mini features
Assembly Automation Volume 23 . Number 2 . 2003 . 198–206
vehicle’s working life. It also had to be able to resist road vibration for the entire life of the truck, typically 500,000 miles. Araldite 4910 was used to bond Menzolit 5320 Class A SMC substrates in a secondary process, rather than riveting or welding. Bonding provided better vibration resistance than mechanical fasteners and also reduced stress concentrations. Each assembly involved four simple lap joints. Two joints bonded stiffeners running the full length of the assembly on each side. Two smaller inserts were also then bonded. The process was completed in ambient curing conditions. ‘‘We chose to use Araldite 4910 over other adhesives because it gave us a number of processing, performance and health and safety advantages, by eliminating the need for a highly volatile primer’’, explained Gordon McDonough, Industrial Engineer, Polynorm Plastics. ‘‘It was suitable for this application as the adhesive was available in cartridges and required only simple
pre-treatment of the SMC. There was also no heating of the jigs which allowed us to produce parts more cost effectively and still have a short time in handling strength.’’ The bonded joints also successfully withstood a range of tough destructive testing conditions, including heat deterioration, water and humidity tests and thermal cycling, as well as meeting Scania’s own Environmental Standards STD4158 and STD 4159. Polynorm Plastics (UK) Ltd continue to use Araldite 4910 for this application and are currently bonding around 1,000 parts per annum. For further information on the Araldite specialist range of SMC bonding materials and the Araldite 2000 range of industrial adhesives, visit company’s Web site: www. adhesives.vantico.com Further information about Vantico adhesives can be found on the company’s Web site: www.adhesives.vantico.com
206
New products
Assembly Automation Volume 23 . Number 2 . 2003 . 207–211
New products ‘‘FieldGate’’ links fieldbuses to ethernet: Provides simple integration of up to three fieldbus systems into IE networks Keywords Fieldbus, Ethernet
With the new FieldGate module, Softing offers a family of rail-mounted gateway solutions enabling fieldbus networks to link directly with industrial ethernet. FieldGate makes it easy to connect enterprise-wide office and IT networks with automation systems at the plant floor level where fieldbuses are now common. Currently, FieldGate is available with connectivity for PROFIBUS, CAN, CANopen, DeviceNet and FOUNDATION fieldbus. Each FieldGate module has up to three ports for the selected fieldbus, providing ethernet connectivity for up to three independent networks. Compact design, low cost and self-configuration ensure that FieldGate can be added easily to existing communications structures to extend existing networks or integrate new ones (Plate 1). As well as allowing direct data exchange between network hierarchies, FieldGate can provide a link from a higher level PC or central workstation to support tasks such as network configuration, device parameterization, diagnosis and production data acquisition. In effect FieldGate allows a high level computer to ‘‘see’’ a single network Plate 1 A picture, called PROFIGATE, showing FieldGate against an industrial background, is available electronically from www.ggh.co.uk/ clients.htm
structure right down to I/O devices on the plant floor. FieldGate supports ethernet transfer rates of 100 and 10 Mbits/s. Protection class IP20 is met. Because of its compact structure, FieldGate fits easily into narrow switch cabinets. Voltage supply is 24 V DC. For convenient integration into PC networks, a range of suitable OPC servers is available. In addition, an integrated web server offers internet functionality for extra support during commissioning and service.
Loadpoint’s direct drive rotary tables are the first for precision tasks with angular position resolution better than 0.00002 Keyword Bearings
Loadpoint Bearings has launched a revolutionary range of air bearing rotary tables employing a new design of direct drive system to achieve previously unattainable levels of resolution on angular position, better than 0.00002 (Plate 2). This level of accuracy means that the rotary tables are ideally suited to demanding tasks such as dicing silicon wafers and high precision grinding and machining operations. Developed originally for Loadpoint’s own use – and proven in trials on the company’s own dicing machines – the rotary tables are believed to be the first commercially available product of their type to integrate a direct drive motor. The motor itself – a brushless DC torque unit – is a very recent Plate 2 Loadpoint’s direct drive rotary tables are the first for precision tasks with angular position resolution better than 0.00002
207
New products
Assembly Automation Volume 23 . Number 2 . 2003 . 207–211
development. It is key to the air bearing table specification, enabling a design, which is low in height and free from cyclic errors and mechanical wear. The new design resolves a host of issues associated with mechanically driven air bearing tables. Traditionally, the drive on these units has been provided by such devices as worm and wheel, pinch wheel, cable or harmonic drives. Each has associated disadvantages, the harmonic drive is physically large and has limited resolution on angular position whereas the others produce cyclic errors and are prone to wear, both of which impair operating accuracy. The Loadpoint design overcomes all of these problems, achieving a resolution on angular position better than 0.00002 (a parameter which is critically important on large diameter rotary tables) and axial and tilt motion errors of less than 0.25 mm and 0.013 mrad, respectively. The system resolution achieved is a function of the encoder type and the interpolation of its signal by the drive. A wide choice of encoder types offering different levels of resolution means that the user has considerable application flexibility in this respect. Loadpoint is producing the new rotary tables in a range of sizes up to a maximum diameter of 0.5 m. These can be with or without drives to suit different application requirements. A typical specification is shown in Plate 2. The table, 320 mm in diameter, is currently used to support silicon wafers for dicing. It has an axial stiffness of 100 Nmm, a load capacity of 1,350 N and can generate 75 Nm peak, 6.5 Nm continuous torque. For further information contact: Frank Wardle, Loadpoint Limited Chelworth Industrial Estate, Cricklade, Swindon, Wiltshire, SN6 6HE, UK. Tel: 01793 751160; Fax: 01793 750155; E-mail: all@ loadpoint.co.uk; Web site: www.loadpoint. co.uk
and functionality, is now available from Tyco Electronics. Specifically designed to provide superior performance and versatility, and incorporating several unique features – including a quick change mounting fixture for a variety of insertion heads – the floor-standing device reduces cycle times, is perfectly suited to large board sizes or high production volumes, and represents a significant step forward in insertion technology. Unlike any other component insertion machine, Tyco Electronics’ P300 can operate with up to three insertion tools, for three different products, delivering unrivalled levels of flexibility and cost-effectiveness. Incorporating a sophisticated in-line handling system, and semi/fully automatic transfer belt for loading and unloading, the advanced device also offers optional stepper or AC servo motors for X-Y table functionality, and a sophisticated insertion head for significantly reduced cycle times. With a maximum insertion area of 400 600 mm (15.75 23.6 in.), the ultra-reliable P300 features optical sensors for automatic position correction, force monitoring of every cycle, and advanced PCB thickness measurement with adjustment of insertion depth, to guarantee consistent, high-quality insertions. With a PCB turning unit for simple insertion on the solder or component side, the P300 has a standard SMEMA interface enabling easy integration into PCB transfer systems, or use with standard PCB input/output buffers. For further information contact: Holger Nollek, Tyco Electronics AMP. Tel: +49 (0)9851/903-800; E-mail: hnollek@ tycoelectronics.com; Web site: www.tooling. tycoelectronics.com/europe
New insertion machine represents significant step forward in PCB Assembly technology Keywords Electronics, Assembly, Printed circuit board
A new automatic insertion machine, delivering superb levels of speed, flexibility
Modular terminal is ultimate communicator Keyword Pneumatics
Festo has launched a new remote I/O system to accommodate multiple protocols suited to all electrical and pneumatic controller levels. The modular electrical terminal, called the CPX (Plate 3) is compatible with all leading Fieldbus open protocols and adaptable to many customer-specific standards. The CPX terminal dramatically simplifies connections between sensors, valve terminals 208
New products
Assembly Automation Volume 23 . Number 2 . 2003 . 207–211
Plate 3 The CPX terminal dramatically simplifies connections between sensors, valve terminals and electrical controllers, and make costly custom designs unnecessary
and electrical controllers, and makes costly custom designs unnecessary. In many applications, CPX eliminates the need to run sensor and switch cabling back to terminal boxes in a central control cabinet, thereby reducing installation time and costs. The reduced wiring requirement is a practical advantage of particular benefit to customers in the food and process automation industry. CPX features a unique ‘‘sandwich’’ construction comprising a sub-base bus board, an intermediate electronic interface, and a modular connection plate. CPX can be used with the Festo CP installation system to achieve faster cycle times by creating decentralised structures comprising pneumatic valve terminals, the electrical CPX unit and pneumatic actuators. Installation is further simplified by the optimised design, which features a physically separate connection panel and electrical section to allow permanent wiring but still permits fast replacement of the electronics module – ideal for rapid fault rectification. Festo offers configuration software for the CPX unit, allowing customers to specify a single part number for the entire unit, which is then delivered pre-assembled and tested. Electrical modules offer a wide range of connection formats including M12, M8, M12 8-pin, Haraxw, CageClamp or multi-pin D-sub connectors – all to 1P65 or 1P20 classes – and feature a range of selectable digital and analogue I/O configurations. CPX also conforms to the Festo Plug and Workw standard and can accommodate up to nine electrical modules plus a Fieldbus node.
In addition to its connectivity advantages, CPX adds a new dimension to automation projects with sophisticated valve condition monitoring, diagnostic capabilities and remote set up built in. Faultfinding diagnostics for errors operate both at a local level and across a bus connection, improving error traceability beyond node level down to the individual valve and I/O. Clear fault messaging is critical to machine uptime and productivity, so CPX delivers meaningful messages, e.g. ‘‘open circuit’’, ‘‘short circuit’’, ‘‘under voltage’’, etc. The diagnostic capability is combined with a memory function that stores the 40 most recent faults, along with data on the start and end of incidents. Power and communication supplies are isolated to facilitate error diagnostics even after an emergency stop. Powerful set up capabilities also allow users to set parameters for pulse debounce and pulse extension across the Fieldbus. Prices for the CPX are deliberately competitive. List prices are approximately: £9 per digital I/O; with Fieldbus modules from £150. For further information contact: Nicola Meadway, Festo Ltd, Automation House, Harvest Crescent, Ancells Business Park, Fleet, Hampshire, GU51 2XP, UK. Tel: 01252 775000; Fax: 01252 775001; E-mail:
[email protected] Compact, two axis motion controller provides OEMs with benefits of ethernet connectivity at low cost Keywords Ethernet, Motion control
Naples Coombe is providing OEMs with the benefits of compact, two axis motion control and ethernet and RS232 connectivity – all at low cost, on its new Galil DMC 1425 motion controller (Plate 4). The DMC 1425 provides powerful links for easy, high speed and long distance communications between any number of computer hardware devices located on a LAN, including MMI interfaces, PCs, PLCs, motion controllers and vision systems. In addition, the controller also supports the Modbus protocol for easy communication with I/O devices. By supporting the 10-base T ethernet protocol, the extremely compact (95.25 127 mm) DMC 1425 provides a practical solution for decentralised control
209
New products
Assembly Automation Volume 23 . Number 2 . 2003 . 207–211
Main Street, Chaddleworth, Berks, RG20 7EH. Tel: 01488 638488; Fax: 01488 638802; E-mail:
[email protected]; Web site: www.naplescoombe.com
Plate 4 Naples’ compact, two axis motion controller provides OEMs with benefits of ethernet connectivity at low cost
New light curing instant adhesives combine the best of UV and instant curing technologies Keyword Adhesives
applications where space is at a premium. The motion controller supports both stepper and servo motor operation – providing linear and circular interpolation of up to two servo axis – and is equally effective in standalone applications, where its non-volatile memory enables it to be used without a host computer. Although about half the size and price of multi-axis controllers, the DMC 1425 sacrifices nothing in performance, offering the OEM an impressive array of features to handle even the most challenging applications. These include advanced PID compensation, an extra encoder input for electronic gearing, uncommitted digital I/O and analogue inputs, forward and reverse limits, and high-speed position latch. Modes of motion offered include: point-to-point positioning, jogging, contouring, electronic gearing and electronic CAM (ECAM). In common with single axis controllers in the ECONO series, programming of the DMC 1425 is considerably simplified with two-letter, intuitive commands and a full set of software tools such as WSDK for servo tuning and analysis, ActiveX tool kit for visual basic users and a C-programmers tool kit. The package of high functionality, compactness and low cost that the DMC 1425 provides, makes the product, the ideal OEM solution for applications in industry sectors such as machine tools, manufacturing, medical equipment, aerospace, materials handling, packaging, semiconductor manufacturing, food processing, textiles and test and measurement. For further information contact: John Naples, MD, Naples Coombe Ltd,
Henkel Loctite Adhesives Ltd announces a major breakthrough in adhesive technology with the launch of two new light curing instant adhesives. The new range comprises Loctite 4304 low viscosity, and Loctite 4305 high viscosity adhesives. These clear, one-part, solvent-free systems, offer a rapid cure of the exposed adhesive using UV light, whilst achieving an instant adhesive bond without the use of heat or the need for racks to cure products over a long period (Plate 5). Supplied as 28 g or 454 g packs, the new adhesives are suitable for use up to 80 C and have a fixture time of 30-45 s when used to bond steel and only 5-10 s for bonding ABS. The unique combination of extremely high bond strength, together with a rapid and complete cure, provides an ideal means of improving both the output and performance of many industrial bonding operations. Suitable for a wide range of applications including the bonding of polycarbonate windows to ABS housings, tamper proofing and loudspeaker assemblies, their excellent Plate 5 Loctite 4304 low viscosity, and Loctite 4305 high viscosity adhesives, are supplied as 28 g or 454 g packs
210
New products
Assembly Automation Volume 23 . Number 2 . 2003 . 207–211
cosmetic appearance makes them perfect for the bonding of logos and plastic parts in a wide range of industries. They are also ideal for the manufacture of medical equipment and are biologically tested and approved to ISO 10993, for use in medical devices. Henkel Loctite’s new light curing instant adhesives couple the rapid light cure of an acrylic adhesive with a secondary cyanoacrylate cure, effectively combining the adhesion and cure of standard instant adhesives with the depth of cure and surface curing characteristics of UVs. With the introduction of light curing instant adhesives, Henkel Loctite has provided a new range of options for the bonding of many types of materials in a wide variety of engineering applications. For further information contact: Henkel Loctite Adhesives Ltd, Watchmead, Welwyn Garden City, AL7 1JB, UK. Tel: 01707 358800;Fax:01707358900;E-mail:customer.
[email protected]; Web site: www.loctite. co.uk
A motor the size of a pencil tip! Keyword Motors
The maxon RE 8 has been designed for applications which require a powerful motor unit to be packed into a very tight space – in situations where there can be no compromise on precision and reliability. Only 8 mm in diameter, 16 mm long and weighing less than 4 g, it is nevertheless capable of achieving an impressive nominal power rating of 0.5 W. Its ceramic shaft is as thin as a pencil lead – just 0.8 mm in diameter – and turns in two sleeve bearings. The superiority of the ceramic material over steel in terms of its electrical insulating properties allows for a very compact commutator design. Bearing wear is also minimal in comparison with that of systems using steel components (Plate 6). High speed and virtually everlasting life are guaranteed, thanks to the precious metal brushes, which ensure constant and low-contact resistance between themselves and the commutator – even after a prolonged standstill. The ‘‘heart’’ of the motor is the worldwide patented ironless rotor, System maxonw, which benefits from maxon’s unique coil
Plate 6 The maxon RE 8 has been designed for applications which require a powerful motor unit to be packed into a very tight space – in situations where there can be no compromise on precision and reliability
winding. Advantages of its lightweight construction include an unusually low mass inertia and very fast acceleration. There is no magnetic detent at all and also minimal electromagnetic interference. Combined with a neodymium permanent magnet – one of the best magnetic materials currently available – the ironless winding makes the design of these maxon motors much more efficient than those of their competitors. To add to its versatility, the RE 8 is compatible with maxon planetary gearheads – which can be attached for applications where the given motor speed is too high, and the torque too low. It can also be fitted with an encoder system, if needed, although its integral Hall effect sensors will be sufficient for most speed and position detecting purposes. ‘‘Among other uses, the characteristics of the maxon RE 8 make it perfect for ‘light chopping’ in a wide variety of laser-based equipment’’, says Keith Ellenden, CEO of maxon motor uk. ‘‘With its small size and weight, minimal current requirement and exceptional reliability, it is also ideally suited to life-preserving medical applications such as body-worn insulin pumps.’’ For further information contact: maxon motor uk ltd, Maxon House, Hogwood Lane, Finchampstead, Berks, RG40 4QW. Tel: 01189 733337; Fax: 01189 737472; E-mail:
[email protected] Further technical details are also available from the maxon Web site: www.maxonmotor. co.uk
211
Internet page
Assembly Automation Volume 23 . Number 2 . 2003 . 212–213
addresses different aspects of nickel-based welding. Overall, a rather poor site, most of which is still ‘‘under construction’’.
Internet page Keywords Internet, Adhesives, Joining, Welding
http://www.dynea.com Dynea International
http://www.millerwelds.com Miller Electric
Based in Finland, Dynea International develops and manufactures high quality adhesive and surfacing solutions for the wood working and other industries. They are also involved in industrial resins, overlays and oil field chemicals. Dynea has production sites and technology centers in 26 countries around the world, with a total of approximately 3,400 employees. This is a well presented site which is easy to navigate and contain plenty of product information.
Established in 1929, Miller Electric now manufactures TIG, MIG and stick welders; wire feeders; welding guns; plasma arc cutters; resistance spot welding equipment; automation welding controls; welding power generators; and welder training materials. They formed an alliance with Illinois Tool Works (ITW ) in 1993. This is another impressive site which is well worth visiting.
http://www.dow.com The Dow Chemical Company Since the early 1970s, Dow has been providing innovative chemical, plastic and agricultural products and services. Their products include performance plastics for the automotive industry, emulsion polymers, water soluble polymers, polystyrene, polypropylene, epoxys, and a range of agricultural products. They have customers in over 170 countries, serving food, transportations, medicine and health, building and construction industries. An excellent site, containing many product datasheets in PDF format.
http://www.specialmetalswelding.com Special Metals Welding Products Company The Special Metals Welding Products Company supplies a large range of nickel-base welding consumables. This is commonly used for joining high-nickel alloys and highperformance steels, joining or repair of cast irons; and welding dissimilar metals. They also supply a range of consumables for joining stainless steels, aluminium and aluminium alloys, and copper and copper alloys, along with nickel-base filler wires and weldstrip, flux-coated electrodes, flux-cored wires, and fluxes. For 40 years the company has been providing an intensive 3 day forum that
http://www.shure-glue.com Shure-Glue Systems, Inc. A well presented site from a company that offers hot-melt systems and high-quality, low-cost compatible replacement parts for packaging applications.
http://www.polyfoam.cc/ Polyfoam Products, Inc. Established in 1989, Polyfoam Products, Inc. manufactures polyurethane foam and adhesive formulations, along with polyurethane dispensing equipment. Their products include Polysetw, Stormcheckw and Tite-Setw adhesives. The site contains product information, technical data, links to other sites if interest and a list of affiliations.
http://www.loctite.com/ Loctite Corporation Established in 1953 in a basement laboratory and with only a single product, Loctite has grown into a global giant with operations in over 80 countries, a broad range of high-technology sealants, adhesives and coatings, and 4,200 employees. With over 50 regional Web sites to choose from, this site is huge. IT contains a vast on-line product catalogue, a virtual tour through their ‘‘Worldwide Design Handbook’’, and technical articles, data sheets and manuals in PDF format.
212
Internet page
HOT SITES http://www.lordadhesives.com/ Lord Corporation Lord Corporation has three main business divisions, the Chemical Products Division, the Mechanical Products Division, and the Materials Division. Between them they design, formulate, manufacture and market high performance adhesives, polyurethane coatings, electronic adhesives, speciality rubber chemicals, bonded elastomer assemblies, magneto-rheological fluids, and devices to control shock, vibration and noise. This is a very good site, containing some excellent application stories and background information.
Assembly Automation Volume 23 . Number 2 . 2003 . 212–213
http://www.devcon.com/ ITW Devcon Devcon is part of Illinois Tool Works (ITW) Performance Polymers. They manufacture and sell adhesives, sealants and repair products to OEMs, supplying aviation and aerospace, electronics, medical, pharmaceutical, and industrial markets. Devcon’s range of adhesives includes epoxy, methacrylate, cyanoacrylate, anaerobic, and silicone technologies. Their Plastic Steelw has been an alternative to welding and brazing for over 40 years. This is another superb site that contains technical data sheets, application guidelines, material safety sheets, selector guides and much more. Jon Rigelsford
213
Book reviews
Assembly Automation Volume 23 . Number 2 . 2003 . 214–215
include abbreviations, conversions, symbols and signs, and Web sites on plastics. The ‘‘Injection Molding Handbook’’ will prove to be invaluable to students, product and mold designers, process engineers and technical managers alike. Its comprehensive coverage makes it suitable for anyone wanting to learn more about the materials, design considerations and processes of injection molding.
Book reviews Injection Molding Handbook. 3/e Dominick V. Rosato, Donald V. Rosato and Marlene G. Rosato Kluwer 2001 1457 pp. ISBN 0-7923-8619-1 £245 (hardback) Keyword Injection molding
The Welding of Aluminium and its Alloys
The ‘‘Injection Molding Handbook’’ is an immense reference text covering all aspects of injection molding, its related technology and design considerations. Chapter 1 provides a clear description of the Complete Injection Molding Process, from machine characteristics and molding basics, to training programs. Detailed coverage of the many different Injection Molding Machines is given in chapter 2 and followed by a chapter on Plasticizing (the heating and melting of the plastic to be injected into the mold). Chapters 4-6 discuss Molds to Products, the fundamentals of Designing Products and Molding Materials, respectively. They address many topics including plastic melt behaviours, mold components, design and material optimisation, material selections, and recycling. Chapter 7 presents the many methods and techniques available for successful process control. Design Features that Influence Product Performance, Computer Operating and Auxiliary Equipment and Secondary Operations are addressed in chapters 8-10, while Troubleshooting and Maintenance techniques are discussed in chapter 11. The following two chapters discuss Testing, Inspection and Quality Control, and Statistical Process Control and Quality Control, respectively. Chapter 14, Costing, Economics and Management, presents topics including cost analysis methods, financial plant management, and materials management. Specialised Injection Molding Processes are presented in chapter 15, while Injection Molding Competition is addressed in chapter 16. The final chapter of the book provides a useful summary while four appendices
Gene Mathers Woodhead Publishing 2002 236 pp. ISBN 1-85573-567-9 £95.00 (hardback) Keywords Welding, Aluminium
This book provides a clear analysis of the metallurgical principles for the welding of aluminium and its alloys. It is a practical guide which has avoided the use of mathematics to describe the effects of welding. Chapter 1, an Introduction to the Welding of Aluminium, describes the characteristics of aluminium, its product forms (wrought and cast), and common welding definitions. Strengthening mechanisms, aluminium weldability problems, and strength loss due to welding, are addressed in chapter 2, Welding Metallurgy. Material Standards, Designations and Alloys are discussed in chapter 3 and introduce designation criteria, alloying elements, the CEN designation system, specific alloy metallurgy and filler metal selection. The following two chapters address Preparation for Welding, and Welding Design, respectively. Subjects discussed include storage and handling, cutting techniques, cleaning and degreasing, welding speed, edge preparation and joint design, and fatigue strength of welded joints. Chapters 6-9 introduce different welding techniques. These include TIG and MIG welding, plasma-arc welding, laser welding, electron beam welding, and resistance welding. Welding Procedures and Welding Approval are discussed in chapter 10. The final chapter of the book addresses Welding Defects and Quality Control, and describes common defects in arc welding and a range of non-destruction testing methods.
214
Book reviews
Assembly Automation Volume 23 . Number 2 . 2003 . 214–215
Four appendices are included providing information about British and ISO standards related to welding and aluminium, physical, mechanical and chemical properties at 20 C, principal alloy designations for cast products, and alloy designations for wrought products. Overall, ‘‘The Welding of Aluminium and its Alloys’’ is a superbly written text book containing illustrations which are clear and relevant to the text. It is suitable for those new to aluminium welding, as well as students of mechanical engineering and materials science course, welding and production engineers, design engineers and production managers.
and misalignment, while visual inspection, ultrasonic and radiographic testing are commonly used non-destructive testing methods. Chapter 3, Service Conditions, addresses the commonly occurring stresses which affect the lifetime of the system. These include internal pressure, residual stress, misalignment, bending, thermal contraction, and cyclic loads. Failure modes, large-scale tests and coupon test are amongst the subjects presented in chapter 4, Performance Assessment of Butt Fusion Welds in Polythylene Pipes. The remaining four chapters provide Discussion, Conclusions, Acknowledgements and References. The remaining two reports address various aspects of ‘‘Ultrasonic and Radiographic NDT of Butt Fusion Welded Polythylene Pipes’’ and ‘‘Assessment of Ageing Properties and Residual Stresses in Thermoplastic Welds’’, respectively. The third report comprises nine chapters with the first three on Introduction, Objects and a discussion on the Programme of Work. Chapter 4 presents the Experimental Procedure for assessing the ageing properties of the thermoplastic welds and describes the materials, welding techniques, ageing procedures for PEEK (polyetheretherketone) and mechanical tests employed. Chapter 5, Results and Discussion, covers the development of welding conditions for PEEK, ageing studies and residual stresses. The remaining four chapters provide Conclusions, Further Work, Acknowledgements and References. Overall, this is a well written and clearly illustrated reference text. Although it covers specific areas on plastics welding, and will be of most use to people involved in that area, many of the concepts apply equally to the welding of other materials. Jon Rigelsford
Welding of Plastics: Core Research from TWI M.J. Troughton, G.S. Booth, I.J. Munns, G.A. Georgiou, S.M. Tavakoli and R.H. Leggatt Woodhead Publishing (Abington Publishing) 2000 153 pp. ISBN 1-85573-519-9 £115 (paperback) Keywords Welding, Plastics, TWI
This volume presents three research reports from TWI which addresses very specific problems associated with the welding of plastics. The first report, ‘‘Structural Integrity of Butt Fusion Welded Polythylene Pipes – A Review’’ comprises eight chapters and concludes that current tests are inadequate for assessing the service performance of butt fusion welded polythylene pressure pipes. It introduces polythylene pipes and their applications and discusses Defects and Defect Detection in chapter 2. Sources of defects include weld beads, cold welds, contamination at the weld interface,
215
Patent abstracts
Assembly Automation Volume 23 . Number 2 . 2003 . 216–217
At least one of the electrodes is a circular cylinder that can rotate around its cylindrical axis. A suitable electric current is then passed between the electrodes as to cause the parts to be welded together. As the welding electrodes track the line to be welded, the sliding frictional force generated between the surface area of the circular cylinder and the metal part resting against it, cleans the surface of the metal part. The process is particularly well suited for welding sand-blasted molybdenum foils and tungsten wires together.
Patent abstracts Structural polyurethane adhesive Keywords Patent, Adhesives Applicant: Lord Corporation, USA Patent number: US6,423,810 Publication date: 23 July 2002 Title: High strength, long-open time structural polyurethane adhesive and method of use thereof
This patent describes a two-part polyurethane structural adhesive having a long open time, and a method of bonding structural components. The adhesive composition comprises an isocyanate or isocyanate prepolymer and a polyol mixture. The polyol mixture contains a polyol having secondary hydroxyl functionality by way of a long chain secondary polyol, forming at least 75 per cent of the weight of the mixture. The secondary polyols may be a polyether polyols, polyester polyols, polycaprolactone polyols, polycarbonates, acrylic polyols, polybutadiene polyols, polyalkylene triols, or a tetra-ols. The adhesive exhibits at least 25 min of open time at 32 C and cures to high strength under ambient conditions. It can be used to adhere structural substrates such as reinforced plastics, metals, wood, and glass, as used in the construction, marine and automotive industries.
Resistance welding
Ultrasonic welding Keywords Welding, Patent, Motorola Applicant: Motorola, Inc., USA Patent number: US6,176,953 Publication date: 23 January 2001 Title: Ultrasonic welding process
A method of ultrasonically welding a film between two thermoplastic members is presented. Each of the two thermoplastic members have a mating part of a complex joint. The joint comprises a shear joint that cuts away part of the film when the two members come together, and a mash joint that bonds the film and the two members together during the ultrasonic welding process.
Gluing substrates Keywords Gluing, Substrate, Patent Applicant: OTB Group B.V., The Netherlands Patent number: US6,440,240 Publication date: 27 August 2002 Title: Method for gluing optical disc substrates together
Keywords Welding, Patent Applicant: Patent-Treuhand-Gesellschaft fur elektrische Gluhlampen mbG, Germany Patent number: US6,448,532 Publication date: 10 September 2002 Title: Method for resistance welding metal parts
Current resistance welding techniques often result in poor weld quality when welding metal parts with soiled or roughened surfaces together. Therefore, an aim of this invention is to provide a resistance welding process for welding metal parts together where at least one of the parts has a soiled or roughened surface. Initially, the metal parts to be welded together are overlapped and clamped into position between two welding electrodes.
This patent describes a method for gluing two substrate discs together to produce data storage devices such as digital versatile discs (DVDs). The disc elements are held horizontally and positioned above one another so that a gap is formed between them. A liquid adhesive is applied to the inner area of the gap, in the form of separate strands in the circumferential direction on the facing sides of the disc elements. While the elements are brought together, they are rapidly rotated. The combination of the centrifugal force acting on the adhesive and the gravitational force of the upper disc, allows a homogeneous and bubble-free adhesive layer of a predetermined thickness to be created between the disc elements.
216
Patent abstracts
Crack resistant weld Keywords Welding, Patent Applicant: General Electric Company, USA Patent number: US6,410,165 Publication date: 25 June 2002 Title: Crack resistant weld
This patent presents a method for weld at least two dissimilar, metallic alloys to form a weld that is free of cracks. The method incorporates a pure (99.00 per cent minimum by weight) nickel fill-wire, integrally assembled into the joint
Assembly Automation Volume 23 . Number 2 . 2003 . 216–217
between the two alloyed metals to be joined. Welding of the joint results in the formation of a nickel rich region within the weld that prevents the weld against cracks. The alloys joined by this method in the patent are an iron-based, low expansivity, gamma-prime strengthened superalloy, and a high carbon, powder metallurgical tool steel high in refractory metal alloying agents. Such a welded joint can be used in the fabrication of a rotating anode bearing shaft assembly for use in an X-ray generating machines. Jon Rigelsford
217
Diary
Assembly Automation Volume 23 . Number 2 . 2003 . 218-219
ICAR 2003 (C)
Diary
30 June-3 July Coimbra, Portugal 11th Int. Conf. on Advanced Robotics UJC Nunes, University of Coimbra. Fax: +351 2399 406672; E-mail: urbano@ ist.uc.pt
Conferences and exhibitions Key: C = Conference, E = Exhibition, S = Seminar, W = Workshop If you would like further information about any of the conferences or exhibitions featured in the Diary section, please contact the organisers for that particular event. Editorial Note: if you are aware of any local, national or international seminars, exhibitions or conferences, the Editor would be pleased to receive this information as early as possible in order to include it in this section of the journal.
LAMDAMAP ’03 (C) 1-4 July Huddersfield, UK Laser Metrology and Machine Performance Mrs Helene Pickles. Tel: +44 (0) 1484 473266; Fax: +44 (0) 1484 472340; E-mail:
[email protected] VIE 2003 (C) 2003
ICRA 2003 (C) 12-17 May Taipei, Taiwan Int. Conf. on Robotics and Automation Conference Secretariat Tel: +886 2 236 222 09; Fax: +886 2 236 578 87; E-mail:
[email protected]/tw; Web site: www.icra2003.org
Int. Robots and Vision (C + E) Int. Sym. on Robotics. 3-5 June Chicago, USA Heather Straight, P.O. Box 3724, Ann Arbor, MI 48106, USA. E-mail: hstraight@ robotics.org; Web site: www.ifr.org; www.robots-vision-show.info
17-20 June Rhodes, Greece 11th Mediterranean Conf. on Control and Automation K. Valvanis. Web site: http://med03.rasip. fer.hr
ISATP 2003 (S) 9-11 July Besancon, France IEEE Int. Sym. on Assembly and Task Planning Web site: http://cfao.ulb.ac.be/isatp2003/
ISR 2003 (C + E)
MED 2003 (C)
7-9 July University of Surrey, Guildford, UK Visual Information Engineering VIE 2003 Secretariat, IEE Event Services, Savoy Place, London WC2R 0BL, UK. Tel: +44 (0) 20 7344 5476; Fax: +44 (0) 20 7497 3633; E-mail:
[email protected]; Web site: http://conferences.iee.org /VIE2003/
CIRA 2003 (S) 16-20 July Kobe, Japan IEEE Int. Sym. on Computational Intelligence in Robotics and Automation Y. Nakauchi, National Defence Academy. Fax: +81 468 44 5911; E-mail: nacauchi@ nda.ac.jp
AIM 2003 (C) 20-24 July Kobe, Japan Int. Conf. on Advanced Intelligent Mechatronics 218
Diary
Assembly Automation Volume 23 . Number 2 . 2003 . 218-219
S. Sugano, Waseda University. Fax: +81 3 5272 0948; E-mail: sugano@paradise. mech.waseda.ac.jp; Web site: www. aim2003.org
IEEE/RSJ Int. Conf. On Intelligent Robots and Systems Web site: www.iros2003.org/
MFI 2003 (C)
2004
29 July-1 August Tokyo, Japan IEEE Conf. on Multisensor Fusion and Integration for Intelligent Systems Katsushi Ikeuchi. E-mail:
[email protected] MMAR 2003 (C) 25-28 August Meidzydroje, Poland 9th IEEE Int. Conf. on Methods and Models in Automation T. Kaczorek. E-mail: Kaczorek@ isep. pw.edu.pl
ECC 03 (C) 1-4 September Cambridge, UK European Control Conference Fax: +44 (0)20 7240 8830; E-mail: ecc03@ iee.org.uk; Web site: http://conferences. iee.orh/ECC03/
Sensors and their Applications (C) 2-4 September Limerick, Ireland Institute of Physics Web site: http://physics.iop.org/IOP/Confs/ SAXII/
IEEE Sensors 2003 22-24 October Toronto, Canada Web site: www.ieee.org/sensors
IROS 2003 (C) 27 October-1 November Las Vegas, USA
INTERKAMA (E) 16-20 February Dusseldorf, Germany Solutions for Automation in Production and Business Processes Web site: www.INTERKAMA.com
ISR 2004 (C + E) 23-26 March Paris, France Int. Sym. on Robotics Soline Sommers, SYMAP Tel: +33 (0) 1 49 68 54 77; Fax: +33 (0) 1 49 68 54 84; E-mail:
[email protected]; Web site: www.isr2004.com
MACH 2004 (E) 19-23 April NEC, Birmingham, UK Machine Tools and Manufacturing Technology Web site: www.mtta.co.uk
IEEE Int. Conf. On Robotics and Automation (C) 26 April-1 May New Orleans, USA Toshio Fukada. E-mail: Fukada@mein. nagoya-u.ac.jp
GD 2004 (C) 5-9 September Toulouse, France Gas Discharges and their Applications E-mail:
[email protected]; Web site: http://GD2004.ups-tlse.fr
219
