A Presentation for IMAPS France 2001
“Packaging at the Turning Point - Technology and Management Issues”
John W. Balde, Senior Consultant, Interconnection Decision Consulting, IMAPS Fellow, IEEE Fellow, Flemington, New Jersey U.S.A., balde@IDConsulting.com
Abstract
Packaging is more important than ever, but packaging skills are declining at the OEM companies. Not only manufacturing but design is being outsourced, and packaging skills are moving to the Contract Manufacturers/Designers. But they only do state-of-the art, and do not innovate. Innovation remains at vertically integrated companies, mostly in Japan, but also is moving to Design and Development organizations such as Fraunhofer and IMEC. But there is a tendency to take only small steps - to avoid major changes that would produce big improvements that would mostly accrue to second and other applications thereafter. But only the vertically integrated company makes decisions for the future. Some way has to be found to introduce new technology such as IMPS and LC Polymer materials to electronic packaging.
Background - Outsourcing
We live in a time of great challenges for Electronic Packaging. In fact, it is a time of crisis.
Not because of innovation in packaging - there are many new technologies that are promising and important - MEMS, System-on-a-Package, Bluetooth applications, improving RF cellular design, new electronic applications. Rather the crisis is one of the way packaging is carried out.
Packaging has never been more important. A cellular phone, not quite well designed, can have significantly poorer performance, and have a hard time succeeding in the marketplace. Mainframe Computers newly designed can indeed offer greatly increased performance at lower cost. Many electronic systems and products are better, cheaper, and available in the market earlier than ever before. So what’s to worry about?
It is the way packaging is now carried out that causes the worry.
Outsourcing - The new business model
At the same time that Packaging has been generally recognized as critical to success in the marketplace, the packaging activities of OEMs are screeching to a halt. Company after company is outsourcing not only their manufacturing, but their packaging designs also. Look at the table of activities. Large PE companies have done two things - they have either divested themselves of major portions of their manufacturing, or have outsourced production manufacturing to others.
Whole plants and activities have been cut free - IBM letting go of Printers (Lexmark), Flat Cable (International Flex circuits), PCMCIA CARDS (Celestron), Military electronics (sold to Loral). Lucent has sold off many plants and will sell others soon. Whole parts of their manufacturing such as IC fabrication are no longer a part of the company.
But more important are the outsourcing of the manufacture of products still considered part of the companies necessary products - DEC selling off its printed circuit plant years ago and buying printed circuits, but NEC having its multilayer boards made by Hitachi, Ericsson selling its printed circuit plant to Solectron, etc. But these are components, you say?
We move to a different level when the assembly of the final product is outsourced = Qualcomm having its phones made by Kyocera, Ericsson having its phones made by Flextronics, Motorola, Lucent, many other products made by contract manufacturers.
This is risky for survival reasons. Outsource cellular phone manufacture to Kyocera and very soon Kyocera knows more about the phones than Qualcomm does. It can decide to introduce product under its own name - without Qualcomm, competing with Qualcomm.
But it can do and did something else. If it has the manufacturing capability, and Qualcomm does not, it can decide not to go it alone, but to buy the company that originally outsourced to it. So we have seen Kyocera buy Qualcomm, and take over the product niche. The more common and more insidious effect - loss of design control.
Almost every company outsourcing manufacturing and then design believes it can still innovate at the OEM level by thinking of new design of products, which they will subsequently delegate to manufacture. But that is VERY difficult. Improving a design comes from intimate knowledge of the problems of manufacturing the existing design. Work with a product, solve the problems, and one quickly finds out what works and what does not. Any knowledge of improvements comes to the contract house, not to the OEM.
Any contract manufacturer will tell you that the products as they manufacture them have many changes initiated to reduce costs or simplify manufacturing, or just to make the design work. These improvements become the property of the Outsource house - they are not required to provide these improvements to another company for future bidding for the manufacture of this product. A company is locked into the outsourcer that it chooses early on.
Note, just as these improvements are the property of the outsourcer, the knowledge is not well passed back to the OEM company outsourcing the product. The OEM designer can think up new configurations for the product (better email screens for a cellular telephone, different applications for a Bluetooth module) - but the skill in designing those modules and products resides at the contract engineering house.
Outsourcing Companies and Activities are Becoming Larger Than Many OEMS
Look at the table. Solectron pays billions of dollars to buy new facilities to keep up with demand. In a downturn or tight money market, the Contract Manufacturer takes facilities off the hands of the OEM - and never sells them back. Flextronics expands to increase its manufacturing capabilities. But many other houses increase their work for OEMs - Kyocera, Kitron, Celestica, Sanmina, Xetel, Promex, Ibiden, etc., all are the major manufacturers today - and increasingly competent.
But Outsourced Engineering is state of the art -
NOT INNOVATIVE
This seems paradoxical. If they are so competent, why do they not innovate in technology? The answer is in the contracting process. An outsourcer or contract manufacturer is initially contacted to make a product conceived at the OEM company, and expected to offer a competitive bid for the manufacture, USING THEIR BEST STATE OF THE ART TECHNOLOGY. They may indeed sharpen and hone that technology, but will not be supported or authorized to try out some new manufacturing scheme.
If a Sheldahl had the capability of making high density flex for system-on-a-package applications, there is no contract from the outsource manufacture to try out this technology instead of some more conventional one. If manufacture can be doubled in density using the University of Arkansas IMPS technology - the Contract house does not try this. The same thing for Liquid Crystal Substrates instead of polyimide.
Why is this?
It is because original exploratory effort is undertaken by a large OEMs whenever future applications promise a big payoff, BUT THEY KNOW THAT THE SAVINGS ON THE FIRST APPLICATIONS MAY NOT YET BE REALIZABLE.
IBM developed the multilayer co-fired ceramic technology because it needed the interconnect capability not possible with organic multilayer boards. NTT developed the additive copper on polyimide organic board technology because telephone applications need the increased capability. IBM Japan developed the surface laminar additive technology to make possible finer lines and higher density needed by today’s electronics.
But these were actions of vertically integrated companies - whose ability to recognize the need for innovation was driven by the end products of the company. There is no such far-out focus for contract manufacturers.
Coping Without Vertically Integrated Companies
If we see more and more designs outsourced, there are fewer companies able to serve as the developers and first users of new technology to pave the way for their future needs. Japan has those companies - and rightly preserves the possibility for innovation.
But even Japan does not do so without help. JAPAN has MITI, the Ministry of Technology and Industry, which funds exploratory programs for new technology. It will put out for bids on some new developments, and the results of the exploratory work are published and made available to vertically-integrated companies to pursue and further develop. And the technology is adopted.
The US does the same, through DARPA contracts to demonstrate new technology. But there is a difference - there is not the implied push for some company to use that technology thus demonstrated. The answer is there, looking for an adoptee. But what does the rest of the world do?
The Europeans have a good answer
In Europe there has been a rise of companies that do innovative design and development for various OEMs. Look at the table. Leading technology centers such as Fraunhofer IZM in Berlin, IMEC in Belgium, and IVF in Sweden do advanced development work for many clients, and truly innovate. Other companies such as Kitron do advanced work, and some advanced work happens at the University of Linchping and other places.
That same development work is less in the US. MCNC in North Carolina has done contract design work, notably for Rockwell, with some innovations, as has the HiDEC center of the University of Arkansas. Georgia Tech, for all its highly visible industry leadership in new materials and understanding of electronic design, has strangely done very little in contract design for clients. The leadership is clearly in Europe.
But there is still a problem in funding for innovation.
Fraunhofer and IMEC get the funding, but so many others are expected to design only using the state of the art. There are innovative new technologies that have been explored and found worthy that have difficulty getting used. There is an OEM tendency to be comfortable only being second! Let someone else be first; then we will jump in. But if everyone wants to be second, there is no progress! SOMEONE HAS TO BE FIRST.
Let me list some of the possibilities that are waiting on the vine to be picked.
Underutilized New Technologies:
IMPS Technology, Integrated Mesh Power System
IMPS Technology, the proposal of the University of Arkansas, uses the space between signal traces on a circuit board to place power and ground traces. These power and ground traces would be woefully inadequate to provide low impedance support wiring unless they are interconnected at every intersection of vertical and horizontal traces.
This is possible because laser drilling produces the vias, and the vias are virtually free. Pin together the network of power and ground leads at every circuit crossing and the result is almost equal to the performance of the more usual supporting power and ground planes in a multilayer board.
But a multilayer board is no longer needed to get good performance. Put the power and ground IN THE SIGNAL PLANE, AND A TWO SIDED CIRCUIT CAN PERFORM AS WELL AS A 4 LAYER MULTILAYER BOARD, as long as the density of the 4-layer board is not required. Suddenly a two sided flex circuit can replace the boards of MCM-L modules. Cost savings are 50%!
Liquid Crystal Polymer Substrate Material
Another development waiting for a first user is the use of Liquid Crystal Polymer instead of Polyimide for these flex circuits and for other boards. The liquid crystal is not only 1/5 the cost of polyimide, it has 1/5 the moisture absorption, 5 times the stability or characteristic impedance, and 1/5 the expansion forces that cause via misregistration. Reduce the variations in X-Y dimensions and the vias can be downsized, and the circuit traces can be closer together. The end result is a very great reduction in manufacturing costs, with an increase in circuit capability.
No takers. Why? It DOES require doing something non-standard, which might save much for future designs, but may not realize those savings for the very first application.
MEMS devices have been developed to produce variable capacitors. But one of the principal applications has yet to be developed. If RF MEMS devices can be used to effect frequency control at the RF frequencies of wireless telephony, the stability of the circuits can be improved. MEMS is for analog work, and the RF circuits of phones are analog.
This one is probably near fruition - possibly as I speak there will be an application. But it has taken longer than normal without the OEMs as the driving force.
Continuing Development Needed Accurate and Stable Passives in Organic Technology
MEMS for electronics - oscillators, varactors replaced by voltage sensitive switches, etc. The work is taking place, but the utilization is infrequent.
Conclusions
Accept the leadership of Japan as the location of the only remaining vertically integrated companies.
Or support and develop the design centers = the Fraunhofers of the world. Hope for increased similar activity in the US.
Be aware of the risks of delegating design competence to others - and the possibilities of losing the company to the contract manufacturer.
For more information on IMPS and Liquid Crystal Technology see:
"Packaging Technologies That will Control the Future," a Keynote Address by J.W. Balde at Flextronics Conference, Sunnyvale, September, 1998.
| Table of Contents | Previous Page | Next Page |
|
