Déjà vu all over again: It feels all so familiar with MEMS poised to cross the chasm just like analog did 20 years ago.
A perfect storm is brewing in semiconductor design and it reminds of a period about 20 years ago when many of the same forces were at work. The storm is all about challenges and opportunities to expand a market for a specific type of semiconductor technology, and the main thing missing is coherent infrastructure or ecosystem to facilitate it.
I’m first referring to the time when analog design emerged from being a black-magic art controlled by a relatively small group of wild haired geniuses who were generally kept isolated from mainstream logic designers. The designers at that time had not just heavy electrical engineering talent, but also possessed deep understanding of process physics and mathematics. They used arcane tools and methodologies and no one could quite figure out how they got the analog and mixed-signal designs they created to work. Analog was just plain different than digital logic design.
There were no real standard practices and EDA tools were just emerging to support this craft, especially in digital design. The breakthrough tool was Analog Artist from Cadence and it provided a much more structured way to develop analog designs with the use of a design environment. The challenge that was met over time was to link the electrical design with the actual physical implementation. This allowed for faster design cycles and designs that were better optimized for the process capability. Another big missing piece was a proven infrastructure, particularly in the area of semiconductor manufacturing process, to make analog design an efficient, predicable and more economically viable practice. Luckily, the commercial foundry business model was really taking off and companies like TSMC were looking for new things with which to differentiate their fabs. They were happy to work with the tools vendors to help fabless designers figure out how to incorporate analog information into the flow.
I was at Cadence then and we worked long and hard not just on tools, but the whole set of requirements to build an analog design process from front to back. Our biggest breakthrough may have been the process design kit (PDK), which we developed in conjunction with the leading manufacturers like TSMC. It sounds simple now, but the notion of a standard way to describe and incorporate critical proprietary process technology for the AMS designer was revolutionary at the time. Today the PDK and accompanying parametrized layout cells (Pcells) are an accepted way of life.
Flash forward twenty years and exchange MEMS for analog-mixed signal design. The opportunity for MEMS is ripe as companies eye the virtually untapped consumer market to apply MEMS capabilities to (think gyroscopes and accelerometers in iPads or Nintendo Wiis, and every cellphone on earth – a huge step forward from sensors in airbags and ink jet printers). The research firm iSuppli just came out with eye-opening forecasts for sensors in smart phones, predicting a quintupling of the market for this one category thanks to an increasingly demand to support motion command and navigation applications. Overall, most analysts peg the MEMS market to be a $15 billion market in the next five years.
But the growth is tied to consumer applications and that will require significant improvements in how MEMS are designed. The tools are there and there is a lot of IP floating around but there is no structured ecosystem in terms of the design methodologies and manufacturing processes. It’s an ad hoc process with a lot of manual hand-offs and fine tuning required. The costs and time frame associated with developing MEMS (as high as $40 million and as long as 10 years) are out of whack with the dynamics of the consumer market. It feels a lot like analog design circa 1990.
MEMS’ biggest liabilities are 1) the lack of a seamless connection to other parts of the IC flow because MEMS must interact with electrical components; 2) the traditional “one product, one process” mentality that doesn’t efficiently scale or support design re-use; 3) the need for foundation IP; and 4) specialized foundries, that don’t allow for a fabless business model. These factors have limited MEMS to a tightly controlled domain of expert designers and specialized manufacturers. Hmmm, sounds similar to where we were with analog 20 years ago.
With MEMS poised to cross the chasm from a market opportunity standpoint, the time has come to democratize how we design and manufacture these amazing devices. At this year’s Design Automation Conference, there was a “Birds of a Feather’ Meeting on this very topic, and representatives from the MEMS industry, foundries, EDA suppliers and other interested parties came together to discuss the issues involved in proliferating MEMS-enabled design.
The meeting, albeit a small one, is a good sign. We need the type of collaboration between tool vendors and manufacturers that resulted in the PDK breakthrough in analog twenty years ago. We need the ability to exchange IP and manufacturing data in a standard way, re-use process technology for multiple applications, and generally make MEMS design more accessible to traditional custom IC designers.
A foundry ecosystem with integrated design environment, reference flows, silicon IP libraries, and “PDKs” will break the “one process, one product” tradition, and MEMS design will become more accessible to fabless companies. It will have the added benefit of generating income for the foundries and drive some business to the mature process nodes that MEMS is likely to involve. We will see MEMS integrated in SoCs, truly achieving systems on chips.
The good news is that things are happening in this space. Coventor has taken the lead from a tools and flow standpoint, launching a platform that links to the custom IC electrical world and the system design space. Their tool suite hides the complexity of the multi-physics, 3D models and exotic materials used in MEMS from designers mostly familiar with electrical circuitry. TSMC has done a 180-degreee reversal on its stance on MEMS, now firmly embracing it in its roadmap and actually producing MEMS for a new wave of fabless MEMs companies such as InVenSense, which came out of nowhere to become the largest supplier of gyroscopes to the consumer market last year. It’s also encouraging to see the recent “more than CMOS” partnership between IMEC and TSMC that includes MEMS as a key focal point for its collaboration.
I am bullish on the MEMS opportunity and can envision history repeating itself based on the analog design revolution that I was a part of. The same will happen in MEMS and the end result will be more feature-rich products that enrich our every day lives in ways we can only dream of today. The promise of things to come is still a source of wonder for me.
About the author:
Jim Hogan has worked in the semiconductor design and manufacturing industry for more than 35 years gaining experience as a senior executive in electronic design automation, semiconductor intellectual property, semiconductor equipment and fabrication companies.
Hogan currently is the managing partner of Vista Ventures LLC and is an active strategic consultant to public and private technology companies. Prior to Vista, Hogan was a general partner at Telos Venture Partners and senior vice president of business development at Artisan Components Inc., now part of ARM Holdings PLC. He held senior executive engineering, marketing and operational management positions at Cadence Design Systems, Inc., National Semiconductor Corporation and Phillips Semiconductor. At Cadence, he was an executive fellow, president of Cadence Japan, corporate vice president of marketing, and corporate vice president for field operations.
Hogan holds a B.A degree in mathematics, a B.S. degree in computer science and an M.B.A all from San Jose State University. He serves on the Board of Advisors at San Jose State's School of Engineering, and on the board of directors of several technology startups.