Solving the jigsaw

The need for concurrent design and manufacturing development in the creation of microelectromechanical systems (MEMS) is self evident in the technology’s name. It’s that ‘mechanical’ word. However, during the most recent phase of the MEMS ‘hype cycle’, many ideas and concepts went forward where manufacturability was not fully explored. This is not a universal truth. The MEMS based devices in Texas Instruments’ Digital Light Processors, Analog Devices’ accelerometers and Freescale Semiconductor’s gyroscopes already ship in large quantities to many satisfied customers. But these examples show what big players have achieved using closely guarded MEMS processes, typically developed at and run through their own fabs. By contrast, much innovation in MEMS comes through start ups, research organisations and universities with far less access to process research and deep pockets. “There’s a substantial cost involved in getting the manufacturing right, not just the mask, but also throughout product development. It has to start alongside the initial design work,” says Andrew McCraith, head of marketing and business development for MEMS oscillator start up Silicon Clocks. “Because a new company faces a lot of financial restrictions, you could say that many have come out with [MEMS] products where they had not really done all that work.” The view of this relative newcomer is shared by a leading semiconductor industry veteran. Ray Burgess, former head of Motorola SPS, is now ceo of MEMS switching start up TeraVicta Technologies. He describes why earlier attempts to develop his part of the market failed.