How fabless works for MEMS: A chat with InvenSense’s VP of business development

July 31, 2012 -- Jengyaw "Joseph" Jiang, VP of business development, InvenSense Inc. (NYSE:INVN), spoke with Solid State Technology about the company’s fabless approach to micro electro mechanical system (MEMS) innovation and its Nasiri-Fabrication (NF)-Shuttle, which launched in 2012 and will soon do a third run.

Photo. Joseph Jiang, InvenSense.

The NF Platform integrates CMOS and MEMS on silicon. The InvenSense NF Platform has been used to produce over 200 million MEMS units to date.

NF-Shuttle allows where mask costs to be split amongst multiple users that purchase real estate on the same mask. The users can validate designs on silicon with a much lower upfront cost than prototyping on their own. InvenSense launched the first shuttle run in January 2012 with select universities, including Stanford University and University of California Berkeley (UC Berkeley). Its second run occurred in May, and the company is currently preparing for a third in December. Stanford and UC Berkeley secured spots on both of these additional shuttles.

Figure. InvenSense's NF Platform.
(Click the image to enlarge.)

The December shuttle is the first that will be opened up to the industry at large. “Opening our proprietary MEMS fabrication process to industry and universities benefits the whole MEMS industry,” said Jiang. The platform can be used for pressure sensors, accelerometers, microphones, etc. It is very difficult to develop MEMS currently, with the 1 product/1 process model, Jiang noted.

As a fabless company, InvenSense works with multiple foundries (GLOBALFOUNDRIES and TSMC) to produce its MEMS devices. This combines fairly standard, high-volume manufacturing processes with innovation in the MEMS design. By sharing the shuttle with other MEMS developers, InvenSense expects the participants to have lower engineering costs and gain access to top-tier foundries. Smaller MEMS companies often work with smaller foundries, which cannot be cost-competitive with the top-tier foundries, nor offer the same capacity, Jiang said. InvenSense benefits by having its manufacturing platform gain adoption throughout the MEMS industry, and licensing it to other MEMS makers.

The fabless model also frees InvenSense from some of the worries about disruptions, such as those from natural disasters, slamming its supply chain to a halt. This is a clear advantage in the semiconductor space, which we are applying to MEMS, Jiang said.

MEMS have always had to compete on cost, size, and performance, much like semiconductors. As MEMS hit milestones in these areas, volumes are steadily increasing. In the future, securing foundry capacity in larger wafer sizes with high yielding processes will need to be a primary goal of MEMS makers.

-- Meredith Courtemanche, digital media editor, meredithc@pennwell.com