Analog Survival Means Learning To Be Adaptable

A dapt! That’s always a good notion in a period of change. It becomes a matter of customer relations, long-term vision, product pull from deep in the future, and the ability to draw a roadmap from here to there. For example, traditional suppliers of silicon-based analog products are adapting their product lines to microelectromechanical systems (MEMS).

Analog Devices and Wolfson Microelectronics surprised me last fall with briefing requests related to MEMS microphone announcements. That wasn’t completely odd. In 2006, Analog Devices acquired AudioAsics, a Danish company, with intellectual property related to high-end hearingaid microphones. In January 2007, Wolfson acquired Oligon, a MEMS sensor startup, which has intellectual property related to manufacturing microphone membranes and related technologies.

At the time, I couldn’t do much except connect two dots: ADI’s announcement and the AudioAsics acquisition, which I’d learned about at the 2006 Audio Engineering Society conference, but written about in the context of the challenges of managing remote design centers. I asked myself whether it made sense for semiconductor companies to pursue micromachines at all. As it turns out, they definitely should make that effort.

In fact, you will find that analog semiconductor companies are pursuing other non-traditional, complementary areas. But for the moment, let’s concentrate on MEMS microphones and approach the story of their development as a paradigm. At that point, I wondered why these companies were investigating MEMS mikes at all. The technology didn’t seem to fit with their main product lines.

Wegener said that eventually, those handheld original device manufacturers (ODMs) think you will line up to buy products that record surroundsound to be played on hometheater systems. To capture that surroundsound, he noted, the handheld products need phased microphone arrays, and electrets are too big and delicate. They need MEMS, he said.

It also wouldn’t hurt if they could perform lossless compression in hardware ahead of their DSP, and that compression is Samplify’s forte. Well, that’s an extreme blue-sky vision. Today’s MEMS mikes are nowhere near that compact.

Yet even in this generation, one can sense a trend in that direction. Both ADI’s and Wolfson’s briefing materials included foils on beam-steering. ADI explains that beam steering is key to using multiple mikes in portable automobile GPS receivers. The beam-steering enables the GPS to ignore extraneous audio sources—cell-phones, radios, road noise—to better pickup driver voice commands.

“The ADMP421 consists of a MEMS microphone element, an output amplifier, and a fourth-order S-? modulator. The digital interface allows for the pulse density modulated (PDM) output of two microphones to be time-multiplexed on a single data line using a single clock,” according to ADI’s ADMP421 digital-output MEMS microphone’s datasheet (see the figure).

This indicates the intention to use pairs of the mikes in applications that could include beam-steering. As for size, the ADMP421s have a 3- by 4- by 1-mm form factor, so they’re not ready for large arrays, but the potential is clearly there. For more, see the datasheet at www.analog.com/static/imported-files/data_sheets/ADMP421.pdf.

Wolfson’s WM110, WM120, WM110E, and WM120E MEMS mikes are analogoutput devices. The WM120s have a smaller footprint, and the “E” versions (“E” for “enhanced”) have a tighter ±1-dB sensitivity tolerance, microphone-to-microphone. That’s a breakthrough spec emphasized by Wolfson in the briefing, noting that it eliminates a time-consuming calibration step during manufacturing for OEMs that are implementing beam-steering in applications such as GPS systems.