PORTLAND, Ore.—The world's smallest MEMS chip—a tiny 700 square micron digital microphone—was fabricated recently by Bosch Sensortec GmbH for its subsidiary Akustica Inc. The CMOS die integrates a mechanical microphone diaphragm in a sea of mixed signal CMOS circuitry that culminates in a standard PDM (pulse density modulation) output.
The MEMS microphone market topped 695 million units in 2010, up more than 57 percent from the 441 million units shipped in 2009, according to IHS iSuppli (El Segundo, Calif.), which predicts that shipments will rise to more than 1.7 billion units, or $471 million, by 2014. Applications for MEMS microphones include mobile phones, touchscreen tablets, PCs, laptops, netbooks, video- and audio-recorders, Bluetooth headsets and camera modules.
The first chip to come from Akustica since its 2009 acquisition by Bosch Sensortec, the design team was able to squeeze another 30 percent out of the die's area—which had already been shrunk from two- to one-millimeter square—to qualify the new digital microphone as having the world's smallest MEMS die at just 700 square microns.
"Bosch's experience in shrinking high-performance circuitry enabled us to reduce the size of the die while also maintaining its low noise and improving is sensitivity," said Akustica's CEO and general manager, Stefan Finkbeiner. "For directionality and noise cancelation, it is also thin enough that you can fit multiple mics into the bezels of any PC, laptop, netbook or even the smallest tablet computer."
Akustica also plans to downsize its current 3.8-by-4.7-by-1.3 millimeter package further for handheld devices wishing to fit high-fidelity MEMS microphones into spaces too small for conventional electret microphones, such as ultra-thin cell phones. The 840-by-840 micron die is just 380 microns thick, allowing it to comfortably fit in even the smallest industry standard packages.
Using Bosch's CMOS process, Akustica has shrunk the size of its die by 80 percent, yielding five-times as many chips per wafer.
Akustica's advantage going forward will be the ability to fit more mics into a given form factor, thus allowing more products to benefit from microphone arrays, with attendant ambient noise reduction on communication up-links and headphone and Bluetooth earphone products. With more mics, talker tracking becomes easier in conference situations.
Would there be any cost implications due to the technological rigour involved? What would be the impact on the yield? And will the discount due to more chips per wafer be helpful in bringing down the cost?
It seems there is not a big architectural breakthrough but just another shrinking in geometry. So, I would expect more companies can build the same kind of MEMS mic in this size and soon the record can be broken by a smaller geometry. An era of mic array is here!
Your reasoning is sound-and I concur-but Akustica has been ahead in die size from the beginning and keeps shrinking it ahead of competitors. Most cell phones, however, are buying their MEMS mics from Knowles Acoustics-their traditionally supplier of electret microphones. Akustica has yet to crack the cell phone market, but that may change now that they are a part of MEMS giant, Bosch Sensortec.
I could see the move to smaller phones and perhaps using an array of mics for improved audio due to noise reductions (as stated above). The benefit is multi-faceted: smaller die = lower cost and better placement options due to PCB layout flexibility. does the new mic have better or worse frequency response?
A smaller microphone could be quite useful in some previously unexplored applications. It could reduce the area needed in phones, BUT the one thing that this world would not gain any benefit at all from would be a smaller cell phone. My reasoning is thus: the microphone will be even farther from the speakers mouth, and the ear-piece volume will still be low, as a way to save battery power. The result will be even more people jabbering loudly on their cell phones, sharing details of their conversations with all the passers-by, who have no desire at all to be involved with the private lives of these cell phone users. So use the new small devices for new medical applications and new security systems applications, and even for new video game console applications, but not to make a smaller cell phone.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.