In the area of MEMS microphones the world is moving to high signal to noise ratio (SNR), according to Gunar Lorenz, head of technical marketing for consumer sensors at Infineon Technologies AG. This is because the growth applications, particularly voice interface equipment will require it. The improved membrane design is one way his company will achieve it, he said.
Speaking at the MEMS and imaging sensors summit organized by SEMI Lorenz characterized three generations of moving membrane capacitive MEMS microphones.
Early MEMS microphones had a membrane that moved in relation to a single back plate with a typical SNR of 65dB.
Current designs have dual perforated backplates – or rather a front plate and rear plate – with the membrane in between them and moving in relation to them to increase the sensitivity and dynamic range to about 70dB.
The third generation is the sealed dual membrane where what were the backplates in the previous design are unperforated and become the moving the membranes and what was the membrane becomes the static backplate in the middle of the sandwich. Lorenz said this arrangement could achieve 75dB. In voice activated applications this could translate to 98 percent accuracy in word recognition versus 85 percent with the current configuration.
By being sealed this provides protection against dust and moisture ingress which were previously disadvantages of the MEMS capacitive approach.
When asked about piezoelectric MEMS microphones that have already been used to achieve high SNR and dust and moisture protection, Lorenz said such MEMS have challenges with drift and temperature variation, with reliability and phase shift.
The next improvement will likely come from software and algorithms, said Lorenz in response to a question from his audience.
Related links and articles: