Bio-Inspired Flow Velocity Microphone: Acoustic Simulation of Possible Encapsulating Packages

Abstract

With the emergence of voice assisted devices in the consumer’s electronics industry like smart speakers and wireless earbuds, there is a need to design voice capture systems that are more robust to background noise and reverberation. Current directional far field audio capture systems are  based on omnidirectional microphone arrays or pressure gradient microphones with dual ports that provide bidirectional polar pickup patterns. The fundamental sensing mechanism of such systems  is based on ‘acoustic pressure’ sensing which is scalar quantity in nature, with no directional information. On the opposite, the use of thin bio-inspired nano hair follicles as sensors (mimicking the jumping spider auditory sensing mechanism, for example), enable  “acoustic particle velocity” sensing that is vectorial quantity carrying spatial information. This inherent directional nature of acoustic particle velocity enables the development of  novel and compact directional microphones.

The current research focuses on the acoustic design and optimization of the sound ports and signal paths for the encapsulating packages that will house the flow velocity element. It aims at defining the optimal package design to successfully integrate this kind of velocity microphone into various consumer use cases by retaining or improving the acoustic performance of the bare (a.k.a. un-encapsulated flow sensing element.  Several package configurations have been investigated, starting with the ideal  “out-of-plane” sensing (open pipe channel with inlet and outlet port), while the technical challenges and the possible design recommendations were derived from multiphysics modeling conducted using COMSOL (pressure acoustics & thermoviscous) simulations. The key acoustic performance targets of MEMS microphone, such as shifting resonant frequencies above audio frequency range to achieve flat sensitivity and directivity, for such out-of-plane configuration are presented with appropriate validation against analytical expressions.