Articles techniques et présentations

When studying the behavior of a MEMS or electret microphone it is important to consider the design of the inlet or sound channel (tube) that defines how the acoustic signal gets into the microphone, as well as the design of other internal acoustic components such as acoustic chambers and openings between them. The performance of the microphone, for instance in terms of frequency response and reliability, will depend on such components. In many industrial fields, including professional audio and hearing aids, often lumped element (LE) models [1] are used as a quick and convenient tool to predict the behavior of audio transducers, such as MEMS microphones. These models provide such useful insights as sensitivity to sound and intrinsic noise of the device. These key performance indicators depend on accurate representation of the microphone model by lumped elements including acoustic masses and resistances of internal components of a particular device. To increase the accuracy, end corrections to the acoustic channel length [2], for example, for spouts and sound inlets, need to be included in the LE models. These end corrections describe how an acoustic channel is open to the outside environment, for instance a channel with an open end or a channel that ends with an opening in the infinite plane, called baffle. In this abstract, a COMSOL finite element study has been performed representing two typical cases: an acoustic channel terminated to the open space, representing a spout, being a tube with open end, and acoustic channel terminated to the baffled space, representing an inlet, being an opening in an (almost) infinite plane. The simulation results have been compared to closed form solutions for the end correction values typically used in the field. Nevertheless, to the best of the authors knowledge, only end correction values for the acoustic mass (not for acoustic resistance) have been found in the literature. In both cases of acoustic channel terminated to the open space and acoustic channel terminated to the baffled space the end correction coefficients calculated via COMSOL for the acoustic mass have been found in agreement with the literature. Further, as far as the real part of the acoustic impedance, representing the acoustic resistance of the channel, is concerned, the end correction coefficient is typically assumed equal to the one of the acoustic mass. However, COMSOL simulations reveal that this assumption is not true anymore for the case of acoustic channel terminated to the baffled space. In detail, in case of acoustic channel terminated to the baffled space a difference of 24% between the end correction coefficients for acoustic mass and acoustic resistance has been found.