Today Micro Electrical Mechanical Systems (MEMS) microphones are available in a range of electronic consumer products such as smart phones, tablets, smartwatches and surveillance cameras. The MEMS microphones are usually attached to a circuit board with a hole that lets sound propagate through, as well as additional wave-guides which alter the MEMS microphones original frequency response. The MEMS microphone and additional wave-guides are in the same size order as the thermal and viscous boundary layers. These are called non-ideal losses and are usually not considered when dealing with large scale acoustical systems. The only way to predict the impact of these losses is the use of Finite Element software.

The objective of the work is to model the thermoviscous losses when the waves propagate through narrow regions. The system of study is the Axis Network Camera P1367 and the study focuses on the acoustic path into the microphone. The first aim is to model the acoustic path along with the MEMS microphone to produce a frequency response that matches the measured frequency response of the different configurations for the sound-guide. A second aim is to find the configuration which produces the most desirable frequency response.

Several measurements with different configurations were made, such as varying the length and radius of the sound-guide hole. All measurements were performed in an an-echoic chamber. Thereafter, a FEM model was created of the simplified acoustic path and the different configurations that were performed in the measurement were compared with the simulated results.

The simulated frequency responses differ in terms of where the resonance frequency occurs, but the configurations of the sound-guide match the overall behavior when comparing the simulated and measured results. The most optimal configuration of the acoustical path is obtained.

The simulated model requires more work in terms of obtaining a better matching frequency response, most importantly the MEMS cavity. The real geometry of the MEMS sensor cavity did not produce the same frequency response as the one in the data sheet for the specific microphone used for this study, thus a fictive cavity was introduced to produce the desired frequency response. The model did succeed in capturing the overall behavior as well as when the configuration was altered.