Refine
Document Type
- Article (4)
Language
- English (4)
Has Fulltext
- yes (4)
Is part of the Bibliography
- yes (4)
Keywords
- MEMS (4)
- Mikrofon (4)
- Flip-Chip-Technologie (1)
Institute
This paper describes a novel method for trimming a low noise, differential micro-electromechanical system (MEMS) microphone with a state of the art signal to noise ratio (SNR) of 66 dB(A) and a sensitivity of -38 dBV/Pa @ 1 kHz after completing the fabrication process. The method allows compensating the electroacoustic variations caused by process tolerances and reducing specification limits of the microphone sensitivity from ±3 dB to ±0.5 dB. Trimming is done by programming the bias supply voltage of the capacitive membrane-backplate-system and the gain factor of the output pre-amplifier. Microphone sensitivity can be modified within a range of 11.2 dB. Thus the standard deviation of the sensitivity of a production batch could be reduced from 0.97 dB to 0.11 dB. Furthermore the SNR could be increased to 66.5 dB(A).
MEMS microphones must be further miniaturized for use in mobile devices. So we developed packaging technologies for MEMS microphones using flip chip technology instead of wire bonding. The first two generations have the sound hole in a ceramic interposer at the bottom side of the package. Now a new flip chip microphone package with the sound hole on the top side has been developed. The new package technology combines a large acoustic reference volume for good signal to noise ratio with small size. Compared to other microphone packages the sensitivity is improved by 3 dB. The front volume is small to avoid resonances in the acoustic frequency range.
Capacitive MEMS microphones are now widely used for mobile phones and headsets. The frequency response of MEMS microphones is influenced by the package geometry. This can be applied to tune the performance of the microphone in the system. Packages with small sound inlet holes are acoustic low-pass filters. The cut-off frequency is determined by the diameter of the holes. In previous work we did not find sufficient agreement between a lumped element simulation and the measured frequency response of the microphones. Now we measured the acoustic resistance of small sound inlet holes directly. The measured resistances agree well with calculated values. The frequency response of MEMS microphones was measured for different sound hole diameters and lengths. The results are in good agreement with calculated transfer functions of the package.
In this paper a new method for sensitivity recalibration of capacitive MEMS microphones is presented. Recalibration can be applied to compensate ageing or environmental influences. Recalibration can be done by measuring the sensitivity for only one bias voltage after a stress test. A microphone with a variable bias voltage can measure its pull-in voltage. Unfortunately, the drift of the pull-in voltage does not correlate with the sensitivity drift. So a true self-calibration without a defined acoustical test signal was impossible.