Temperature drift compensation of a tuned low stiffness MEMS accelerometer based on double-sided parallel plates

This paper presents a self-calibration and compensation method for a MEMS accelerometer whose effective stiffness is electrostatically tuned to about 2.79 N/m with the help of the double-sided parallel plates (DSPP) capacitors. An excitation signal in the form of an AC reference displacement is introduced to drive the proof mass and the response amplified by the mechanical stiffness and the readout gain is demodulated for the calibration of bias drift. Experiments results show that the uncompensated bias offset of the accelerometer has a first-order temperature drift coefficient of 0.066 mg/°C, which is significantly reduced to 0.008 mg/°C by a manual compensation based on the above calibration. The Allan deviation analysis of the long-term measurements shows that the bias instabilities of both the uncompensated and compensated outputs are close to each other and the Allan deviation at an integration time of 3000 s is suppressed from 86.77 μg to 36.36 μg with the compensation. It reveals that the temperature drift of the proposed accelerometer can be precisely calibrated and compensated by the proposed approach.