A Background Test and Calibration Technique of Nonlinear Mismatches in MEMS Oscillating Accelerometers

MEMS accelerometers have benefits of low cost and small volume; however, their long-term stability is limited by the fabrication mismatches. This article proposes a self-test and calibration technique for differential MEMS accelerometers, to reduce the effect of flicker noise in reference voltages due to nonlinear stiffness mismatch in two channels. Through detecting the frequency response induced by the sinusoidal test signal, the oscillating amplitude of the two channels is properly adjusted to meet a theoretically derived noise cancellation condition. A numerical model is established based on the practical mechanical parameters of the MEMS structures to simulate the proposed system and optimize the key design parameters. Finally, a prototype including a MEMS accelerometer chip and an FPGA-based control circuit is implemented to verify the suggested approaches. Compared with that without the proposed technique, the 1-h bias instability of the silicon oscillating accelerometer (SOA) is reduced from 6.31 to
$3.44 \mu \text{g}$
. At different reference voltages from 1.09 to 1.33 V, the decrease in instability is between 21.7% and 45.5%.