Robust digital calibration engine for MEMS inertial sensor systems

The precise measurement of acceleration is usually performed by MEMS capacitive sensors, which provide very low signal amplitudes. Therefore, a robust and highly accurate read-out system is required to process these output signals. In this paper we present an innovative concept for relaxing the requirements on the analog components of the read-out system by using digital calibration techniques. The design uses adaptive filters in a correlation-based calibration system, which continuously detects errors and corrects the digital output signal without interrupting normal sensor activities. Permanent and time-variant linear gain errors can be corrected and the complete sensor system becomes more robust against changing ambient conditions. The proposed concept has been successfully verified to correct different gain errors in a 350nm CMOS sensor interface.