Research Of Temperature Compensation And Optimal Closed-loop Design For High Precision Capacitive MEMS Accelerometer

Micro-electro-mechanical systems(MEMS)accelerometer is a kind of inertial sensor used for acceleration measurement.Advantages such as small size,low power consumption,high compatibility with IC and batch fabrication make it possible for silicon micro-machined accelerometer to be widely utilized in civil and military markets.This thesis focuses on the study of real-time temperature compensation and closed-loop system optimization of high-precision capacitive MEMS accelerometer.To reduce the bias drift against temperature,a real-time temperature compensation algorithm is proposed.And aimed at reducing power consumption,an analog circuit based on diode-quad bridge for capacitive sensing is designed and realized.Besides,an automatic test system of amplitude-frequency and phase-frequency characteristics is designed to quickly calibrate the open-loop characteristics of the accelerometer,under the guidance of which the closed-loop system is further optimized.The major work and contributions are given as follows:1)Two different capacitive sensing method,dual-channel modulation method and diode-quad bridge demodulation method,is analyzed,compared,and realized in miniaturized PCB,with the latter highlighting simpler circuit form and saving power consumption by 85%.The open-loop accelerometer utilizing ring diode circuit has quite good performance,with the equivalent noise acceleration of 12.6 μg/√Hz,and the bias instability of 5.9 μg.2)A real-time temperature compensation algorithm for a force-rebalanced MEMS capacitive accelerometer which relies on the linear relationship between the temperature and its dynamical resonant frequency is proposed.An extra phase-locked loop(PLL),where we inject a drive signal to excite the accelerometer into resonance and track the real-time resonant frequency,is added to the conventional force-rebalanced loop to compensate for the temperature-induced output drifts.Measurement results demonstrate that the bias offset of the accelerometer has the first-order temperature coefficient of 3.54 mg/℃,and that is significantly reduced to 0.05 mg/℃ after compensation.Besides the improved bias instability of 1.6 μg,the long-term drift is suppressed more than one order of magnitude in the Allan deviation plot,from 35 μg to 2.4 μg after compensation,which reveals the restraining of the temperature-related performance deterioration.3)Aiming at the problem that the dynamic parameters of accelerometers fluctuate greatly due to the error of micromachining and thus closed-loop parameters of single accelerometer need to be manually adjusted,the automatic frequency sweep test scheme for accelerometer’s dynamic parameters is designed.The accelerometer’s open-loop characteristics can be quickly obtained using the scheme,laying the foundation for later batch calibration and fast closed-loop design.After the frequency sweep test,the open-loop characteristics analysis and closed-loop research are carried out in the accelerometer system based on the ring diode detection scheme.The theoretical open-loop parameters agree well with the measured characteristics,and a customized PI controller to optimize pole zero placements is accordingly designed for the force-rebalanced closed loop.The characteristics of the closed-loop accelerometer are also tested using the automatic sweep scheme with a measured bandwidth of 37.2 Hz,which exactly matches the settings in the closed-loop design.