A High-Precision, Low-Power Signal Demodulation System of Charge-Based Sensor Based on the Fusion of FPGA and AFE Circuits and Its Application in the Detection of Piezoelectric Accelerometers

Charge-based sensors are widely used in vibration testing, inertial navigation, rotational speed testing, and so on. The output signals of the charge-based sensors are usually weak, so modulation and demodulation methods are often used to bring the low-frequency output signals to high frequencies to reduce low-frequency noise and interference and then demodulate the high-frequency carrier signals to obtain useful information. Most of the demodulation systems use multiplication methods, where amplitude fluctuations in the reference signal affect the stability of the demodulated signal and lead to errors, and resource consumption of the multiplier leads to high power consumption. To address the above problems, we propose a high-precision, low-power charge-based sensor signal demodulation system based on the fusion of field-programmable gate array (FPGA) and analog front-end (AFE) circuits: The FPGA circuit uses a switching demodulation and an infinite impulse response (IIR) low-pass filter for demodulation. The AFE circuit uses Hanning-coordinate rotation digital computer-discrete Fourier transform (Hanning-CORDIC-DFT) algorithm for demodulation. The data judgment algorithm controls the disconnection of the two circuits by determining the variance of the output signal to achieve low-power demodulation with a minimum power consumption of 79 mW. The data fusion algorithm uses an adaptive variable-weight Kalman filtering fusion algorithm to fuse the outputs of two circuits to achieve ultrahigh-precision signal demodulation and improve other demodulation performance. Take vibration acceleration measurement as an example. The demodulation error of the FPGA and AFE circuits is 0.8504% and 0.836%, respectively. After data fusion, the error is reduced to 0.63%. The system provides a high-precision and low-power solution for charge-based sensor signal demodulation.