Digital self-learning calibration system for smart sensors

Design and realization of a digitally controlled closed loop calibration system for smart sensors, capable of self-learning, is reported. Closed loop design enables analysis of sensor properties and optimization of calibration procedure. Dedicated software was developed, enabling control of acquired data and calibration procedure. Resulting statistical data provide closed loop feedback variable for failure analysis. Calibration system enables digital temperature compensation by acquisition of calibration points, calculation of sensor polynomial coefficients and storing the calculated data in the sensor memory. Due to the proposed digital temperature compensation method, temperature output error of pressure sensors was significantly reduced (typically from 0.15% FSO/°C to 0.05% FSO/°C, based on calibration of 29482 manifold absolute pressure sensors). Modular approach results in reduction of calibration time (typically down to 42 s/sensor), thus enabling mass calibration.