Digital non-parametric characterization method for MEMS resonant sensor

This paper reports a new digital circuit architecture to identify the resonant frequencies of a micro-resonant sensor. A generic and configurable model of micro-resonant sensors has been used to evaluate the efficiency of combining a pseudorandom excitation with a non-parametric harmonic retrieval algorithm all within a digital characterization circuit. Estimation of the resonant frequencies is based on the eigenanalysis of the autocorrelation matrix of the sensor´s output. An original discrimination scheme compensate for the over-estimation on the number of harmonics. Results indicate that this architecture can spectrally identify resonant sensors with up to 4 mechanical resonant modes with an accuracy of a 100ppm on the frequency estimation. Implementation of this characterization circuit would allow micro-resonant sensor identification, calibration and default compensation