Simulating and monitoring the resonant frequency of MEMS for failure detection and prediction

This paper discusses the necessity of failure detection and failure prediction in MEMS devices an proposes a method based on monitoring the resonant frequency. Most of the failures change the Young-modulus, the spring constant or the mass of the moving elements and thus the resonant frequency while others have an effect on damping or cause electrical leakage influencing the amplitude of the vibration. We propose architecture for the circuit, which enables to detect light anomalies and switches the failing or degrading but still functioning parts of the MEMS off. The method and architecture are applicable for many MEMS devices, not only for resonant ones. Finite Element Analysis (FEA), simulations and experiments are supporting the theory.