Adaptive Algorithm for Interpretation of Low-frequency Noise DFT

摘要

In the electronics low-frequency noise is often used for reliability estimationudof semiconductor devices. Spectral power density of this noise proportionaludto 1/ f, where γ is the spectral exponent. In [1] shown, that forudspectral analysis of non-white noise exists optimum resolution bandwidth.udTotal mean square error of the estimate is minimal when analysis filterudbandwidth is optimum. In [1] it is shown that DFT could be represented asudresult of processing by filters with different bandwidths by means of frequencyudaveraging.