Fault feature exaction method for the circuit based on Haar wavelet filter banks

According to features of analog circuit test response with pseudo random stimulus signal, fault feature exaction method for the circuit based on wavelet filter banks is presented in the paper. The frequency spectral of the test response signal is decomposed into some parts equably or unequally by Haar wavelet filter banks. Comparing each sub-band signal of response of the different circuit faults, we can find the sub-band with the maximal difference among different fault responses. The fault feature exaction is achieved by the maximal difference sub-band. Because the test response is addressed in different band, fault character exaction method based on wavelet filter group has high precision and high resolving power for parametric fault.