Wavelet Analysis for the Detection of Parametric and Catastrophic Faults in Mixed-Signal Circuits

Methods for testing both parametric and catastrophic faults in analog and mixed-signal circuits are presented. They are based on the wavelet transform (WT) of the measured signal, be it supply current (I-PS) or output voltage (V-OUT) waveform. The tolerance limit, which affects fault detectability, for the good or reference circuit is set by statistical processing data obtained from a set of fault-free circuits. In the wavelet analysis, two test metrics, one based on a discrimination factor using normalized Euclidean distances and the other utilizing Mahalanobis distances, are introduced. Both metrics rely on wavelet energy computation. Simulation results from the application of the proposed test methods in testing known analog and mixed-signal circuit benchmarks are given. In addition, experimental results from testing actual circuits and from production line testing of a commercial electronic circuit are presented. These results show the effectiveness of the proposed test methods employing the two test metrics against three other test methods, namely, a test method based on the root-mean-square value of the measured signal, a test method utilizing the harmonic magnitude components of the measured signal spectrum, and a method based on the WT of the measured signal.