A Fault Dictionary Method for the Diagnosis of Catastrophic Faults Using Power Supply Current

Abstract

Fault diagnosis of analog circuits is still more difficult than fault diagnosis of digital circuits. It is a bottleneck in the development of electronic circuits. Catastrophic faults are the major failures of analog circuits. Output voltage is traditionally used to diagnose analog circuits. A new signal, named dynamic power supply current is proposed for fault diagnosis of analog circuits in this paper. And an AC fault dictionary method is presented. Firstly, the dynamic power supply current of analog circuits is sampled under an ac input stimulus. Then eight fault features of power supply current are extracted. These fault features include four features in time-domain and four features in frequency-domain. Secondly, the ambiguity sets of every fault feature are built and every ambiguity set is assigned with an integer code. Thirdly, these fault features and potential faults are used to construct the integer-coded fault dictionary. The elements of fault dictionary are the integer codes. This fault dictionary can be used to diagnose the circuits-under-test. As an example, the proposed method was used for fault diagnosis of a band-pass filter. We diagnosed this filter using the dynamic power supply current and the traditional output voltage, respectively. The diagnosis results of two kind of signals were compared. The comparing results indicate that the dynamic power supply current is suitable for fault diagnosis of analog circuits.