Ambiguity Sets Determination for Fault Diagnosis of Analog and Mixed Signal Circuits

Abstract

A Simple Binary Classifier for finding ambiguity sets or elements and potetially faulty components in analog and mixed signal circuit under test (CUT) is proposed in this paper. Node Voltages are used as features for classification. The classification criteria is based on the Euclidean distance, threshold and score metrics of the features. Threshold value for classification is estimated based on the performance metric, F1 Score of the binary classifier. Two bench mark CUTs, second order LPF and an 8-bit Digital to Analog Converter are used to evaluate the performance of the binary classifier used.