Fault Diagnosis of Delta Sigma ADC using National Instruments-Multisim and Implementation using SVM

This paper targets on the Fault Diagnosis (FD) to optimize ruggedness and accuracy of systems used in circuits of general purpose and industrial applications. Here FD is implemented for delta sigma ADC with all possible faults and covers ability in detecting, identifying and localization of defects. It is designed to check faults in the circuit by treating it as the issue of pattern recognition and clarified using recent Machine Learning (ML) techniques. Functional test (FT) is preferred compared to structural test (ST) for diagnosing complicated circuits as FT authenticates the conditions of the circuit under test (CUT) based on the test cases and specifications of the devices in the circuit. FD using Fuzzy logic, Neural Networks and Fault Dictionary are not accurate or give unconvincing results after diagnosis and may require more time to debug. Support Vector Machines (SVM), a ML approach with appropriate kernel functions optimize the accuracy of diagnosis so that correct repair actions can be taken which reduces the price of product. This recommended method of FD gives improved results than previous methods with the foundation of ML and the concept of generalization.