Bounding signal probabilities for testability measurement using conditional syndromes

Abstract

An algorithm for bounding the random pattern testability of individual faults in a circuit is proposed. Auxiliary gates for bounding the testability are constructed, converting the problem into one of determining the signal probability at the output of the auxiliary gate. The results presented are in terms of lower bounds of the testabilities of faults. The bounds generated by the algorithm can be used by designers to identify pseudorandom pattern resistant faults, to enable them to modify the circuit structure to make the faults easy to detect, and, hence, to increase the fault coverage. >