Efficient pattern generation for transition-fault diagnosis using combinational circuit model

An efficient diagnosis procedure to distinguish non-equivalent transition faults and identify equivalent transition faults is proposed. This procedure consists of three main methods. The Fault Inactivation Method (FIM) generates diagnosis patterns to distinguish fault pairs by inactivating one fault and detecting the other for each fault pair, the Fault Pair Filter Method (FPF) quickly identifies a large portion of equivalent-fault pairs after FIM, and the Fault Propagation Method (FPM) generates diagnosis patterns for the remaining distinguishable fault pairs and identifies equivalent-fault pairs by initializing both faults in each pair simultaneously and creating distinguishable faulty responses. Experimental results show that only 6 out of more than 7.48*10⁸ fault pairs in ISCAS89 benchmark circuits cannot be handled by this method, i.e., a diagnosis resolution of higher than 99.999999%is achieved.