A strategy for detecting faults in sequential machines not possessing distinguishing sequences

The problem treated here is that of detecting faults in digital equipment by applying input sequences at the input terminals and observing output sequences at the output terminals. The checking of digital equipment by input/output tests applied at the terminals is motivated by current and future usage of large-scale integration techniques which make internal test points generally inaccessible for testing purposes. The modeling of digital equipment by finite-state sequential machines and then designing fault-detection tests based on the state table is a general approach. The difficulty is that it results in very long experiments for large state tables, particularly for the case in which the state table does not possess a distinguishing sequence. that is, an input sequence for which the response uniquely identifies the initial state. This paper presents a strategy for designing more efficient fault-detection tests for machines not possessing distinguishing sequences.