Deep reasoning approach to sequential circuit fault diagnosis

The authors describe an algorithm for the location of faults in synchronous sequential circuits. It is based on the `deep reasoning´ approach to circuit fault diagnosis. The application of the deep reasoning or `first principles´ approach to circuit fault diagnosis has been shown to be an effective and powerful diagnosis technique when applied to combinational circuits. However, its application to sequential circuits has proved to be difficult and the candidate generation procedure has been reported to become indiscriminate when applied to this kind of circuit. The authors show how the deep reasoning approach to circuit fault diagnosis can be applied to synchronous sequential circuits by incorporating the circuit´s output dependency on the present state of storage state devices into the diagnosis process. They achieve this by creating a combinational model of the sequential circuit with the feedback loops broken and by introducing probing points to measure state values. The application of the extended diagnosis procedure to the combinational model of a sequential circuit gives encouraging results