A new algorithm for diagnosis-oriented automatic test pattern generation

Production testing does not only aim at detecting faulty devices, but its goals are often to repair the element or to investigate the cause of failure, so as to tune the manufacturing process. Diagnostic testing is thus becoming the object of attention both in industry and academia, thanks also to the increased power of tools like fault simulators, testability analysers, and ATPGs. Diagnostic testing has two aspects: assessing the diagnostic properties of a given test pattern set or generating test patterns having such properties. This paper deals with the latter aspect. An ATPG algorithm, the Δ-algorithm, generating a pattern able to distinguish between two faults, is described and its preliminary results obtained on a set of benchmark circuits are reported