Abstract
This paper describes a fault identification algorithm for combinational and full-scan sequential circuits called FLOSPAT - Fault Localization by Sensitized Path Transformation [1,2]. The goal of fault identification is to localize a fault to the fewest possible gates and to determine the Boolean functions realized by those gates. Instead of choosing a fault model, FLOSPAT uses fault-independent sensitized path tracing [3] to localize functional deviations. Sensitized path transformation is used to adaptively generate test vectors which improve the diagnostic resolution. The output of FLOSPAT is used for physical defect diagnosis by cross-referencing gate-level defect dictionaries generated by the contamination-defect-fault mapper CODEF [4,5,6].
