The principles of inductive fault analysis (IFA), a technique for the determination of a list of the possible faults in an integrated circuit, ranked according to their probability of occurrence, are reviewed and criticised. It is pointed out that IFA should be upgraded to achieve reasonable flexibility in dealing with different MOS technologies and that the Monte Carlo approach, used for determination of fault probabilities, cannot practically provide reliable fault ranking. A possible approach to more algorithmic and less technology-dependent IFA and an alternative to the Monte Carlo method for evaluation of fault probabilities are discussed, together with a simple application example.
|Tipologia ministeriale:||Articolo su rivista|
|Appare nelle tipologie:||1.1 Articolo su rivista|