PSEUDORANDOM TEST-LENGTH ANALYSIS USING DIFFERENTIAL SOLUTIONS

As the size of VLSI circuits increases, the use of random testing is b ecoming more common, One of the most important aspects of random testi ng is the determination of the test pattern length that guarantees a h igh confidence of fault detection, Generally, random test length is es timated by assuming that the set of test patterns applied is purely ra ndom, The assumption is not completely correct in applications where l inear Feedback shift registers (LFSR's) are employed to generate input vectors, In this paper, we have developed a test (Markov) model which faithfully reflects the pseudorandom behavior of test patterns, and a ll detectable single stuck-at faults (instead of the worst single stuc k-fault only) are considered, The required test length is then determi ned by solving differential equations to achieve the specified test co nfidence, Based on the test model, analysis is first dedicated to the two-fault case, results are then extended to the k-fault analysis wher e k greater than or equal to 3. The test length thus determined is sma ller than that derived based on the random pattern assumption, and tes t costs can be greatly reduced.