Using ordered partial binary decision diagrams, forced basis nodes, and the blocking method for observability to calculate bounds on testability analysis measures

by Andrew H. Crofts

Institution: Texas A&M University
Department: electrical engineering
Degree: MS
Year: 2012
Keywords: electrical engineering.; Major electrical engineering.
Record ID: 1957322
Full text PDF: http://hdl.handle.net/1969.1/ETD-TAMU-1995-THESIS-C76


PARFAIT is a testability analysis (TA) program that introduces two novel concepts to increase speed and accuracy of results. The first is the blocking method, a linear algorithm for calculating observability that targets single path propagations while also accounting for multiple paths that are near the fanout stem. Second, forced basis nodes are created by making selected primary inputs constant in an effort to eliminate reconvergent fanout. This allows many nodes to achieve total reconvergence status and be used as pseudo-Pls to achieve increased accuracy in functional representation. PARFAIT is one of the few TA programs that provides bounds on testability measurements. By using ordered partial binary decision diagrams (OPBDDS) to represent functional information, the reconvergent fanout problem is minimized and the accuracy of the results is known. In addition, it is the only TA program to combine this feature with a user-defined tradeoff between speed and accuracy.