An algorithm for diagnosing two-line bridging faults in combinational circuits

A deductive technique for diagnosis of bridging faults

“…In about 25 of the cases, the diagnosis is partial; that is, only one of the lines of the bridge can be determined with certainty. In such cases, and if the resolution is so large that bridging fault simulation cannot be performed, then the diagnosis procedure can be followed with other techniques [5,6,7,8] using the candidate list to improve the resolution. Note that these resolutions were obtained using a compacted stuck-at test set.…”

“…Further, the graph reduction procedure is linear in the size of its vertex set (jVI j). Techniques such as those used in [5,6,7,8] require either the storage of stuck-at fault dictionaries or the simulation of stuck-at faults during diagnosis. As seen in columns 12 and 13 of Table 2, the storage requirements for dictionaries can be very large, and the simulation time is about an order of magnitude larger than that required for the diagnosis procedure.…”

“…The corresponding node sets are shown within the curly brackets. The intersection graph has similar structure to the initialization graph proposed by Chakravarty and Gong [8]. However, there are important differences.…”

“…Further, the construction and storage requirements of fault dictionaries may be prohibitive. Chakravarty and Gong [8] describe a voltage-based algorithm that uses the wired-AND (wired-OR) model and stuckat fault dictionaries. The wired-AND and wired-OR models work only for technologies for which one logic value is always more strongly driven than the other.…”

“…The path-trace and the intersection graph are constructed and processed dynamically during diagnosis. [5,6,7,8] can be used to improve the resolution.…”