[previous] [next] [contents]    Test vectors for short circuit detection in full BST interconnects

The set of test vectors presented in table 1 was produced by a modified "self-diagnosis" algorithm and guarantees complete short-circuit fault detection capability. This algorithm also guarantees that the sequence of logic values captured in any short-circuited interconnect (the so-called STR, sequential test response) is different from the sequence of logic values applied to any interconnect (the so-called Sequential Test Vector, STV). For diagnostic purposes, this is equivalent to say that aliasing syndromes cannot occur, since any faulty SRV never belongs to the set of STVs. The other type of syndrome challenging fault diagnosis occurs when disjoint sets of short-circuited interconnects produce the same SRV. This other type is known as confounding syndrome and its avoidance requires a larger number of test vectors.

ptv[0]: 010101010001000001010101
ptv[1]: 010101010010000010101010
ptv[2]: 011010100001000001011010
ptv[3]: 100110100001000010100101
ptv[4]: 101001100010000001100110
ptv[5]: 101010010010000010011001

Table 1: Set of test vectors generated for short-circuit fault detection among full-BS interconnects.

The 6 horizontal test vectors shown in table 1 represent the logic values applied simultaneously to the 24 interconnects of this type (each bit stream has 24 bits) and are therefore normally referred to as Parallel Test Vectors (PTVs). The STV applied to each interconnect may be read in each column, starting with the STV applied to net 0 in the leftmost column.