Aliasing errors in multiple input signature analysis registers

Aliasing errors in linear feedback shift registers used as multiple-input signature-analysis registers (MISARs) for self-testing networks are investigated analytically. The authors derive the final value for aliasing errors for a MISAR with the same probability of an error on each of the different inputs. They also derive the final value for aliasing with the same register if there are different probability values of an error on each of the different inputs. This analysis assumes that the different inputs are all independent of one another. The analysis is based on the Markov model. In either case, the final value is shown to be 1/2^k where k is the length of the shift register