Efficient Design of Self-Checking Checker for any m-Out-of-n Code

The use of self-checking checkers in the design of highly reliable systems has many significant advantages. It allows errors to be detected upon occurrence without testing, whether the error is caused by a permanent or intermittent fault. However, there are relatively few codes for which efficient self-checking checkers have been designed. In this paper we present procedures for designing efficient self-checking checkers for m-out-of-n codes (i.e., codes where each valid code word consists of m bits with value 1 and n – m bits with value 0). Codes for arbitrary values of m and n are considered. Realizations which require significantly less logic than previously known realizations are presented for all cases except n = 2m and m = 1. These checkers are totally self-checking for all single and unidirectional multiple stuck-type faults. They are also very easy to test compared with previously presented realizations. Saving in logic complexity and testing complexity of 70 to 97 percent is demonstrated.