On one method of formation of optimum sum code for technical diagnostics systems

Abstract

The article provides a method of formation of the sum code with minimum total number of undetectable errors in data vectors. The idea of building this code is based on the principle of weighing of data vector bits, obtaining the weight of data vector and following modifications of its value. New code has the same number of bits in check vectors, as classic Berger code, however, it also has high detection ability and, that is by no means unimportant, in the area of low multiplicities. The article gives the description of modified weight-based sum code characteristics, as well as the results of experiments for organization of concurrent error detection (CED) systems for the set of reference combinational circuits LGSynth'89.