New sum code for effective detection of double errors in data bits

The authors of this article provides a formation method for the sum code with improved characteristics for error detection in data vectors, comparing to the similar performances of classic Berger codes. Error detection efficiency improvement is achieved due to modification of weighted transition sum code according to specified rules. New modified sum code has a minimum total number of undetectable errors in data vectors, as well as decreased, in comparison with Berger code, number of double undetectable errors of data bits.