We are using a matrix to detect depressions of a membrane keypad. From a perspective of human factors what is suggested debounce time? Thanks, Doug