A systematic approach for testing a DAC mismatch shaping method

DAC mismatch shaping for multibit Delta-Sigma A/D-converters has been a very popular topic in scientific publications. In recent years, most of these publications have proposed various methods on how to modify a popular data weighted averaging scheme so that it would produce less spurious tones. Many of these publications reveal different styles on reporting the performance merits. A simulation result based on only one set of random differential nonlinearity errors for the unit DACs is hardly valid. A group of simulations based on several randomizations is more informative, but hardly repeatable. This paper proposes some ground rules, which improve comparability of various methods.