Abstract-Dynamic element matching (DEM) is capable of providing good average linearity performance in matching critical circuits in the presence of major component mismatch, but the approach has received minimal industrial adoption outside of Σ−∆ structures because of challenges associated with implementation of a required randomizer and because of the time-local nonstationarity. This paper presents a DEM approach to analog-to-digital converter (ADC) testing in which low-precision DEM digitalto-analog converters (DACs) are used to generate stimulus signals for ADCs under test. It is shown that in a testing environment, this approach provides very high precision test results, and timelocal nonstationarity is of no concern. In addition to traditional random DEM techniques, a deterministic DEM (DDEM) strategy that eliminates the need for a randomizer is introduced. The performance of the DDEM method is established mathematically and validated with detailed simulation results. Furthermore, the DDEM method requires far fewer samples to achieve the same level of average linearity than the random DEM approach. It is demonstrated that both the random DEM and DDEM methods can be used to accurately test ADCs with linearity that far exceeds that of the DAC used as a signal generator. This technique of using imprecise excitations and DEM to test much more accurate ADCs offers potential for use in both production test and built-in self-test environments where high linearity test sources are difficult to implement.Index Terms-Analog-to-digital converter (ADC) testing, built-in self-test (BIST), dynamic element matching (DEM), integral nonlinearity (INL) testing.