Dual cycle shift data-weighted averaging technique for multi-bit Sigma-Delta Modulators

Abstract

Data weighted averaging (DWA) is the most popular dynamic element matching (DEM) technique to reduce the effects of DAC nonlinearity in multi-bit sigma-delta modulators (MBSDM). However, due to the non-symmetry of the unit-elements in multi-bit DAC, in-band signal-dependent tones always appear when a low oversampling ratio (OSR) is employed in traditional DWA. Therefore, a novel DEM technique, dual cycle shift DWA (DCSDWA), is proposed to deal with the tone problem in this paper. Compared with other modified DWA techniques, DCSDWA is a more effective one with simple implementation. Using the proposed technique in a 4th-order MBSDM with an OSR of 32, the maximum in-band tone is reduced to 6 dB when a random DAC-element mismatching error of 0.5% is assumed. In addition, it results in a better linearity of signal-to-noise distortion ratio (SNDR) versus different input signal and a wider dynamic range (DR). The simulation results prove that the proposed DCSDWA is suitable for low OSR applications (OSR¿8) with a good performance.