A delta-sigma-modulation multi-channel network with improved noise-shaping characteristics

A delta-sigma-modulation multi-channel network is proposed for an analog-to-digital converter application. Weighting coefficients at summing nodes and digital filters obtaining the final output are designed so that adding the second channel cancels the first channel's quantization noise. The network suppresses the quantization noise leakage due to mismatch while improving the shaping characteristics. Increasing the weighting coefficients and the number of channels enhances the signal-to- noise ratio (SNR). Our signal-level simulation showed that for a 4-channel network, the SNR increased by around 25 dB, or an effective number of bits (ENOB) of 4 bits, compared with a 1-channel reference network for an oversampling ratio (OSR) of 60. Also, a preliminary circuit implementation demonstrated improved noise-shaping characteristics.