Stability analysis and optimization of CT DSMs with an NS SAR quantizer by a redistributed noise shaping technique

Abstract

Continuous-time (CT) Delta-Sigma modulators (DSMs) with a passive noise-shaping (NS) successive-approxi-mation-register (SAR) analog-to-digital converter (ADC) as their internal quantizers are a promising ADC architecture for small area, high resolution and energy efficiency. A standalone passive NS SAR ADC can attain a stronger NS effect by placing left half-circle (LHC) poles in its noise transfer function. However, a CT DSM with an LHC-pole NS SAR quantizer has not been reported before, as it is difficult to stabilize. To address this issue, this article proposes a Redistributed Noise Shaping (ReNS) technique, along with an analysis method, to enable the use of an LHC-pole NS quantizer in a CT DSM for a stronger overall NS effect while maintaining a large input maximum stable amplitude. A CT DSM prototype designed in a 65 nm CMOS process demonstrates the effectiveness of the ReNS technique. With 20x oversampling, the quantization noise within the signal bandwidth (BW) and beyond 10x BW are redistributed into the frequency range between 1x and 10x BW. In simulations, the design achieves a signal-to-noise-and-distortion ratio of 73.4dB over 5MHz BW while consuming 1.19mW.