← 返回 JSSC 论文列表JSSC 2023第9期Data Converters55nm CMOSNeural Network Accelerator
A 1 V 107 µW 15-Bit Pseudo-Pseudo-Differential Incremental Zoom ADC Zhaonan Lu H
一种1V供电、107µW功耗的15位伪伪差分增量缩放ADC
89.9dB SNR/0.378ms转换时间/1.07µW@1V
模数转换器增量型ADC伪差分结构低功耗设计CMRR增强
▸创新点1:采用伪伪差分(PPD)结构,通过两个单端3位SAR ADC和一个单端三阶增量型ADC处理差分输入信号,有效提高共模抑制比(CMRR),同时降低功耗和面积开销。
▸创新点2:引入三相位时钟技术,消除正负输入信号采样之间的半周期延迟,显著提升ADC的CMRR性能,优化了差分信号处理的同步性。
▸创新点3:混合SAR与增量型ADC架构,结合SAR ADC的高速度和增量型ADC的高精度,在1V电源下实现107µW超低功耗和89.9dB SNR的高性能指标。
▸创新点4:采用55nm CMOS工艺实现,在0.378ms转换时间内达到180.8dB Schreier FoM,展示了该设计在低功耗高精度ADC领域的卓越性能。
Abstract
This article presents a 15-bit pseudo-pseudo-
differential (PPD) incremental zoom analog-to-digital converter
(ADC). It employs two single-ended (SE) 3-bit successive-
approximation-register (SAR) ADC and a third-order SE incre-
mental 16 ADC to process a pair of differential input signals.
A novel three-phase clock helps eliminate the half-cycle delay
between the positive and the negative input sampling, boosting
this work’s common-mode-rejection ratio (CMRR). Fabricated in
55 nm CMOS technolog