← 返回 JSSC 论文列表
📄 下载 JSSC 原文 PDF
JSSC 2023第12期Data ConvertersSAR ADCNeural Network Accelerator

A Rail-to-Rail 12 MSs 913 dB SNDR 941 dB DR Two-Step SAR ADC With Integrated Inp

提出一种带集成输入缓冲器的两步式SAR ADC,采用预测电平移位技术实现轨到轨信号摆幅,性能优异。
12 MS/s, 91.3 dB SNDR, 94.1 dB DR, 6.6 Vpp差分摆幅, 4 MHz输入频率, 30.41 mW功耗
SAR ADC输入缓冲器轨到轨预测电平移位高线性度
创新点1:预测电平移位输入驱动技术(方法创新) - 该技术通过预测输入信号的电平变化并动态调整驱动电平,实现了轨到轨信号摆幅的驱动,无需增加缓冲器供电电压,显著提高了线性度和驱动能力,适用于高采样率场景。
创新点2:集成输入缓冲器设计(电路创新) - 通过将输入缓冲器集成到ADC芯片中,有效降低了输入负载,减少了外部驱动电路的复杂度,同时优化了热噪声性能,支持6.6 V峰峰值差分信号摆幅。
创新点3:两步式SAR ADC架构(系统创新) - 采用两步式逐次逼近寄存器(SAR)架构,结合高精度和高速特性,在12 MS/s采样率下实现了91.3 dB SNDR和94.1 dB动态范围,功耗仅为30.41 mW,性能领先同类设计。
创新点4:低功耗高性能设计(系统创新) - 通过优化电路设计和系统架构,在保证91.3 dB SNDR和94.1 dB动态范围的同时,将功耗控制在30.41 mW,实现了高能效比,适用于便携式和低功耗应用场景。
Abstract
Input buffers can be used to reduce the input load of high-resolution discrete-time (DT) Nyquist analog-to- digital converters (ADCs), which can be challenging to drive, particularly at high sampling rates, because of the large input sampling capacitance needed to reduce thermal noise. An input driving technique called predictive level-shifting is proposed to drive rail-to-rail signal swing without increasing the buffer supply rail for high linearity. This article demonstrates an easy-to- drive