← 返回 JSSC 论文列表JSSC 2021第10期Data Converters40nmNeural Network Accelerator
A CT 22 MASH ADC With Multi-Rate LMS-Based Background Calibration and Input-Ins
一种采用多速率LMS背景校准的CT 2-2 MASH ADC，具有高温度和工作频率变化容忍度。
40nm CMOS, 8MHz带宽, 79dB DR, 78.5dB SNDR, 7.3mW模拟功耗, 16mW数字功耗
连续时间ADC多级噪声整形背景校准LMS算法温度稳定性
▸基于LMS算法的多速率背景校准技术
▸结合后置调节器和伪随机噪声注入的数字校准方法
▸优化的第一调制器设计以减少输入信号泄漏
Abstract
This article presents a continuous-time (CT) multi-
stage noise-shaping (MASH) delta–sigma ( ) analog-to-digital
converter (ADC) with enhanced tolerance to temperature and
operating-frequency variations through on-chip multi-rate (MR)
background calibration based on the least-mean-square (LMS)
algorithm. The proposed digital calibration, which combines MR
operation, post-conditioners, and pseudorandom-noise injection
at the quantizer input, efficiently and accurately matches dig-
ital transfer