The Design and Implementation of Low-Pass Multibit Delta-Sigma Modulators

碩士 === 國立臺灣大學 === 電機工程學研究所 === 89 === Abstract Oversampling techniques based on delta-sigma modulation are widely used to implement the interface between analog and digital signals in VLSI systems, such as digital audio systems. This type of systems requires a large dynamic range (i.e., a...

Full description

Bibliographic Details
Main Authors: Tzu-Chien Hsueh, 薛子建
Other Authors: Shen-Iuan Liu
Format: Others
Language:en_US
Published: 2001
Online Access:http://ndltd.ncl.edu.tw/handle/54557132412328256876
Description
Summary:碩士 === 國立臺灣大學 === 電機工程學研究所 === 89 === Abstract Oversampling techniques based on delta-sigma modulation are widely used to implement the interface between analog and digital signals in VLSI systems, such as digital audio systems. This type of systems requires a large dynamic range (i.e., about 16 ~ 20 bits) at low-frequency bandwidth of 22 kHz. The delta-sigma modulation approach is relatively insensitive to imperfections in circuit components and offers numerous advantages for the realization of high-resolution analog-to-digital (A/D) converters. In particular, oversampling architectures is a potentially power-efficient means of implementing high-resolution A/D converters because they reduce the number and complexity of the analog circuits in comparison Nyquist-rate converters. Furthermore, they allow the performance requirements, and thus most of the power dissipation, to be concentrated in the input stage of a converter. For a multibit DS modulator, the performance is directly related to the linearity of the internal multibit DAC in the feedback path. Various dynamic element-matching techniques have been proposed to circumvent the nonlinearity of the internal DAC. By using these techniques, the DAC noise is also shaped like quantization noise in delta-sigma modulators. In this thesis, we pay attention to the data weighted averaging algorithm, which ideally can achieve first-order DAC noise shaping. The implementation of the DWA algorithm is much less complexity than others, and this algorithm can provide very high-resolution delta-sigma analog-to-digital converters for DC measurements. However, the DWA algorithm induces tones into the baseband resulting in degradation of the modulator’s performance. In this work, we propose an improvement to suppress the performance degradation caused by tones. This thesis describes the results of research into the design of the oversampling multi-level delta-sigma modulators implemented by switch-capacitor circuits using the modified data weighted averaging algorithm that provide the performance required for high fidelity, digital audio applications. The experimental modulators described herein can deliver a high dynamic range over a 22 kHz bandwidth and have been fabricated in standard 0.35 mm CMOS technologies.