高精度SAR型ADC结构及基于LMS的自适应精度提升算法

Traditional SAR ADC structure and weight error calibration method are not suitable for high resolution SAR ADC, this have already become a key limiting factor which block the resolution increasing of the SAR ADC. In order to solve this problem, the following innovative research will be done:1、The new universal structure for the high resolution SAR ADC will be presented:①redundant structure will be used to realize digital calibration for SAR ADC,②digital calibration and segmentation are combined to realize precisionimprovement. The problem that thetraditional SAR ADC structure can only use analog calibrationalgorithm which leads to the increasing of circuit complexity and low calibration precisionwill be solved by the presented new structure.2、LMS based self-adaptive weight error universal extraction and calibration algorithm will be presented:① LMS based self-adaptive weight error universal extraction model will be created, ②weight error universal calibration model will be created.The problem that thetraditional analog weight error extraction and calibrationalgorithm need assistant circuit can be solved;The problem that the digital calibration algorithm presented in recent years of complicating design are difficult to be applied to high resolution SAR ADC will also be solved by the presented new LMS based self-adaptivealgorithm. 3、High resolution SAR ADC using the new presented structure and calibration algorithm will be designed.The presented weight error extraction algorithm can be popularly applied to all kind of ADC for weight error extraction. And the presented weight error digital calibration algorithm can be applied to all redundant structure ADCs.

传统的SAR型ADC结构和权重误差校准方法对高精度SAR ADC不适用，已经成为制约SAR型ADC精度提升的一个关键限制因素。为此，进行以下创新研究：1、提出新的高精度SAR型ADC通用结构：①利用冗余结构实现SAR型ADC的数字校准，②利用数字校准技术和分段结构实现精度提升。解决传统SAR型ADC结构只能采用模拟校准方法导致电路复杂度高、校准精度低的问题。2、提出基于LMS的自适应权重误差通用提取及校准方法：①建立基于LMS的自适应权重误差通用提取模型，②建立权重误差通用校准模型。解决传统模拟的权重误差提取和校准方法需要辅助电路的问题；解决近年来数字校准方法设计复杂度大，很难应用于高精度SAR型ADC的问题。3、研制采用新结构新校准方法的高精度的SAR型ADC。提出的权重误差提取方法可以推广应用到所有类型ADC的权重提取中；提出的权重误差数字校准方法可以推广应用到所有的冗余结构ADC中。

A/D转换器是连接模拟世界和数字世界的桥梁，将自然界的连续的模拟信号转换为计算机等数字设备可以处理的离散的数字信号，A/D转换器的精度和线性度越高，转换后的信号跟真实的模拟信号就越接近。由于制造工艺的偏差，限制了A/D转换器的精度和线性的提升，这成为高精度A/D转换器实现的瓶颈。.本项目提出了一种通用的高精度A/D转换器结构，以及相应的精度及线性度提升算法。提出的基于分段冗余结构，打破了精度提升与版图面积的固有关系，能够极大的降低版图面积和功耗。提出的分段参数模型，极大的较少了校准需要的样本数量，校准速度快，精度高（例如：对于16-18位A/D，比传统直方图测试快500倍，SNR达到96dB以上，SFDR达到118dB），并且不会降低转换速率。