基于自适应预测滤波理论的模数转换器设计方法

A Method to Improve the Dynamic Range of Analog-to-Digital Converter

基于自适应差分量化理论,提出了在不增加转换位数的情况下提高模数转换器(ADC)动态范围的一种设计方法。该方法通过量化输入与预测的差值来获得预测增益,从而扩大ADC的动态范围;通过差分量化值与预测值相加得到输出信号,并将该分别采用LMS和RLS算法进行处理,得到输入信号的预测值,计算机模拟结果证实,上述两种算法均能使模数转换器的动态范围提高约25dB.比较而言,RLS算法收敛速度更快,收敛性更好,但计算量较大.

This paper introduces a novel method, which can provide impetus for ADC performance improvements. By digitizing a residual signal, instead of the original received wideband signal, we show that the Signal-to-Noise Ratio (SNR) to the output signal of ADC can be significantly increased. Such residual signal is accomplished by predicting the received signal and subsequently subtracted from the overall received signal. Predicting signal is obtained from two prediction algorithms: Least-mean-square (LMS) and Recursive-least-square (RLS). The various computer simulated performance evaluation results justify the effectiveness of the proposed methodology.