刚开始接触ADC，3阶，过采样64,1位量化的模型，效果很差，请各位给些改进建议。

order=3，M=64

SNR

Architecture is OK, but the spectrum seems very bad.
I think you should change the integrator coefficients.

I have calculated several times about the coefficients compared with the outcome H(z) from Matlab. Is there any possibility that the simulation .m file does not fit with my architecture for I only change the parameter such as the order, osr. I do not well understand the file which I borrowed from others.

Maybe, It's possible

thanks

三阶及以上1bit 量化很容易产生量化器过窄，请尝试减小输入信号幅度或改为多bit 结构。

thanks i will try again