I have modelled a 3:rd order 3-bit delta sigma modulator in the matlab-toolbox
and transformed it to continuous time. The continuous time modulator is modeled
in Cadence using verilog-A. I have checked the impulse responses of both loopfilters in cadence and in matlab and they match so the linear systems should
be equivalent. However when I close the loops and simulate the modulators in
matlab and Cadence and plot the fft and calculate the SNR with matlab, there is
a difference. The SNR differs by 1 dB and the power spectral densities are slightly off. Does anybody have an idea about the origin of this difference? I have checked all parameters in both models many times and they should be correct so I find this
really strange.