Hi,

I modeled a first-order Sigma-delta Modulator with 4-bit quantizers using verilogA in Cadence, I have used the delsig toolbox to get the Loop filter parameters (implemented in CIFB topology, OSR = 100, BW =10MHz, No. of Quantization levels = 16). The peak SNR predicted by the toolbox for the first-order modulator is 77.56 dB (for an input of 0.89FullScale) but for 0.85FS I got an SNR of 72.65dB.  

Now I am trying to model a MASH 1-1-1 ADC by replicating the above first-order modulator. But for this I get an SNR of 81.3 dB (for 0.85FS), while the theoretical SNR for a 3rd-order sigma-delta modulator is 143.05dB (for 0.85FS). The SNR that I get is similar to the one I get to the MASH 1-1 configuration implemented with the same first-order modulator. That is, I am not able to get the effect of the third modulator. I can clearly see noise-shaping if I plot the spectrum at the output of stage1 modulator but the noise shaping is not so prominent at the output of the stage2 and stage3 modulators (picture attached).

Is it the expected behaviour at the ouput of these stages? Is there something I can do it improve noise shaping at the output of stage2 and stage3.

Thanks.