Sigma delta modulator simulink model troubleshooting

[eladla]

Hi all,
I am currently working on a sigma delta modulator design. I have done all the math and extracted coefficients. I am now implementing my design in SIMULINK. The design is a 4th order CIFF. I am having one issue. When I ran the simulation the first time I got a negative SNR. When I removed the DAC and made it so the quantizer out put was Vref or -Vref (instead of 1 and 0), I got the SNR I expected from the theoretical calculations. I see a DAC in most of the references I looked at. Any ideas why this is? Am I doning something wrong?

Thanks you ppl!

 

[ElEngineer]

Maybe you have a gain element of 1/Vref after the DAC?

 

[eladla]

Thank you for your reply.
No, I have a gain element calculated by SDToolbox.
I`m now wondering...
Maybe the output of a fully differential feedforward modulator should be 1 or -1. How else will we be able to encode both possitive and negative inputs?

 

[ElEngineer]

I think that you are right. If you have negative signals you have to have the feedback with negative signals too.

 

[eladla]

The feedback DAC converts the output to Vref or -Vref, but my problem is that I have to set the output of the quantizer to 1 and -1 (in stead of 0 and 1) or else I don`t get the calculated SNR.
Any ideas?

Thanks.

 

[VitaMT]

SNR will be positive if you will not consider DC point.
In my design I use calcSNR from SDToolbox:
[snr,ptot]=calcSNR(yy1(1:N),f,3,fB,w,N);
instead of [snr,ptot]=calcSNR(yy1(1:N),f,1,fB,w,N);
And in all demos from SDToolbox output of the modulator is bipolar.