Questions about sigma-delta harmonics in modulator

[intuition]

sigma-delta harmonics

Hi, I'm a designer of 2rd order sigma-delta modulator. I have some questions while implementing fully-differential op with discrete cmfb as below:

1. How the op harmonics effect the modulator output? What harmonic level could be acceptable for 100dB SNR modulator?

2. What's the standard way to estimate harmonics for fully-differential op with discrete cmfb? Because of cmfb using clock, the output signal are neither so pretty for spice fft analysis nor Matlab analysis.

Can anybody give some advices for either question? thanks a lot!

 

[svensl]

Re: sigma-delta harmonics

Try https://www.designers-guide.org/Analysis/delta-sigma.pdf.
It might help.

 

[boeysue]

Re: sigma-delta harmonics

The dspguide seems not found that topics.
And should you descripe your design?

Is that ADC or DAC?

 

[intuition]

Re: sigma-delta harmonics

The dspguide seems not found that topics.
And should you descripe your design?

Is that ADC or DAC?

Hi, it's a 3bits sigma-delta ADC modulator design. I've already constructed a 2nd order cascaded with switch-capacitors and 2 fully differential op, but the op's maximum output swing could reach 70% vdd, for bigger the swing, larger the harmonics. So I want to estimate the effect of harmonics at op's output to this modulator.

But for using discrete common mode feedback, I can't tell if the harmonics is contributed by op's linearity, or by common-mode feedback clock..Should I add these two together.. or just consider op's linearity?.. or .. the harmonics are not so critical in sigma-delta modulator design..??

 

[extraord]

Re: sigma-delta harmonics

harmonic is not a key feature in sigma-delta modulator,because it will be shrinked by negtive feedback of integrator. A/1+AF.

and you can also find out the harmonic Distortion of op amplifier by Spectre pac analysis.(when there is a discrete time cmfb).

 

[intuition]

Re: sigma-delta harmonics

Thanks. I've done some simulation at Matlab, just put a harmonic signal at integrator output. If I put -60dbv harmonic at first integrator output, it will depressed to -100dbv at sdm output, a 40dbv depression for 2nd order case. So I think even if harmonics occurs, it would be shrinked by noise shaping. And for the second integrator, the harmonics just doesn't matter at all.. Regards