Need Help: Design of 14 bits SC Sigma Delta ADC

[joujou]

Hi,

I designed a SC Sigma Delta ADC with its different compenents: TG, feedback TG,OTA, Latch Comparator .
But when I analyzed its spectrum to have 14 bits , I obtained these results:



The first one is a superposition of all spectrum:
**The ENOB of the modulator with a real OTA (13.31 bits) is better than the ENOB of the modulator with an ideal OTA (13.04 bits) : illogical result !!
** All spectrum contain a third order harmonic!!
The second one is a zoom on the bandwith
Someone can help me to analysis these result?
Please Let me Know if you have some comments

Joujou

 

[tony_taoyh]

I think, you need to check the setting of the ideal OTA: the main thing is gain, gm, and slew rate.

The third hamonica is due to the non-linearity. Normally, it is caused by the gain non-linearity and switch's on-resistance non-linearity.

Hope this information helps.

 

[joujou]

Hi tony_taoyh,

Thank you for your help and sorry for my late replay
I think that the probléme is on the transmission Gate because when I test the spectrum of the ideal modulator i used an ideal amplifior

My request now is:
What are the simulations that you should do to confirm the origin of the third order harmonic?

joujou

 

[tony_taoyh]

The simulation skill apply to any circuit:
You can use transient simulation to catch some non-linear problem: for example, the output swing of the OTA, the non-linearity of the ON-resistance. However, be careful on the result: Because, device mismatching is not included for the transient simulation. Some side calculation is required for this one.

Another one is: the noise effect is not included in the transient simulation. Notes: in the latest cadence tool, the vendor try to include it, using some random number generator.