Designer’s Guide Community

BackerShu

Community Member

Offline

Posts: 64

How to simulate the fully differential opamp with SC CMFB?
Aug 30^th, 2009, 9:25pm

After adding the SC CMFB circuit to a fully differential opamp, I can only check whether the output common mode voltage could stay at the expected value with a certain time of tran. simulation to determine how the SC CMFB circuit works. Could someone tell me how to exactly find out its infulence on the opamp's AC performance GBW, PM, etc. by simulation? In another word, how to simulate the GBW and PM of a opamp with SC CMFB?
I have tried pss and pac simulation in SpectreRF. But the simulation result donesn't resonable. So I doubt this method may not work or I did it wrong way.

Methods using any tools would be helpful!

Thx in advance!!

Back to top

Thanks and Regards!
BackerShu

https://www.linkedin.com/pub/shu-guanghua/26/b72/711

IP Logged

tyanata

New Member

Offline

Posts: 2

Re: How to simulate the fully differential opamp with SC CMFB?
Reply #1 - Sep 3^rd, 2009, 5:50am

I have same problem. I have designed amplifier which in AC simulation with ideal CMFB (using system from several VCVS) achieves about 75dB gain. Using PSS and PAC with real SC CMFB I achieve about 33dB. It is strange that in spectre.out file for DC levels for OUTP and OUTN are about 300mV output N cascodes are in triode region, but if I plot OUTP and OUTN from PSS analysis it is seen that they achieve required from design CM. So it seems that there is some option to be mark PAC to takes steady state DC OP after stabilization time, but not DC OP after beginning of simulation. My clock is 2MHz. I entered for tstab 500us which is quite enough and tstart is 200us.
My be your problem is same!

Back to top

IP Logged

Ken Kundert

Global Moderator

Offline

Posts: 2384
Silicon Valley

Re: How to simulate the fully differential opamp with SC CMFB?
Reply #2 - Sep 3^rd, 2009, 9:18am

BackerShu,
PSS followed by PAC, PXF, and PSTB are definitely the way to go. If the result are wrong it, it is probably because you are making a mistake in how you apply them. Perhaps if you describe what you are doing, we can point out the flaw.

Tyanata,
In DC analysis the clocks are not operating, so the common-mode bias point is not being controlled by the common-mode feedback. As such, you DC results seem to be reasonable, but also useless. You need to focus on determining what is limiting the gain in the real circuit. Have you tried running PAC analysis on the idealized circuit? If you did and still saw 75dB of gain, it would point to the problem being in the circuit rather than the simulation.

-Ken

Back to top

IP Logged

BackerShu

Community Member

Offline

Posts: 64

Re: How to simulate the fully differential opamp with SC CMFB?
Reply #3 - Sep 9^th, 2009, 7:26am

tyanata:
It seems that we do encounter the same problem. The operating frequency of the SC_CMFB in my opamp is 50MHz. I tried setting the tstab at 10us or even larger, but the result remain unresonable, the voltage gain changes from 125dB in the AC simulation to -45dB in PAC simulation.
BTW, where to set the tstart, I just couldn't find the parameter? what does the parameter mean?

Ken:
I tried the opamp with ideal CMFB or the SC_CMFB in the way that tyanata tells, but the result makes no differnent and both are wrong. I think I do make some mistakes in the simulation settings. I'll send both you and tyanata a file to specifically describe how I set the simulation condition and hope this will give a clue to solve the problem that bother me so much.

Thank you very much for your attention and helpful replies!!

Back to top

Thanks and Regards!
BackerShu

https://www.linkedin.com/pub/shu-guanghua/26/b72/711

IP Logged

BackerShu

Community Member

Offline

Posts: 64

Re: How to simulate the fully differential opamp with SC CMFB?
Reply #4 - Sep 11^th, 2009, 6:57pm

After some further simulation, I did get the some simulation results which seems reasonable. But still with DC Gain and GBW reduction, I extend the tstab to 100us. The results makes no improvement.
I write a file about how I did the simulation and gave the results along with my question. Hope this helps to solve the problem.

Regards!

Back to top

SC_CMFB_Simulation_Pro.rar

Thanks and Regards!
BackerShu

https://www.linkedin.com/pub/shu-guanghua/26/b72/711

IP Logged

BackerShu Community Member Offline Posts: 64	Re: How to simulate the fully differential opamp with SC CMFB? Reply #5 - Sep 11^th, 2009, 10:27pm the SC_CMFB circuit in the design is shown in the attchment. Does it has something with the DC Gain and the GBW reduction?
Back to top	SC_CMFB.jpg Thanks and Regards! BackerShu https://www.linkedin.com/pub/shu-guanghua/26/b72/711 IP Logged