stb and ac analysis does not match!?

Hi Guys,

I have some problems to understand what is wrong in the stb analysis or why in my mind it do not match with the Ac analysis and feedback theory.

To do stability analysis I always have done AC analysis and use the bode theory to understand the stability of op-amps or feedback systems.

When I designed the the first full-differential op-amp I have heard about the cmfb and stb analysis. I did it in this time and I agreed with the tool (Spectre -Cadence).

But now I want to understand and i did a simple OTA with a unitary feedback. To compare the to methods, I ran ac analysis with vdc source at input, I put ac = 1 and measured the open loop gain. it give-me 28 dB. (The input of the OTA was well biased and the transistors are in saturation region).
After I closed the loop and put a vdc source between the loop ( ac = 1, dc = 0), I run a Ac analysis and measured the close loop gain, as the OTA is a buffer configuration the close loop gain is equal to 0 dB.
With the same configuration, I ran the stb analysis than measured the loop gain that was 82 dB. I think it is wrong because the loop gain is the product between open loop and the close loop gain. So it would be 28 dB!!

Anyone can help me to understand this!

I read this paper that look good to understand the advantages between Ac and stb analysis.
http://www.cdnusers.org/community/virtuoso/resources/ams/tp_AllegroMicro_CDN_Final.pdf

Regards,

This is kind of weird... I've used both methods and they always match. Could you post the schematics for both configurations??

diemilio

Hi,

The schematics are attached,

Rafael.

Ok, testbenches look fine...

For what freq range are you running both the AC analysis and the STB??? Could you plot both the open loop AC response and the Loop Gain (20dB Mag and Phase) that Stb is giving you??

Keep in mind that when you're doing the open loop sim you're not loading your opamp with anything, whereas for the STB analysis you have an output load equivalent to the input capacitance of your circuit so that changes the location of the poles and zeros. The DC gain shouldn't be affected though, (unless you have a pole or a zero @ freq = 0, which is not the case here) that's why I find it a little bit surprising

diemilio

Hi,

I am sending the results.
As you said, the load will not change the DC gain, to make sure I put 1pF load that is bigger compare to the input capacitance.

When I looked the stb analysis, the Gain and the band-with is different from ac analysis.

Stb and ac analysis was ran from 1 to 1G, 20 point per decade.

Can you give one example!?

Rafael.

A gain of 25 dB seem rather low for this opamp... Are you 100% sure all your transistors are in saturation and that the biasing conditions are at the right place when you do the AC analysis?? Remember, not only you need them in saturation, but you also need the output voltage to be equal to the voltage of the reference you're connecting to the negative input terminal so that you can make an apples to apples comparison with the stb results.

It seems to me that the biasing conditions of your amplifier are changing when you add the feedback, which means the amp might not be properly biased.

diemilio

Added after 4 minutes:

BTW... are you setting the Vdc of the Voltage source for the stb analysis to 0??? Also, remove the value of Vac, just in case this might screw things up (but I don't think this is the case)

diemilio

Added after 6 minutes:

Here's an example...

diemilio

Hi Diemilio,

Thank for your help. I know this kind of opamp give one gain more then 28dB, more then 60 dB! depending on the biasing and size of the transistors.

I did not pay attention that was the different on the biasing of second stage of opamp. The difference on the biasing between the open and the close loop crashed the gain and changed the results.

I tried to put in the same condition to have the same results and now it matched.

Tanks again,

Rafael.

I saw the example now, thanks again.