The Designer's Guide Community
Forum
Welcome, Guest. Please Login or Register. Please follow the Forum guidelines.
Feb 27th, 2020, 10:47pm
Pages: 1
Send Topic Print
STB vs AC analysis - different result (Read 305 times)
Andrew94
New Member
*
Offline



Posts: 2

STB vs AC analysis - different result
Dec 14th, 2019, 5:02am
 
Hello.
First I'm new to this forum. Great pleasure to meet all of you Smiley

My question is:
I have an opamp (one stage with differential input and current mirror as a active load - the basic configuration). I wanted to measure the DC Open-Loop Gain, Closed-Loop Gain, PM & GM.

First I have made STB sim(with no C as a load at output): I put the opamp in the voltage follower configuration and in the loop i put a DC voltage source with 0 value as a probe. I have made the sim and the result for PM, for example, was PM=41.2deg.

Second I have made AC sim(with no C as a load at output): I put the opamp in the voltage follower configuration and in the loop i put an inductor L=1G H and a capacitor to ground C=1G F. I have made the sim and the result for PM, for example, was PM=-85.8deg.

Strange. So I applied a step at input and in the voltage follower configuration(same, with no C as a load at output) I looked at output to see the wave form, and the wave form looks like for a PM of around 40-45.

https://i.ibb.co/RpVf3Bj/STB-vs-AC-loop.png


I put the image below.
*in the image there are 3 simulation for both STB in the right and AC in the left, with 3 different values of output C as load.
All of them are different, hence offering different results.
* also in the image it can be seen the step response, looking almost like a PM around 45.

What is the explanation ?
STB and AC should show the same result or not ?
Back to top
 

STB_vs_AC_loop.png
View Profile   IP Logged
Ken Kundert
Global Moderator
*****
Offline



Posts: 2247
Silicon Valley
Re: STB vs AC analysis - different result
Reply #1 - Dec 14th, 2019, 2:59pm
 
It is virtually impossible to accurately measure loop gain with a simple AC analysis. You should never use AC analysis to measure loop gain.  One of the reasons why you don't get the right answer is that you have changed your circuit. Specifically, by adding the filter you are changing the loading at the input as seen by the feedback network.

-Ken
Back to top
 
 
View Profile WWW   IP Logged
Andrew94
New Member
*
Offline



Posts: 2

Re: STB vs AC analysis - different result
Reply #2 - Dec 17th, 2019, 4:17am
 
Ok, I understand your point of view.

But:
1. Which one says the correct result for PM,GM,Open-Loop or Close-Loop Gian, 3db BW, unity gain freq ? AC or STB ?
 
1.1. Because in a corner simulation, or monte-carlo, both will show different values.

1.2. You can not make the calculations for all the designs you are doing, just to check which one is giving you something near the right answer.
      And in this way we return to the question 1 and 1.1.



2. Why L-C method is showing different answer ? Although all the books are using it.

2.2. AC method (with L-C), basic tell you the "Gain and Phase" of the signal, that means in other words, the gain form the output all the way to the input !!!!
     On the other hand, STB give you the "Loop Gain", that means the gain inside the loop or the gain of the loop. So just from theoretical point of view, STB is not giving you any information about the gain, !!! attention !!!, from the output all the way to the input.
Back to top
 
 
View Profile   IP Logged
Ken Kundert
Global Moderator
*****
Offline



Posts: 2247
Silicon Valley
Re: STB vs AC analysis - different result
Reply #3 - Dec 17th, 2019, 2:17pm
 
For loop gain, gain margin, and phase margin, use STB analysis.

For closed-loop gain, use AC.

For open-loop gain, don't bother. It is meaningless. Nobody uses an opamp open loop.

Why does L-C method show a different answer from stb? It is wrong.  Why is it wrong? Because you changed the circuit.

Why do all the books recommend the L-C method? Because that is the way it was done before we had the stability analysis. They were never accurate, but it was the best we could do.

It is important to realize that you are not trying to measure the characteristics of the opamp. Rather you are trying to measure the characteristics of your circuit.  And if you change your circuit you change your results. So you always make measurements with the loop closed and unmodified.  If you are interested in gain, input or output impedance, PSRR, etc. you measure it with an AC or XF analysis. If you are interested in noise, you use a noise analysis. And if you are interested in stability, you use a stability analysis.

-Ken

Back to top
 
 
View Profile WWW   IP Logged
Frank Wiedmann
Community Fellow
*****
Offline



Posts: 658
Munich, Germany
Re: STB vs AC analysis - different result
Reply #4 - Dec 18th, 2019, 1:09am
 
You can find a lot of information on loop gain simulation (probably more than you ever wanted to know) on my webpage at https://sites.google.com/site/frankwiedmann/loopgain
Back to top
 
 
View Profile WWW   IP Logged
Mos
Junior Member
**
Offline



Posts: 13

Re: STB vs AC analysis - different result
Reply #5 - Feb 8th, 2020, 11:55am
 
Hi I have a question on stb analysis. Is there a recommended location on inserting the iprobe for stb analysis.
I am trying to simulate a voltage regulator. If I insert at the gate of error amp(Diff pair gate) I get one response and if I insert at the gate of pass transistor I get different response.
Which one to trust and is there any theory behind where to insert the iprobe?
Thanks
Back to top
 
 
View Profile   IP Logged
Frank Wiedmann
Community Fellow
*****
Offline



Posts: 658
Munich, Germany
Re: STB vs AC analysis - different result
Reply #6 - Feb 10th, 2020, 2:48am
 
The location of the iprobe for the stb analysis does not matter if there is only one loop in the circuit and all signals are referenced to the global ground. If there are several loops in the circuit, you must insert the iprobe into a wire that is common to all loops (see https://kenkundert.com/docs/cd2001-01.pdf for details). If the signals connected to the iprobe are referenced to a node that is different from the global ground, you must enter the node name in the field "Local Ground Name" of the setup form for the stb analysis.
Back to top
 
« Last Edit: Feb 10th, 2020, 4:13am by Frank Wiedmann »  
View Profile WWW   IP Logged
Mos
Junior Member
**
Offline



Posts: 13

Re: STB vs AC analysis - different result
Reply #7 - Feb 11th, 2020, 7:23am
 
Thanks Frank for the document.
Does stb analysis changes between loop-based and device-based algorithm.
Or does it use one of them all the time?

Thanks
Back to top
 
 
View Profile   IP Logged
Frank Wiedmann
Community Fellow
*****
Offline



Posts: 658
Munich, Germany
Re: STB vs AC analysis - different result
Reply #8 - Feb 12th, 2020, 8:37am
 
The stb analysis can use either the loop-based or the device-based algorithm, depending on what you specify as the probe instance (see https://support.cadence.com/apex/techpubDocViewerPage?path=spectreuser/spectreus...). I (and probably most other people as well) only use the loop-based algorithm, however.
Back to top
 
 
View Profile WWW   IP Logged
Pages: 1
Send Topic Print
Copyright 2002-2020 Designer’s Guide Consulting, Inc. Designer’s Guide® is a registered trademark of Designer’s Guide Consulting, Inc. All rights reserved. Send comments or questions to editor@designers-guide.org. Consider submitting a paper or model.