Author Topic: How can I obtain amplifier circuit loop gain?  (Read 3731 times)

0 Members and 1 Guest are viewing this topic.

Offline Lucky-LukaTopic starter

  • Supporter
  • ****
  • Posts: 240
  • Country: it
How can I obtain amplifier circuit loop gain?
« on: January 16, 2020, 08:06:24 pm »
In the last few days I've tried to understand how to get the loop gain (Aol*ß) of my amplifier circuit. I've received several suggestion. I would like to hear your opinion on this matter. Are these 2 methods equivalent? Are they correct?
The results look the same.
Cheers


Memento audere semper.
 

Offline Benta

  • Super Contributor
  • ***
  • Posts: 6420
  • Country: de
Re: How can I obtain amplifier circuit loop gain?
« Reply #1 on: January 16, 2020, 09:02:29 pm »
"Loop gain" makes little sense. You can measure open-loop gain (x - x open) or closed-loop gain (x - x connected). Please clarify.
 

Offline Lucky-LukaTopic starter

  • Supporter
  • ****
  • Posts: 240
  • Country: it
Re: How can I obtain amplifier circuit loop gain?
« Reply #2 on: January 16, 2020, 09:09:37 pm »
I have already said what it is...
Memento audere semper.
 

Offline Wimberleytech

  • Super Contributor
  • ***
  • Posts: 1134
  • Country: us
Re: How can I obtain amplifier circuit loop gain?
« Reply #3 on: January 16, 2020, 09:24:32 pm »
I typically use the inductor method.  As an alternative do the test in two steps.  First close the loop and get the node voltages at the input, then open the loop and drive those voltages into the inputs...yielding...hopefully...the same operating point. 

But to your question, I would say your two methods are mostly equivalent for ac analysis.
 
The following users thanked this post: Lucky-Luka

Offline magic

  • Super Contributor
  • ***
  • Posts: 7453
  • Country: pl
Re: How can I obtain amplifier circuit loop gain?
« Reply #4 on: January 16, 2020, 10:34:00 pm »
You have one of the sources backwards and phase is 180° off.

The gigafarad capacitor always seemed intuitively right to me. The latter is weird, and who would want to edit the plots to divide by V(x) every time one switches to AC analysis. I don't like it.

"Loop gain" makes little sense. You can measure open-loop gain (x - x open) or closed-loop gain (x - x connected). Please clarify.
There is even a wikipedia article on that ;)

It makes sense. For example, unity loop gain frequency rather than unity gain frequency is the one you would care about in an amplifier permanently wired for high closed loop gain and decompensated correspondingly. Expect to see this jargon in anything related to audio amplifiers.
 

Offline Wimberleytech

  • Super Contributor
  • ***
  • Posts: 1134
  • Country: us
Re: How can I obtain amplifier circuit loop gain?
« Reply #5 on: January 16, 2020, 11:26:55 pm »


It makes sense. For example, unity loop gain frequency rather than unity gain frequency is the one you would care about in an amplifier permanently wired for high closed loop gain and decompensated correspondingly. Expect to see this jargon in anything related to audio amplifiers.

Indeed...it does make sense.
When designing the compensation for an amplifier, you plot loop gain vs phase to find phase margin.  Most of the time the goal is unity-gain stability and for that case open-loop gain and loop gain are equivalent.
 

Offline Cerebus

  • Super Contributor
  • ***
  • Posts: 10576
  • Country: gb
Re: How can I obtain amplifier circuit loop gain?
« Reply #6 on: January 17, 2020, 12:04:46 am »
I see you're using LTSpice.

Got to wherever your copy of LTSpice has stored its 'examples' directory and open "LoopGain.asc" - it contains an example of using "Middlebrook's method" in LTSpice. There is also a "LoopGain2.asc" in the same directory showing Tian et al's method of doing essentially the same thing using a slightly different method. I personally found the latter easier to use.
Anybody got a syringe I can use to squeeze the magic smoke back into this?
 
The following users thanked this post: Lucky-Luka

Offline H713

  • Regular Contributor
  • *
  • Posts: 180
  • Country: us
Re: How can I obtain amplifier circuit loop gain?
« Reply #7 on: January 17, 2020, 07:31:56 am »
I know this is off-topic, but if you actually plan to build that you will need something considerably beefier than a 2N4401/2N4403 to drive those 2N3055s to any reasonable power level. The current gain of a 2N3055 at high currents is something like 5. Those little TO-92 package drivers aren't going to like that.
 

Offline Lucky-LukaTopic starter

  • Supporter
  • ****
  • Posts: 240
  • Country: it
Re: How can I obtain amplifier circuit loop gain?
« Reply #8 on: January 17, 2020, 10:42:04 am »
what's the purpose of the big inductor?
Memento audere semper.
 

Offline magic

  • Super Contributor
  • ***
  • Posts: 7453
  • Country: pl
Re: How can I obtain amplifier circuit loop gain?
« Reply #9 on: January 17, 2020, 11:32:48 am »
Try without ;)

The whole L/C circuit passes output DC level back to the input transistor to establish normal operating conditions, but blocks AC, at least above some nano-hertz frequencies ;)
 
The following users thanked this post: Lucky-Luka

Offline Lucky-LukaTopic starter

  • Supporter
  • ****
  • Posts: 240
  • Country: it
Re: How can I obtain amplifier circuit loop gain?
« Reply #10 on: January 17, 2020, 12:17:06 pm »
I don't understand why you need to block AC
Memento audere semper.
 

Offline Wimberleytech

  • Super Contributor
  • ***
  • Posts: 1134
  • Country: us
Re: How can I obtain amplifier circuit loop gain?
« Reply #11 on: January 17, 2020, 02:54:05 pm »
I don't understand why you need to block AC

Because you want to OPEN the loop for frequencies of interest (everything above the nano Hz range, e.g., as magic said).

 
The following users thanked this post: Lucky-Luka

Online mikerj

  • Super Contributor
  • ***
  • Posts: 3382
  • Country: gb
Re: How can I obtain amplifier circuit loop gain?
« Reply #12 on: January 17, 2020, 03:09:05 pm »
I don't understand why you need to block AC

The whole idea is to remove the feedback at AC to measure open-loop gain/phase, but retain DC feedback to set the correct operating conditions.  If you don't block AC then you are providing negative feedback and therefore won't be measuring open-loop gain. 
 
The following users thanked this post: Lucky-Luka

Offline Lucky-LukaTopic starter

  • Supporter
  • ****
  • Posts: 240
  • Country: it
Re: How can I obtain amplifier circuit loop gain?
« Reply #13 on: January 17, 2020, 04:56:06 pm »
1)Before stating that my request is fundamentally wrong I think you should read my question more carefully Lesolee. Haven't you read the anwers I received?
2)The video was to meant to point out what the loop gain is, since another user didn't know it.

Thanks everybody for the answers.

p.s. LOL. Lesolee's post misteriously disappeared after my reply.
« Last Edit: January 17, 2020, 05:29:38 pm by Lucky-Luka »
Memento audere semper.
 

Offline bson

  • Supporter
  • ****
  • Posts: 2497
  • Country: us
Re: How can I obtain amplifier circuit loop gain?
« Reply #14 on: January 19, 2020, 12:02:33 am »
Vf/Vin, where Vf is the feedback voltage.
 

Offline Cerebus

  • Super Contributor
  • ***
  • Posts: 10576
  • Country: gb
Re: How can I obtain amplifier circuit loop gain?
« Reply #15 on: January 19, 2020, 01:47:52 am »
Vf/Vin, where Vf is the feedback voltage.

What question do you think that is an answer to?
Anybody got a syringe I can use to squeeze the magic smoke back into this?
 

Offline Wimberleytech

  • Super Contributor
  • ***
  • Posts: 1134
  • Country: us
Re: How can I obtain amplifier circuit loop gain?
« Reply #16 on: January 19, 2020, 03:49:21 am »
Vf/Vin, where Vf is the feedback voltage.

What question do you think that is an answer to?

You mean.: "To what question do you think that is the answer?"

Anyhooo, it is actually the answer to the title of the original post.
 

Offline Lucky-LukaTopic starter

  • Supporter
  • ****
  • Posts: 240
  • Country: it
Re: How can I obtain amplifier circuit loop gain?
« Reply #17 on: January 19, 2020, 09:01:36 am »
Apart from the fact that I've already received and answer, how can I obtain Vf?
I don't think that such an answer adds something to the thread.
Maybe you can prove me wrong.
Memento audere semper.
 

Offline magic

  • Super Contributor
  • ***
  • Posts: 7453
  • Country: pl
Re: How can I obtain amplifier circuit loop gain?
« Reply #18 on: January 19, 2020, 09:54:50 am »
You can put an "AC 1" source at the normal input of completely unmodified amplifier, run AC analysis and plot V(output)/(V(in+)-V(in-)), where in+,in- are of course the two inputs to the differential pair. Philosophically speaking, that's the open loop gain.

It will give about the same plot, but I suspect that it may be slightly more optimistic at high frequencies and that's why people use other methods to asses stability of the circuit. Or maybe I'm wrong and it will be exactly the same :-//
 

Offline Lucky-LukaTopic starter

  • Supporter
  • ****
  • Posts: 240
  • Country: it
Re: How can I obtain amplifier circuit loop gain?
« Reply #19 on: January 19, 2020, 10:03:11 am »
It doesn't look good.
Memento audere semper.
 

Offline Cerebus

  • Super Contributor
  • ***
  • Posts: 10576
  • Country: gb
Re: How can I obtain amplifier circuit loop gain?
« Reply #20 on: January 19, 2020, 11:57:51 am »
Vf/Vin, where Vf is the feedback voltage.

What question do you think that is an answer to?

You mean.: "To what question do you think that is the answer?"

Anyhooo, it is actually the answer to the title of the original post.

A "grammar nazi" remark, on a forum where half the people can't spell and one third of the posts don't even even vaguely meet the rules of grammar? You're going to be busy...

Moreover, I don't think it's the answer to anything. There's no Vf defined anywhere here, and if we take the naïve assumption that Vf is the feedback voltage at the negative input then it's obviously wrong. It'd help if the poster clarified themselves.
« Last Edit: January 19, 2020, 12:05:00 pm by Cerebus »
Anybody got a syringe I can use to squeeze the magic smoke back into this?
 

Offline Wimberleytech

  • Super Contributor
  • ***
  • Posts: 1134
  • Country: us
Re: How can I obtain amplifier circuit loop gain?
« Reply #21 on: January 19, 2020, 03:44:41 pm »
Vf/Vin, where Vf is the feedback voltage.

What question do you think that is an answer to?

You mean.: "To what question do you think that is the answer?"

Anyhooo, it is actually the answer to the title of the original post.

A "grammar nazi" remark, on a forum where half the people can't spell and one third of the posts don't even even vaguely meet the rules of grammar? You're going to be busy...

Moreover, I don't think it's the answer to anything. There's no Vf defined anywhere here, and if we take the naïve assumption that Vf is the feedback voltage at the negative input then it's obviously wrong. It'd help if the poster clarified themselves.
Lighten up Cerebus...just having a little fun.
Frankly, I thought you were a bit pejorative in your response, so I was just playing along.

I am an old Texan...you should be better than I.

Anyway, back to the technical stuff.  Why do you say Vf/Vin  does not represent the loop gain?

« Last Edit: January 19, 2020, 04:04:46 pm by Wimberleytech »
 

Offline Cerebus

  • Super Contributor
  • ***
  • Posts: 10576
  • Country: gb
Re: How can I obtain amplifier circuit loop gain?
« Reply #22 on: January 19, 2020, 04:47:27 pm »
Anyway, back to the technical stuff.  Why do you say Vf/Vin  does not represent the loop gain?

What is Vf? No, seriously; it's just been introduced out of thin air.

If Vf is, as a I said a naïve assumption would be, the voltage at the negative input, then Vf/Vin can't be the loop gain.

Let's take the classic feedback equation

Vout = Vin A / (1 + Aβ)

Where Vout and  Vin are obvious, A is the open loop gain and β is the feedback fraction.

The loop gain (by definition) is Aβ, if Vf is as assumed above then Vf = βVout. So (substituting)

Vf / Vin = βVout / Vin ~= 1  and is not = Aβ.
 
The bloody SMF editor stole the keystrokes that should have produced a β  and instead decided that I wanted to save a 1/4 written draft message - bastard!
« Last Edit: January 19, 2020, 05:16:14 pm by Cerebus »
Anybody got a syringe I can use to squeeze the magic smoke back into this?
 
The following users thanked this post: Lucky-Luka

Offline Lucky-LukaTopic starter

  • Supporter
  • ****
  • Posts: 240
  • Country: it
Re: How can I obtain amplifier circuit loop gain?
« Reply #23 on: January 19, 2020, 05:02:30 pm »
@Cerebus
That's what I was thinking.
That's why I think that post doesn't add value to the thread.
I have asked a precise question and I have presented 2 solutions.
I can accept other solutions but it would be nice to detail them a little bit...
Cheers

p.s. English is not my first language so be patient :)
Memento audere semper.
 

Offline Lesolee

  • Regular Contributor
  • *
  • Posts: 223
  • Country: gb
    • publications list
Re: How can I obtain amplifier circuit loop gain?
« Reply #24 on: January 19, 2020, 05:09:10 pm »


The questions were:
(1) Are they equivalent?
Yes, from an theoretical point of view.
(2) Are they correct?
Yes, with the same caveat.

In some sense the ground generator version is less "accurate" since it removes the load of the feedback network, but 22K compared to the 10R load is pretty negligible.
In some sense the floating generator version is more complicated since the plot is a ratio rather than a direct reading, but again that is evidently not an issue.
 
The following users thanked this post: Lucky-Luka


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf