Author Topic: AC analysis results (I think)  (Read 1310 times)

0 Members and 1 Guest are viewing this topic.

Offline JiggerypookTopic starter

  • Contributor
  • Posts: 11
  • Country: ca
AC analysis results (I think)
« on: May 27, 2019, 04:18:34 am »
I have some newbie questions if anyone is willing to answer. I have modified a schematic that I assume was created by Kleinstein and through testing in ltSpice have somehow made it seemingly stable. I have attached a screen shot of what is puzzling me. I copied this test from another post on this forum.

1. What am I looking at? Frequency response or phase relationship or neither?
2. What does it indicate about the circuit? Is it a good result or do I start over?
3. Is there a better test to indicate whether this circuit could be stable under different load conditions?

I would like to build this power supply but am struggling to understand all of the stability issues that may apply. I have read up on bode plots and trying to understand poles and zeros but I'm not there yet. I'm sure that it has to do with leading and lagging voltage and current relationships but then again I could be wrong. I am trying to learn. As far as my understanding goes this test may be meaningless for what I am trying to do.

Please help.

p.s. I don't know how to resize so I hope this is ok.
« Last Edit: May 27, 2019, 04:27:22 am by Jiggerypook »
Every complex problem has a simple, easy to understand, wrong answer.
 

Online iMo

  • Super Contributor
  • ***
  • Posts: 5570
  • Country: va
Re: AC analysis results (I think)
« Reply #1 on: May 27, 2019, 12:07:02 pm »
The PSUs could be analyzed for the stability, with a phase margin output (and gain).

There is a tech-article from ADI(LT) I've been using.

https://www.analog.com/en/technical-articles/ltspice-extracting-switch-mode-power-supply-loop-gain-in-simulation-and-why-you-usually-don-t-need.html

Readers discretion is advised..
 

Offline JiggerypookTopic starter

  • Contributor
  • Posts: 11
  • Country: ca
Re: AC analysis results (I think)
« Reply #2 on: May 27, 2019, 01:15:01 pm »
Thanks IMO, I will read that. It never came up in any of my searches. I just hope that i'm not too thick. lol
Every complex problem has a simple, easy to understand, wrong answer.
 

Offline JiggerypookTopic starter

  • Contributor
  • Posts: 11
  • Country: ca
Re: AC analysis results (I think)
« Reply #3 on: May 27, 2019, 02:55:36 pm »
The PSUs could be analyzed for the stability, with a phase margin output (and gain).

There is a tech-article from ADI(LT) I've been using.

https://www.analog.com/en/technical-articles/ltspice-extracting-switch-mode-power-supply-loop-gain-in-simulation-and-why-you-usually-don-t-need.html
[/quote
I couldn't seem to co-relate a switcher to a linear but this was somewhat helpful.

I assume that the feedback loop in a linear supply would be the error amps. My experience with this circuit was that the current limit was the cause of the oscillations, therefore I attempted to add positive feedback to create some hysteresis as I had seen in one of Dave's videos. This had a secondary effect that allowed me to bring the CV/CC threshold more closely together, improving accuracy.

So, does this mean that I need to inject the test signal into the inverting pin of the current limit op? If so, between C9 and the inverting pin or between R26 and the inverting pin?
Do I just probe the output or output/input?

 

Sorry for all of the newb questions, I'm also just learning ltSpice.



Thanks again for any help, I am trying to learn but right now it's a steep grade and I want to build a stable power supply to further my understanding of electronics.
« Last Edit: May 27, 2019, 02:58:31 pm by Jiggerypook »
Every complex problem has a simple, easy to understand, wrong answer.
 

Offline JiggerypookTopic starter

  • Contributor
  • Posts: 11
  • Country: ca
Re: AC analysis results (I think)
« Reply #4 on: May 28, 2019, 04:18:13 am »
OK, I read the article that IMO referred me to and attempted to apply it to a linear supply but quite obviously I have no clue. I am a total beginner and have little understanding of how to address the stability issues of a linear power supply except to zero in on the problem area and use trial and error methods to achieve stability, as I have done so far.

So, thanks again IMO.
Here are my strange results:
Every complex problem has a simple, easy to understand, wrong answer.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: AC analysis results (I think)
« Reply #5 on: May 28, 2019, 04:50:01 am »
There's no supply voltage because the mains source is AC=0.

Replace it with a DC source ;)

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline JiggerypookTopic starter

  • Contributor
  • Posts: 11
  • Country: ca
Re: AC analysis results (I think)
« Reply #6 on: June 14, 2019, 07:33:18 am »
Well, I took Tim's advice and had results that I could not make sense of. Thank you Tim for your response. It was unexpected. It's obvious that I am too noob to think that I can analyse a circuit such as this using techniques that professionals use in ltSpice. Was I looking at frequency response in a DC system or phase relationships between voltage and current at different frequencies. (confused) and then what does that even mean for stability? (lost).

 I realize the limitations of simulations and do not put my full trust in the results. I played with Circuitmaker decades ago and it was horrible to say the least.
 I have limited background and was sort of asking what I should be looking for, and some guidance in that respect.
 I had previously done a lot of work trying to make that old Greek supply that everybody knows about work, but is slower than snails in January in current limiting, and is faulted in many other ways.
 I came upon Kleinstien's model in a post on this forum and loved it's simplicity and functionality.
 I had been looking at HP drawings and been totally lost.
 I need a power supply to use as a tool to pursue my interest in electronics now that I am in my senior years and have the time. That being said I have a few questions, and I hope that I don't offend anyone on this forum.

1. Is anyone willing to help me on a scale that I can understand so that I may learn? (I never took calculus in high school 50 years ago) but I am a grade 12 graduate and electrician by trade.
2. Has anyone else used this design, or one similar successfully? (examples to learn from)
3. Should I just build it and let the smoke out of multiple components? (Learn the hard way)
4. Do I start too many sentences with "I"?

I don't expect, and am not asking for someone to do the work for me, that's not learning, I just need to be pointed in the right direction.
I love "Daves" videos and have been following the courses from "Digilent" if I am allowed to say that here.
I know that once I feel more comfortable on this forum I will be starting fewer sentences with "I" and maybe even eventually helping someone myself.
I hope that I get as many eyes on this post as I have in my sentences. lol!
Every complex problem has a simple, easy to understand, wrong answer.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: AC analysis results (I think)
« Reply #7 on: June 14, 2019, 08:27:52 am »
To expand that a bit more -- an AC analysis requires DC voltages and bias.  The mains input is a sine feeding diodes -- the diodes will be approximated at the sine's DC setting (which is zero), and the transformer blocks DC anyway, so there's no power.  A rectifier is nonlinear, and nonlinearity just isn't simulated in an AC analysis.

So you need to replace rectifier with a DC source so it gets power at all. :)

Or, do a transient analysis, and figure out the behavior from waveforms.

To get all the rich nonlinear behavior, you have to use transient analysis.  If you want to get AC analysis style results, you basically have to instrument it yourself -- set up sources and AC detectors and let it run for a while to reach steady state.  Big pain. :(

You should still use a DC source, since you'll be waiting forever for the rectifier to start up.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline JiggerypookTopic starter

  • Contributor
  • Posts: 11
  • Country: ca
Re: AC analysis results (I think)
« Reply #8 on: June 15, 2019, 09:10:20 am »
Here is what I have managed. I'm attempting the Middlebrook analysis but it is all very new to me so I have probably made several mistakes. I assumed that the feedback loop for current control would be the inverting input of U2-a on my schematic. If this is not correct, please let me know.

The resulting graph doesn't seem to look like anything I have seen before so I assume that I have done something wrong. Probably a good example of "garbage in-garbage out.

Actually I really don't know what to look for that indicate possibilities of stability.
Do the crossing points indicate where the signals are in phase? What qualities am I looking for to indicate stability?
What role if any does Impedance, Resonance,or Reactance play?

I know, too many questions, read more, etc. I just need to know if I am remotely headed in the right direction.

Every complex problem has a simple, easy to understand, wrong answer.
 

Offline JiggerypookTopic starter

  • Contributor
  • Posts: 11
  • Country: ca
Re: AC analysis results (I think)
« Reply #9 on: June 16, 2019, 06:20:31 am »
OK. I've taken a complex problem, looking for a simple answer which doesn't exist (therefore little response) and it was wrong on my part. thanks IMO and Tim for enlightening me on this subject.
I'm just going to breadboard the thing and hope for the best. I've just got to dream up some good torture tests now and put the scope to use. That's what I call jiggerypooking. Blindly experimenting and learning from whatever results. Who cares about the math anyways. (smoke-crackle-pop!) lol. Just like the zeppelin.
Thank you Klienstien for the circuit to replace the 1984 circuit that I had originally started this project with in 1986 and now finally have the time to finish in my retirement. If it works out you will hear from me, if it doesn't you will hear from me. I never gave up in 1986 I just had no time and the internet was brand new and not very useful. bla bla bla ........ :blah:

Every complex problem has a simple, easy to understand, wrong answer.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf