Author Topic: Help with how this PSU works. (With revised design)  (Read 7506 times)

0 Members and 1 Guest are viewing this topic.

Offline Evangelopoulos PanagiotisTopic starter

  • Contributor
  • Posts: 48
  • Country: gr
Re: Help with how this PSU works. (With revised design)
« Reply #25 on: May 12, 2020, 10:49:58 pm »
About the Pot I can't say until the other problems are corrected in the schematic.

It's a bit late right now so I'll change the schematic first thing in the morning. I'll change the D10 and make sure the rest of the connections are ok.
Thank you for the help so far,
E.P.
 

Offline xavier60

  • Super Contributor
  • ***
  • Posts: 3034
  • Country: au
Re: Help with how this PSU works. (With revised design)
« Reply #26 on: May 12, 2020, 11:14:55 pm »
I dislike seeing Pots used as rheostats or VRs. The simplest way I could figure to avoid it meant using a negative reference.
https://www.eevblog.com/forum/projects/lab-power-supply-the-lost-current/?action=dlattach;attach=400354
HP 54645A dso, Fluke 87V dmm,  Agilent U8002A psu,  FY6600 function gen,  Brymen BM857S, HAKKO FM-204, New! HAKKO FX-971.
 

Offline aheid

  • Regular Contributor
  • *
  • Posts: 245
  • Country: no
Re: Help with how this PSU works. (With revised design)
« Reply #27 on: May 12, 2020, 11:15:21 pm »
I can't see the labels clearly. Can you post it with higher res?

I forget not all have LTSpice :) See attachment. R12 was just to separate the connections in the sim to make the labels on the graph clearer. Note that I tried to make this as a faithful representation of the original schematic.
 

Offline xavier60

  • Super Contributor
  • ***
  • Posts: 3034
  • Country: au
Re: Help with how this PSU works. (With revised design)
« Reply #28 on: May 12, 2020, 11:18:37 pm »
That looks right.
HP 54645A dso, Fluke 87V dmm,  Agilent U8002A psu,  FY6600 function gen,  Brymen BM857S, HAKKO FM-204, New! HAKKO FX-971.
 
The following users thanked this post: aheid

Offline Evangelopoulos PanagiotisTopic starter

  • Contributor
  • Posts: 48
  • Country: gr
Re: Help with how this PSU works. (With revised design)
« Reply #29 on: May 13, 2020, 07:18:20 am »
About the Pot I can't say until the other problems are corrected in the schematic.

Ok I think I corrected everything. It's almost the same as the original.
 

Offline xavier60

  • Super Contributor
  • ***
  • Posts: 3034
  • Country: au
Re: Help with how this PSU works. (With revised design)
« Reply #30 on: May 13, 2020, 07:57:44 am »
A resistor could be put across the 10K Pot, ideally a 1.1K.
A green CV LED can be added. LEDs can and do go open circuit. I would rather have the LEDs not included in the control path, instead have them wired from the opamps' outputs to + control rail with their own dropper resistors.
Make provision for compensation components. A capacitor from ouput to inverting input of the CC opamp.
A capacitor and series resistor from output to inverting input of the CV opamp. 200mil footprint will do for the capacitors.
HP 54645A dso, Fluke 87V dmm,  Agilent U8002A psu,  FY6600 function gen,  Brymen BM857S, HAKKO FM-204, New! HAKKO FX-971.
 

Offline xavier60

  • Super Contributor
  • ***
  • Posts: 3034
  • Country: au
Re: Help with how this PSU works. (With revised design)
« Reply #31 on: May 13, 2020, 08:08:00 am »
If the CC loop is found to be unstable and a compensation capacitor added, it will have the side effect of slowing down the current limiting response to sudden overloads. A diode across R13 should improve the response. Could someone test this in a simulation?
HP 54645A dso, Fluke 87V dmm,  Agilent U8002A psu,  FY6600 function gen,  Brymen BM857S, HAKKO FM-204, New! HAKKO FX-971.
 

Offline Evangelopoulos PanagiotisTopic starter

  • Contributor
  • Posts: 48
  • Country: gr
Re: Help with how this PSU works. (With revised design)
« Reply #32 on: May 13, 2020, 09:16:56 am »
A resistor could be put across the 10K Pot, ideally a 1.1K.
A green CV LED can be added. LEDs can and do go open circuit. I would rather have the LEDs not included in the control path, instead have them wired from the opamps' outputs to + control rail with their own dropper resistors.
Make provision for compensation components. A capacitor from ouput to inverting input of the CC opamp.
A capacitor and series resistor from output to inverting input of the CV opamp. 200mil footprint will do for the capacitors.

A capacitor to stabilize the CC loop sounds like a great idea since the original PSU died from a sudden load I think.

Will a 1uF do the job for both the CC and CV loops? How can I choose the value of the series resistor in the CV loop?

The cathode of the diode across R13 should connect to the inverting input right? 
 

Offline xavier60

  • Super Contributor
  • ***
  • Posts: 3034
  • Country: au
Re: Help with how this PSU works. (With revised design)
« Reply #33 on: May 13, 2020, 09:33:36 am »
A resistor could be put across the 10K Pot, ideally a 1.1K.
A green CV LED can be added. LEDs can and do go open circuit. I would rather have the LEDs not included in the control path, instead have them wired from the opamps' outputs to + control rail with their own dropper resistors.
Make provision for compensation components. A capacitor from ouput to inverting input of the CC opamp.
A capacitor and series resistor from output to inverting input of the CV opamp. 200mil footprint will do for the capacitors.

A capacitor to stabilize the CC loop sounds like a great idea since the original PSU died from a sudden load I think.

Will a 1uF do the job for both the CC and CV loops? How can I choose the value of the series resistor in the CV loop?

The cathode of the diode across R13 should connect to the inverting input right?
That's right about the diode.
100pF and 47K for the CV opamp.
10nF for the CC opamp.
Do some test first to see if the parts are really needed.
I have just realized that R24 will have some affect on loop gain. I'm not certain how it should be dealt with if at all.

:Try to make space for a capacitor across R24.
« Last Edit: May 13, 2020, 09:37:02 am by xavier60 »
HP 54645A dso, Fluke 87V dmm,  Agilent U8002A psu,  FY6600 function gen,  Brymen BM857S, HAKKO FM-204, New! HAKKO FX-971.
 
The following users thanked this post: Evangelopoulos Panagiotis

Offline Evangelopoulos PanagiotisTopic starter

  • Contributor
  • Posts: 48
  • Country: gr
Re: Help with how this PSU works. (With revised design)
« Reply #34 on: May 13, 2020, 09:47:57 am »
That's right about the diode.
100pF and 47K for the CV opamp.
10nF for the CC opamp.
Do some test first to see if the parts are really needed.
I have just realized that R24 will have some affect on loop gain. I'm not certain how it should be dealt with if at all.

:Try to make space for a capacitor across R24.

I added them and If I see they are not needed I'll leave them unpopulated. Same thing for the capacitor across the R24 pot.
 

Offline xavier60

  • Super Contributor
  • ***
  • Posts: 3034
  • Country: au
Re: Help with how this PSU works. (With revised design)
« Reply #35 on: May 13, 2020, 09:58:28 am »
That should cover everything.
Modern LEDs are so efficient, only a small current, as low as 0.3mA.
HP 54645A dso, Fluke 87V dmm,  Agilent U8002A psu,  FY6600 function gen,  Brymen BM857S, HAKKO FM-204, New! HAKKO FX-971.
 

Offline Evangelopoulos PanagiotisTopic starter

  • Contributor
  • Posts: 48
  • Country: gr
Re: Help with how this PSU works. (With revised design)
« Reply #36 on: May 13, 2020, 10:06:32 am »
That should cover everything.
Modern LEDs are so efficient, only a small current, as low as 0.3mA.

Awesome! I learned quite a bit from this project. I shall start the board design now.

Thank you Xavier for the help. I really appreciate it.  :-+
 
 

Offline Evangelopoulos PanagiotisTopic starter

  • Contributor
  • Posts: 48
  • Country: gr
Re: Help with how this PSU works. (With revised design)
« Reply #37 on: May 13, 2020, 05:49:43 pm »
I completed the project thanks to the help of you all.

Let's hope it actually works!  :scared:

The project can be found below incase someone wants to build one. 
https://easyeda.com/EvangelopoulosPanagiotis/Linear-Current-Limiting-PSU

Please excuse my rushed board design.

Have a nice day!
E.P.
 

Offline Dave

  • Super Contributor
  • ***
  • Posts: 1356
  • Country: si
  • I like to measure things.
Re: Help with how this PSU works. (With revised design)
« Reply #38 on: May 13, 2020, 06:11:27 pm »
The fuse protecting the negative clamping diode on the output makes no sense. Let's say the user accidentally connects another power source to your PSU backwards. A large current flows through that diode, the fuse blows. You just saved a 5 cent diode and sacrificed 30€ worth of other components to do it.

If you decide to use a fuse to protect the output, it should be wired in a way that it cuts the circuit when an overload causes it to burst, thus protecting the power supply. In series with the output would be fine (the resistance of the fuse is going to increase the output impedance of the power supply, but it shouldn't be too bad).
<fellbuendel> it's arduino, you're not supposed to know anything about what you're doing
<fellbuendel> if you knew, you wouldn't be using it
 
The following users thanked this post: Evangelopoulos Panagiotis

Offline Evangelopoulos PanagiotisTopic starter

  • Contributor
  • Posts: 48
  • Country: gr
Re: Help with how this PSU works. (With revised design)
« Reply #39 on: May 13, 2020, 06:19:54 pm »
The fuse protecting the negative clamping diode on the output makes no sense. Let's say the user accidentally connects another power source to your PSU backwards. A large current flows through that diode, the fuse blows. You just saved a 5 cent diode and sacrificed 30€ worth of other components to do it.

If you decide to use a fuse to protect the output, it should be wired in a way that it cuts the circuit when an overload causes it to burst, thus protecting the power supply. In series with the output would be fine (the resistance of the fuse is going to increase the output impedance of the power supply, but it shouldn't be too bad).

Damn you're right... Why did I thought it would save the rest?  :palm: I'll put the fuse in series.
 

Offline jaycee

  • Regular Contributor
  • *
  • Posts: 206
  • Country: gb
Re: Help with how this PSU works. (With revised design)
« Reply #40 on: May 13, 2020, 08:44:52 pm »
Jumping in VERY late on this, but essentially the regulator circuit used here is a very common one, I believe Harrison Labs developed it first, and the concept is used in quite a lot of power supplies including the Mastech ones. The neat trick is that the opamp supplies float around the main regulator's +ve output.

Tap switching is simplest and best done by a relay. A simple circuit using a comparator checking the output voltage can be used to know when to switch taps. My own power supply switches between 9VAC and 18VAC taps using a simple two transistor circuit. You would adjust R20/R47 accordingly. "VRAW" in this case is the raw voltage fed to the pass transistor

 

Offline xavier60

  • Super Contributor
  • ***
  • Posts: 3034
  • Country: au
Re: Help with how this PSU works. (With revised design)
« Reply #41 on: May 13, 2020, 11:40:47 pm »
It might be interesting to know that current sourcing PSU designs such as this aren't immediately damaged by applied reverse polarity.
The output transistors might experience much higher than usual dissipation and the output capacitor will eventually burst.
I'm not saying that there is no need for the diode across the output. An example for its need is when PSUs are connected in series.

« Last Edit: May 14, 2020, 08:12:50 am by xavier60 »
HP 54645A dso, Fluke 87V dmm,  Agilent U8002A psu,  FY6600 function gen,  Brymen BM857S, HAKKO FM-204, New! HAKKO FX-971.
 

Offline aheid

  • Regular Contributor
  • *
  • Posts: 245
  • Country: no
Re: Help with how this PSU works. (With revised design)
« Reply #42 on: May 13, 2020, 11:47:16 pm »
Just a thought but one thing I've been missing in my bench supply, which I'm considering hacking in, is a way to have external voltage probes. Idea was to be able to switch from sensing voltage at the output terminals to have sense terminals, to compensate for voltage drop in cables etc.

It'd require an extra opamp control circuit, but can be diode-OR'ed as well so should be relatively easy to add such a design.

Maybe not for rev1 but perhaps something for rev2?
 

Offline jaycee

  • Regular Contributor
  • *
  • Posts: 206
  • Country: gb
Re: Help with how this PSU works. (With revised design)
« Reply #43 on: May 14, 2020, 12:08:57 am »
Just a thought but one thing I've been missing in my bench supply, which I'm considering hacking in, is a way to have external voltage probes. Idea was to be able to switch from sensing voltage at the output terminals to have sense terminals, to compensate for voltage drop in cables etc.

It'd require an extra opamp control circuit, but can be diode-OR'ed as well so should be relatively easy to add such a design.

Maybe not for rev1 but perhaps something for rev2?

Strictly this doesnt require extra opamps, it does require careful routing though. Using separate opamps for sense and error does make it easier though. My own PSU has separate sense amps, and has voltage sense right up to the terminals, but I used separate sense amps primarily because I wanted to feed the output of those to ADC's for metering purposes.

The Elektor "Precison PSU" from 1982 is the only diagram I've seen which shows single opamps and how to correctly do remote voltage sense. Ignore the 723 in this circuit, it is used only as a voltage reference (a bit of a bizarre choice!) - IC2 is doing the voltage control, IC3 is doing the current control

 

Offline xavier60

  • Super Contributor
  • ***
  • Posts: 3034
  • Country: au
Re: Help with how this PSU works. (With revised design)
« Reply #44 on: May 14, 2020, 12:10:24 am »
Just a thought but one thing I've been missing in my bench supply, which I'm considering hacking in, is a way to have external voltage probes. Idea was to be able to switch from sensing voltage at the output terminals to have sense terminals, to compensate for voltage drop in cables etc.

It'd require an extra opamp control circuit, but can be diode-OR'ed as well so should be relatively easy to add such a design.

Maybe not for rev1 but perhaps something for rev2?
You could break the connections where R15 and R24 connect to 0V and -output. Then insert something like a 10Ω, 10uF and back to back 1A didoes all in parallel. There will be a small voltage change depending on if the sense leads are used or not.
Are you simulating the present design?
HP 54645A dso, Fluke 87V dmm,  Agilent U8002A psu,  FY6600 function gen,  Brymen BM857S, HAKKO FM-204, New! HAKKO FX-971.
 

Offline aheid

  • Regular Contributor
  • *
  • Posts: 245
  • Country: no
Re: Help with how this PSU works. (With revised design)
« Reply #45 on: May 14, 2020, 02:17:55 am »
Yeah, the idea with the extra opamp is to limit the terminal voltage to something slightly above the setpoint, in case the sense probes aren't connected to the right part of the DUT or similar.

Of course not strictly needed, just some extra dumbass protection for absent-minded folks like me :)
 

Offline Evangelopoulos PanagiotisTopic starter

  • Contributor
  • Posts: 48
  • Country: gr
Re: Help with how this PSU works. (With revised design)
« Reply #46 on: May 14, 2020, 07:57:53 am »
While I was designing the circuit I thought to my self: "Hey! I can add a switch to the Vsense wire so I can have a remote sense for voltage drop on the leads!"
But then it hit me. What if the sense wires are not connected or connected in the wrong place? It will probably cause the output to rocket to 30V and destroy everything.
So I quickly dropped that idea. But it is something that I might add if there is a need for a v2.0 in the future ;) I might also add the auto tap switching in v2.0 if I see a need for it.

I'll post here again when that time comes. 
 

Offline jaycee

  • Regular Contributor
  • *
  • Posts: 206
  • Country: gb
Re: Help with how this PSU works. (With revised design)
« Reply #47 on: May 14, 2020, 09:55:52 pm »
Sense wires not being connected would be catered for by the resistors that are between the outputs and the sense. You would get some error, but it would stop the voltage rocketing too far away from the setpoint.

Even if you do not need remote sensing, it is worth incorporating it into the board, even if you then just have sense wires connected directly at the output terminals. It improves regulator accuracy considerably.
 

Offline Evangelopoulos PanagiotisTopic starter

  • Contributor
  • Posts: 48
  • Country: gr
Re: Help with how this PSU works. (With revised design)
« Reply #48 on: June 26, 2020, 12:00:43 pm »
Greetings again,

I have good news! The PCBs and components finally came! I've put everything together and powered the PSU up!
No magic smoke escaped!

I only found two "bugs" in the design. First, the two LEDs seem to be the other way around for some reason. :P When CC lights up that means that the voltage is stable and when CV lights up that means that the current is being limited. Just keep that in mind if anyone decides to build this design.

The second "bug" is that I could only get a maximum output voltage of 18V even though my input supply was 45V. I tried changing the 47k (R20) pot but with no luck. After poking around I found that the reason was the 4.7k (R18) resistor. I put a trimmer in parallel and found that with an extra 3.2k resistor in parallel with R18 gives me an output voltage of 43V and after tweaking R20 I got it down to the 42V I wanted.

Anyway, I did some quick tests and it seems to working ok for my taste. So I'm overall very pleased!

Thank you all  for your help!
Have a nice day,
E.P.

P.S. have some pictures :P Note the bodge resistor.
 
The following users thanked this post: aheid

Offline aheid

  • Regular Contributor
  • *
  • Posts: 245
  • Country: no
Re: Help with how this PSU works. (With revised design)
« Reply #49 on: June 26, 2020, 10:02:01 pm »
Awesome, cool to see it built, so thanks for sharing!
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf