Author Topic: EEVblog #221 - Lab Power Supply Design - Part 1  (Read 10090 times)

0 Members and 1 Guest are viewing this topic.

Offline ndictu

  • Regular Contributor
  • *
  • Posts: 210
  • Country: sk
EEVblog #221 - Lab Power Supply Design - Part 1
« on: November 26, 2011, 11:06:25 pm »
I guess you will get to this in the second part but since that is not up yet I have a small question.

At the end when you say that the current from the LM3080 adjust pin through the 2k resistors isn't exactly zero (LT datasheet says 10uA typical), this will do two thing - make the minimum output voltage higher (10uA*2kOhm = 20mV), and also offset it through the entire range.

So, my noob question is, could you fix it by connecting the inverting terminal of the Vset buffer opamp to the adjust pin directly? This will fix the offset on higher voltages. Also, if there was some low power negative supply for the opamp it could also go down to 0V by driving it's output negative.

Is this correct or am I wrong?
 

Offline firewalker

  • Super Contributor
  • ***
  • Posts: 2368
  • Country: gr
Re: EEVblog #221 - Lab Power Supply Design - Part 1
« Reply #1 on: November 27, 2011, 09:47:15 am »
I had post it before. A power supply with an mcu I had implemented years ago.

http://tuxgraphics.org/electronics/201005/bench-power-supply-v3.shtml

This version combines PWM and a DAC (12 bit resolution in total). Everything controlled by an ATmega8 Atmel mcu.

I had made the older version (only with DAC).

http://tuxgraphics.org/electronics/200506/article379.shtml

I think chet_16 implemented the newer version.

Alexander.
Become a realist, stay a dreamer.

 
The following users thanked this post: JoseLog

Offline timelessbeing

  • Frequent Contributor
  • **
  • Posts: 804
  • Country: 00
Re: EEVblog #221 - Lab Power Supply Design - Part 1
« Reply #2 on: November 27, 2011, 10:31:54 am »
sweet tutorial Dave. Please do a switcher next.
 

Offline Mechatrommer

  • Super Contributor
  • ***
  • Posts: 9707
  • Country: my
  • reassessing directives...
Re: EEVblog #221 - Lab Power Supply Design - Part 1
« Reply #3 on: November 27, 2011, 11:16:01 am »
I had post it before. A power supply with an mcu I had implemented years ago.
http://tuxgraphics.org/electronics/201005/bench-power-supply-v3.shtml
This version combines PWM and a DAC (12 bit resolution in total). Everything controlled by an ATmega8 Atmel mcu.
thumbs up!
It's extremely difficult to start life.. one features of nature.. physical laws are mathematical theory of great beauty... You may wonder Why? our knowledge shows that nature is so constructed. We simply have to accept it. One could describe the situation by saying that... (Paul Dirac)
 

Offline BravoV

  • Super Contributor
  • ***
  • Posts: 6781
  • Country: 00
  • +++ ATH1
Re: EEVblog #221 - Lab Power Supply Design - Part 1
« Reply #4 on: November 27, 2011, 11:31:03 am »
Great video Dave !

Just don't forget the cc and cv led indicators. :D

Offline Bloch

  • Supporter
  • ****
  • Posts: 440
  • Country: dk
Re: EEVblog #221 - Lab Power Supply Design - Part 1
« Reply #5 on: November 27, 2011, 11:59:35 am »
I had post it before. A power supply with an mcu I had implemented years ago.

I am a bit confused why it is made like it is.

I mean why not use 1 or 2 op amps and have a "fast" regulating.

Pro: Cost.
Con: Regulation not fast enough ?

Pls comment ...
 

Offline ndictu

  • Regular Contributor
  • *
  • Posts: 210
  • Country: sk
Re: EEVblog #221 - Lab Power Supply Design - Part 1
« Reply #6 on: November 27, 2011, 12:02:18 pm »
I had post it before. A power supply with an mcu I had implemented years ago.

http://tuxgraphics.org/electronics/201005/bench-power-supply-v3.shtml

This version combines PWM and a DAC (12 bit resolution in total). Everything controlled by an ATmega8 Atmel mcu.

I have seen this posted here a few times, most people didn't like there was a micro in the control loop, making CC limiting slow(er). Dave's design is basically fast analog loop with digital settings.
 

alm

  • Guest
Re: EEVblog #221 - Lab Power Supply Design - Part 1
« Reply #7 on: November 27, 2011, 12:52:42 pm »
Last time I checked the tuxgraphics source the control algorithm was very basic, change DAC 1 LSB if voltage/current not within limits. Together with the slow sampling rate, the result was a control loop that could barely track mains ripple, let alone fast transients.

I like Dave's design much better, though I am curious about the amount of noise in constant current mode.
 

Offline firewalker

  • Super Contributor
  • ***
  • Posts: 2368
  • Country: gr
Re: EEVblog #221 - Lab Power Supply Design - Part 1
« Reply #8 on: November 27, 2011, 02:00:56 pm »
It's more a proof o concept that a psu can be made with a single mcu. It has 12 bit res, current sensing and a nice interface.

I had no intention to compare the two designs. I though it was relevant and the article is quite nice.

Someone could use the voltage control (DAC part) and the interface to control an opamp based design (e.g. Dave's) with analog loops e.t.c.

Also in my opinion a beginner can use the article to understand some things about power supplies.

No matter what I believe that with a good supply before the digital control it worth it's money. I would prefer it against any cheap, PRC made psu.

Alexander.

I had post it before. A power supply with an mcu I had implemented years ago.

I am a bit confused why it is made like it is.

I mean why not use 1 or 2 op amps and have a "fast" regulating.

Pro: Cost.
Con: Regulation not fast enough ?

Pls comment ...
« Last Edit: November 27, 2011, 02:09:04 pm by firewalker »
Become a realist, stay a dreamer.

 

Offline Mechatrommer

  • Super Contributor
  • ***
  • Posts: 9707
  • Country: my
  • reassessing directives...
Re: EEVblog #221 - Lab Power Supply Design - Part 1
« Reply #9 on: November 27, 2011, 02:21:13 pm »
i believe your mcu based PSU can be extended to things like pure sine wave inverter etc, but i'm not so sure, just hoping.
It's extremely difficult to start life.. one features of nature.. physical laws are mathematical theory of great beauty... You may wonder Why? our knowledge shows that nature is so constructed. We simply have to accept it. One could describe the situation by saying that... (Paul Dirac)
 

Offline FenderBender

  • Super Contributor
  • ***
  • Posts: 1118
  • Country: us
    • The Solid State Workshop
Re: EEVblog #221 - Lab Power Supply Design - Part 1
« Reply #10 on: November 27, 2011, 07:17:15 pm »
Can the LT3080 be used to regulate a negative voltage? Transformer windings/grounds have to be isolated or something?
 

alm

  • Guest
Re: EEVblog #221 - Lab Power Supply Design - Part 1
« Reply #11 on: November 27, 2011, 07:39:01 pm »
If the power supply is floating, you can make the supply negative by tying the positive output to ground (or any other reference potential). You would typically use pnp pass transistors for negative regulators, but negative regulators are only necessary if you're using a center tapped transformer as input.
 

Offline EEVblog

  • Administrator
  • *****
  • Posts: 31346
  • Country: au
    • EEVblog
Re: EEVblog #221 - Lab Power Supply Design - Part 1
« Reply #12 on: November 28, 2011, 12:04:38 am »
So, my noob question is, could you fix it by connecting the inverting terminal of the Vset buffer opamp to the adjust pin directly?

Spoiler Alert!



Dave.
 

Offline ndictu

  • Regular Contributor
  • *
  • Posts: 210
  • Country: sk
Re: EEVblog #221 - Lab Power Supply Design - Part 1
« Reply #13 on: November 28, 2011, 07:28:00 pm »
Spoiler Alert!

I take that as a yes...

But if I could see that it must be pretty obvious. Maybe you should take some lessons from M. Night Shyamalan about putting twists to your episodes :)
 

Offline l4rtt-1

  • Contributor
  • Posts: 11
Re: EEVblog #221 - Lab Power Supply Design - Part 1
« Reply #14 on: November 29, 2011, 04:43:20 pm »
Question about the final circuit. Why the 1K resistor is needed at the adjust pin?

The 1K resistor at the buffer's output is to protect the op-amp when the transistor is conducting, but what's the other resistor for?

Lauri

edit. One more thing. Why not put the buffer's 1K output resistor between the transistor's emitter and ground. That would prevent shorting the buffer's output when transistor is on. And leave the other adj. pin 1K resistor out of the circuit completly.

edit2. Well I figured it out, my idea wouldn't work, the buffer's output voltage would be present at the adj. pin when the transistor is on and that way completly prevent the current limiting circuit to work.

But still, what's the other resistor for?
« Last Edit: November 29, 2011, 05:02:54 pm by l4rtt-1 »
 

Offline EEVblog

  • Administrator
  • *****
  • Posts: 31346
  • Country: au
    • EEVblog
Re: Dilbert Power supply. Re: EEVblog #221 - Lab Power Supply Design
« Reply #15 on: December 01, 2011, 04:46:32 am »
Can someone tell me how to embed the clip in a post? I am assuming just pasting the YouTube share URL isn't going to cut it.

Yes, just paste the Youtube URL and SimpleMachines does the embedding for you:



Dave.
 

Offline EEVblog

  • Administrator
  • *****
  • Posts: 31346
  • Country: au
    • EEVblog
Re: EEVblog #221 - Lab Power Supply Design - Part 1
« Reply #16 on: December 01, 2011, 11:22:32 am »
Thanks Dave, but are you saying the YouTube URL pastes differently from the share URL I used?

Yes. The forum software only knows and translates youtube.com URL's it does not know what the short URL's are.
Just paste the youtube URL in and it handles it.

Quote
Can you elaborate on the choice of OP Amp please? You mentioned some details about the choice but if it is obvious to most viewers the finer point was lost on me. IIRC you mentioned noise and I assume that lower noise means better output voltage smoothness. Correct me if I am wrong please. Actually this perhaps should've been posted in new blog suggestions. <-hint

It's all to do with offset voltage.
Common opamps like the LM324 have several mV of offset. That's not good in your are trying to get say 1mA accuracy on the current measurement. Not so important on the voltage buffers though.
The TLC2272 or TLC2252 is a fairly cheap precision opamp I like to use as a generic "step up" in performance from the basic LM324 types. They are also rail-to-rail input and output which is nice.

Dave.
 
The following users thanked this post: JoseLog


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf