No Script, No Fear, All Opinion
RSS icon Home icon
  • EEVblog #222 – Lab Power Supply Design – Part 2

    Posted on November 29th, 2011 EEVblog 19 comments


    PART 1 is HERE:
    http://www.youtube.com/watch?v=CIGjActDeoM
    Part 2 of Dave’s constant current linear regulated lab power supply design. An in-depth look at the LT3080 datasheet, the LM334 current source, and a few more circuit tweaks before it’s ready to build in the next episode to see if it works!

    And thanks to John from LT who pointed out this video from LT about the LT3080, given by Bob Dobkin himself, the LT CTO and designer of the LM317!

    http://video.linear.com/11

    Be Sociable, Share!
    • Steve

      Hi:

      Since you’re nearly there, why don’t you take the voltage feedback from the output directly?

    • Maik

      Would be nice to see how you get something like this from theory to breadboard and finally to PCB. Yeah it’s very simple, but maybe we could learn something :)

    • Lestro

      Hummm. I’m wondering… should I quit building my new lab power supply and wait to the end of Dave’s lab PS?…
      Maybe this one will be better!?

      • http://www.eevblog.com EEVblog

        If anything, it will be different!

    • John

      With the help of a friend I was able to design a lab power supply which started out with two LM317s but ended up with one LM317 doing double duty with a transistor in the circuit to limit the current just like in your design. The other major difference is that mine had the shunt resistor on the low end.

      The hardest part for me was figuring out how to add voltage and current displays to my circuit. I ended up using a single ATMEGA328 to drive two 4 digiti 7-seg displays using two ADC inputs for the Isense and Vsense.

      Are you planning on have displays for the current and voltage?

    • someone

      Hey Dave,

      Just watched the video and @35:10 I think you can’t do that, simple because them you lost your current limit feature, even if you exceed your current and and the transistor goes ON your ampop of Vset will compensate.

    • http://lidi.uw.hu/index.php Lidi

      Hi !

      What a coincidence ! I am in designing my own mcu controlled psu too !

      The comparator like current limit is not working for me. Oscillating like hell. I must make that opamp work linear. I mean, if the output current is greater than set current maxmum, the difference is amplified, and the output voltage is decreased with this proportional. And a 220n filter cap is needed for me to the current limit opamp output.
      With these modifications, when the current limit is reached, the output noise increased by few mV only. In the comparator mode, the output is oscillating 0.5 – 1 Vpp !

      So I am curious, waiting the next part. Piece it together, and TURN IT ON ! :D

      • http://www.eevblog.com EEVblog

        Are you using an LT3080?

        • http://lidi.uw.hu/index.php Lidi

          No, Im using PNP darlinton as a pass element. ( TIP147 ) The PNP pass transistor base was pulled down by a small npn, driven by the opa.

    • Hazard

      hi dave, my day job is designing industrial sources and rectifiers, as well as laboratory power supplies … seeing the control scheme of your design, I doubt arises about why you decided to make an error amplifier for current feedback and, in a different way, the error amplifier for voltage feedback and their outputs coupled by resistors, be better to make two identical amplifiers error, and their outputs ORed with two diodes (much similar to of the classic error amplifiers TL494), so will be much more simple to stabilize feedback loops and there is no possibility that interact both error amplifiers …

      Greetings from Argentina!

    • Hazard

      im very sorry of my very poor english, but this fault of the piece of crap of google

      • Julian

        I Think This is about Electronics, not your English. I could perfectly understand it.

        Cheers !

        • hazard

          Hi julian!, jaj, i´m hangry because, My intention was to write in an english better than mine, so use the google translator, also settled the order of the words because it was poorly written, but the damn mess them again agghh!

    • Julian

      Hello David,

      Would it be a bad idea to use a diode in series with the output of the regulator to reduce the output Voltage 0.65v?
      I am aware of the power waste at high currents, but i mean in your design.

      Regards

      -Julian

      • Huan Fai

        Hi Julian,
        If you use a diode may not be able to reduce the output voltage to the voltage you want since the diode drop is inverse to the temperature, rather than using a diode, there is easier to change the voltage control feedback loop directly connect to the output instead of the control pin……I have used that in my power supply….
        http://www.edn.com/article/478139-Servo_loop_improves_linear_regulator_efficiency.php
        I have modify the one in the article above and the one Dave used…outcome that is 80% efficiency and a very clean output voltage…

    • mmj

      Hi David,

      it just happens, that I started to follow your blog. After watching the whole series up to now, I asked myself, why don’t you separate V_control from V_in at the LT3080. If you connect V_control in front of the shunt resistor, you should get a slidely higher voltage margin and a slightly more accurate current measurement for free. The fact, that this is not possible for SOT-223 packages is barely any reason, as this package will not handle the required power for a power supply.
      Did I miss any benefits from your solution?

    • Carl Smith

      The download link does not seem to work. I get a 726kB file.

    • Carl Smith

      Download link works today. One of the reasons I watch the EEVBlog over others is that download link. I can download the file to a flash drive and plug it in my TV and watch from my couch on my 46″ screen from the comfort of my couch….

    • ToddG

      Dave – I’ve become addicted to your videos. Your site is a fantastic learning resource. I’m a total EE noob, but I thought I’d bounce some ideas around here:

      1. Open Source Learning Lab / Toolkit

      A few years ago, I tought myself how to do rudimentary machining by reading a small series of small books by Dave Gingery… they showed how to start with some clay to build a charcoal furnace…you then started casting metal parts…building skills and parts that helped you make a metal lathe. After bootstrapping the lathe, you could then proceed to build other machine shop tools…each one using the previous tooling.

      I think this approach would be awesome for EE! I don’t really know how to start this sort of thing…perhaps it already exists. If anyone is interested, I could create a github account and we could start on this. Some stepping stones could be:

      From your videos…some useful projects seem to be:

      * lab power supply
      * variable resistor box (programmable?)
      * current sink (programmable?)
      * ??? not sure what other projects would be useful, and more importantly, how to order them in order of complexity…

      For me, electronics in college seemed totally artifcial…I had no idea why I was studying these things…and super rushed too. I remember nothing from my electronics labs…That’s why practical projects with an end goal seems so useful to me.

      2. A similar idea, but even smaller would be an open sourcelibrary of small proto-circuits.

      This would be a library of small but common circuits that have pins on the bottom that you could plug into a bread board? You download the open source files, send ‘em off to a printer, and solder them up. Then, as you are proto-typing stuff, you can just take sub-circuit X, and stick it onto your breadboard…confident that it works ‘cuz you already tested it. Perhaps a set of power supplies, maybe a jtag header, dot matrix / led displays with drivers…

      Anyway, if this stuff already exists, could some one point me to it? Also, any useful books for noobs that have useful projects…I’m into building tools to make more interesting things…not making weird noise making gizmos and flashing lights for my cat to play with.

      Thanks again!
      -Todd