Connect with us

Modifying a PC Power Supply.

Discussion in 'Electronic Design' started by Arete, Apr 7, 2008.

Scroll to continue with content
  1. Arete

    Arete Guest

    Can a AT/ATX PS be modified(with components) to function as a variable
    PS (3 to 12V) and capable of handling up to 5 amps?
  2. Most will do 12V > 5A.
    Depending on what you want to use it for, pehap you could
    consider this:
    5 V
    load between 12 V and 5 V = 7 V
    load between 12 V and 3 V = 9 V
    12 V
    This gives you 5 steps, perhaps enough for your application?

    Modifying the control to the 12V goes all he way down to 3 V may not be so simple.
    Could even be easier to add a second switcher on the + 12V.
  3. Guest

    Use with LM338K (pricey, ~$8 @ Mouser/Digikey) for up to 5A, or use an
    LM317 (< $1) for up to 1.5A

  4. Eeyore

    Eeyore Guest

  5. Guest

    Yea, up to 9V then... forgot about that.

    Say... looking at the National LM117 datasheets, I'm looking at some
    neat schematics: "Adjustable 4A Regulator", but that needs a LM308
    that nobody seems to carry anymore. Any ideas for an equivalent?

    Ditto that for 2N2905 and LM195 for the "High Current Adjustable

  6. Adam S

    Adam S Guest

    I once modified an ATX supply (i.e +5V standby supply drives the main
    PWM controller) to go from 0 to 5V @ 25A. It used TL494 and I just
    modified the error amplifier section. I found TL494 based PC power
    supplies are less common these days, with the ever increasing
    power/volume density. Modifying for 0 to %100 should be possible with
    any ATX power supply because the you have the +5V standby. You need to
    somehow inject a positive signal into the voltage control loop and
    therefore adjust its output voltage range %100 to %0.

    Primary side controllers (e.g UC3842 forward converters) will use a
    TL431 on secondary as the voltage error amplifier. Secondary side
    controllers (e.g TL494) use the PWM controller's internal amplifier.

    I was gona publish a simple How-To web page until I realised there was
    no real common ATX power supply design to make my modification useful to

  7. ehsjr

    ehsjr Guest

    Nothing critical about those parts. You can use general
    purpose parts you may have on hand for them. For example,
    you could use a LM358 or LM324 op amp for the 308, a gp PNP
    like the BC327 for the 2N2905 and maybe one of the TIP NPN
    power transistors - like a TIP31C - depending on what's in your
    junk box. (Ok, the LM195 is a bit different than your typical
    NPN power transistor, but you're just looking for a workable
    substitute, here.)

    Take a look at the LM338 - its a "beefed up" LM117 that provides
    up to 5A without external pass transistors.

    When you try to build a high current linear regulator, your bigger
    problem is the "front end". The transformer will need to be rated
    about 1.8 times the current that will be drawn from the supply, so
    a 5A output means a transformer rated at 9 amps secondary or higher.
    Another factor is heat. Say your supply is designed to provide 5
    to 12 volts. That means you must provide about 15 volts DC minimum
    to provide the needed headroom for these chips. If you are drawing
    5 amps at 5 volts, you'll need to dissipate 50 watts of heat, and
    you'll need to keep the 338 temperature below 125C.

    The point is that while the schematics you find may be neat, it is
    not just a matter of soldering things together. Some design work
    beyond what you see in the schematic is needed.

Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day