Umm… Don’t water-cool your power supply

I saw a thing on Slashdot this morning about water-cooling your power supply. One word: Don’t.
I’ve worked inside a power supply twice–both times to replace a dead fan. One time I touched a heatsink that picked up a charge from somewhere–either a voltage regulator or a capacitor. Anyway, it really didn’t feel good. Beyond that, it made me jump.

Not a project you want to undertake if you don’t know what you’re doing. And if you do know what you’re doing, you probably already know it isn’t something you want to do be doing. Anyone who uses the word “electric” to describe something pleasant has obviously never experienced anything electric flowing through them.

I’ll pass, thanks.

If you found this post informative or helpful, please share it!

5 thoughts on “Umm… Don’t water-cool your power supply

  • January 13, 2003 at 12:21 pm
    Permalink

    I’ll pass too. I guess I’m getting old 🙂

    Switching power supplies means that you are not only working with high currents, you are working with high frequencies as well. Not a good combination.

    Anyway, the person who wrote the article must be really bored 🙂

    /Dave T.

  • January 13, 2003 at 5:35 pm
    Permalink

    As someone with experience around the innards of switching power supplies, I humbly concur with both Daves.

    And different models of switchers vary so much internally that IMPO this project absolutely should be left to the professionals, who are prepared to do the amount of checking required to ascertain whether or not a given heatsink is correctly insulated from high voltages: and for that matter whether or not it’s even worthwhile to do. Sometimes, as Dave F. discovered shockingly ;), switcher designers will cut corners on little things like insulating heatsinks electrically.

    Besides, if someone’s thinking about water-cooling a power supply because it runs “too hot”, then it probably shouldn’t even be in the case.

  • January 14, 2003 at 4:22 pm
    Permalink

    I’d just as soon not touch the inside of anything electrical.

    I got a nice 750 volt shock off a piece of gear when I was in the navy. I shorted a rheostat to the side of a CRT. Zap! It put a very real fear of electricity in me.

  • January 15, 2003 at 10:24 am
    Permalink

    Water cooling a power supply is about as brilliant as putting liquid sodium in the secondary of a nuclear reactor, when the primary has superheated steam…

    Oh wait, that’s a soviet submarine power plant design. I suppose they didn’t know what happens when sodium hits water…

    I still don’t understand why power supplies are placed at the top of the tower, with the CPU heat sink typically right below it. This is the reverse of spot cooling – it seems that it is spot heating. There’s so much room in a case, one wonders why they did that. Planned obsolescence?

    Somewhere, a hardware manufacturer is cackling hysterically…

  • January 15, 2003 at 12:58 pm
    Permalink

    Tradition. It’s always been in the top of the tower. And 10 years ago when the only fan in the case was the power supply fan, it was probably the most logical place for it. And towers are still pretty much just desktops flipped over; in 1981 that was definitely the most logical place for one since the power supply fan could get aiflow from the holes in the disk drives and the spaces between them.

    The I-Tee case I talked about elsewhere this week moves the power supply to the bottom though. Now that you mention it, that’s probably the most important thing about that case.

Comments are closed.