Announcement

Collapse
No announcement yet.

Building a load box for ATX PSU fault finding

Collapse
X
Collapse
 
  • Page of 1
  • Filter
  • Time
  • Show
Clear All
new posts
Previous template Next
    #1

    Building a load box for ATX PSU fault finding

    I'm gonna build a load box for testing ATX PSUs.

    This, because if you test it when connected to a motherboard, the motherboard will probably create a bit of interference.
    Besides, if the computer won't turn on, it'd be good to test the PSU in isolation. Would be a time saver in many cases.

    So I had a look at an old AOpen PSU I have around. Its voltages are:
    3.3V 28A
    5V 30A
    12V 15A
    5VSB 2A
    -5V .3A
    -12V .8A

    What I'm after is an average load on all rails.

    Let me just see if I got the equation right:
    U=R*I
    U/R=I
    3.3V / 0.5 ohm = 6.6A.

    Is this correct? Isn't 0.5ohm dangerously close to a short circuit??
    Tags: None
    #2
    Re: Building a load box for ATX PSU fault finding

    For a computer system, no, 0.5 ohms isn't a short. A computer is a low voltage, high current device - why else do you think there are so many paralleled wires in the ATX connector?

    That's why some techs find chipset measurements confusing - it is entirely normal for a power rail to a chipset to read a few ohms without the chipset being shorted.
    Originally posted by PeteS in CA
    Remember that by the time consequences of a short-sighted decision are experienced, the idiot who made the bad decision may have already been promoted or moved on to a better job at another company.
    A working TV? How boring!

    Comment

      #3
      Re: Building a load box for ATX PSU fault finding

      While I appreciate the inquisitive nature of loading a PSU to pinpoint problems, usually it's easier to just swap in a known working PSU. Odds are you won't save any time, the time (and often cost) to design/order/build/hookup the tester will swamp the time it takes to just plug in a different PSU.

      There are other issues with load testers which can mislead. For example a PSU might maintain a certain average voltage but with too much ripple, or have too poor a recovery time to deal with the highly dynamic load which modern PCs present. There's also cross loading issues, you can test a PSU based on some average, peak, or spec'd load on it's label but that load considering the more important 12V, 5V, and 3.3V rails, tends to be a fairly large % different than the load presented by any random PC at startup.

      Comment

        #4
        Re: Building a load box for ATX PSU fault finding

        I have a couple computers lying in the repair pile. Both of them have motherboards that blew up. I've tested their PSUs on other computers and they seem fine.

        I've ordered new motherboards for these computers, but I'd like to do an extra check on the PSU before returning the repaired hardware to my friends/family.

        It seems that the only problem these 2 machines had was in their motherboard, still I'm concerned that their PSUs may be delivering power that is just dirty enough to wear out components over time.

        Hence, I'd like to see how much ripple is on the PSU output, preferably with no MB attached.

        Isn't this is a good idea? Perhaps it's better to use power transistors for resistance? Maybe cheaper and more flexible?

        Comment

          #5
          Re: Building a load box for ATX PSU fault finding

          A big PSU load tester (usable for anything not just ATX PSUs) has been on my project list for a long time. I even got the heatsink for the power transistors. However, it's going to be a pretty big device that i have no need for atm, so it keeps getting delayed.

          The easiest solution is to use a bunch of resistors. It's the least accurate, but for testing the odd PSU you won't need anything more.
          Originally posted by PeteS in CA
          Remember that by the time consequences of a short-sighted decision are experienced, the idiot who made the bad decision may have already been promoted or moved on to a better job at another company.
          A working TV? How boring!

          Comment

            #6
            Re: Building a load box for ATX PSU fault finding

            yeah, resistors are certainly easier than transistors.

            How much power do you recon I have to burn off to reveal common PSU problems? say on a 30A 3.3V rail, am I likely to see the problems by burning 10A, or need I get closer to the max current?

            Comment

              #7
              Re: Building a load box for ATX PSU fault finding

              Take a look here, this is the 80 Plus power supply testing procedure: On the last page an example is given.
              Originally posted by PeteS in CA
              Remember that by the time consequences of a short-sighted decision are experienced, the idiot who made the bad decision may have already been promoted or moved on to a better job at another company.
              A working TV? How boring!

              Comment

                #8
                Re: Building a load box for ATX PSU fault finding

                according to that document one should test at 10%, 50% and 100% of each rails rated power. But that's for determining PSU efficiency. Is this really neccessary for fault finding?

                Comment

                  #9
                  Re: Building a load box for ATX PSU fault finding

                  Fault finding: Test at 100% and if something is faulty it'll blow. Or if the power supply came from a computer and will go back in said computer, first measure the consumption of the computer at the wall, divide by efficiency (assume 70% for crap PSUs, 80% for decent ones), add 20-30% and there you have the load you're supposed to test that PSU at.
                  Originally posted by PeteS in CA
                  Remember that by the time consequences of a short-sighted decision are experienced, the idiot who made the bad decision may have already been promoted or moved on to a better job at another company.
                  A working TV? How boring!

                  Comment

                    #10
                    Re: Building a load box for ATX PSU fault finding

                    A lot depends on what and how much you want to test. If you want to see if a P/S is at least functional, ~10% load on all O/Ps and some LEDs driven by window detectors could tell you if the P/S functions and regulates. For ripple testing, you need to check, at the least, at ~10% load on all O/Ps and at ~100% load on all O/Ps (you also need to BW limit your scope to 20MHz and make sure that what you see is real, not noise picked up by the antenna you think is your scope probe ground lead).

                    An active electronic load is quite a project, and some P/S have 6 O/Ps to be loaded. Folks like Chroma don't charge kilobucks for easily designed cheap crap.
                    PeteS in CA

                    Power Supplies should be boring: No loud noises, no bright flashes, and no bad smells.
                    ****************************
                    To kill personal responsibility, initiative or success, punish it by taxing it. To encourage irresponsibility, improvidence, dependence and failure, reward it by subsidizing it.
                    ****************************

                    Comment

                    Previous template Next
                    Working...
                    X