Testing practice for a Chinese power supply

Discussion in 'Test & Measurement Forum' started by Dodgy Geezer, Jan 18, 2019.

  1. Dodgy Geezer

    Thread Starter Well-Known Member

    Nov 30, 2009
    84
    5
    I bought a Chinese 36V 5A (presumably switch-mode) power supply from Ebay. OK, it was cheap.....

    And because it was cheap, I thought I ought to test it fairly well before connecting it to anything. But how to do this? I hadn't got anything suitable for testing which used 36V... So I connected a spare 2596 drop-down board to the supply, intending to drop the output to 12V and power a cooling fan. That had a useful voltage display on it, but, within a minute of connecting it, as I was adjusting the output voltage to 12V, the display dropped to 0V.

    That was odd. Apparently 0V coming out of the PS, but the 2596 was still displaying with its LEDs, and presumably drawing power. So I took a cheap Chinese voltmeter (are you beginning to see a trend here?) and looked at the input. The voltmeter went haywire for a bit, then registered 0V, and I assumed that the PS had given up the ghost. But when I then tested the multimeter, that seemed to have failed as well. It registered no volts on a new battery. So I had blown up a voltmeter?

    I connected the 2596 board back, and it seemed to work for a while, then drop to 0V. I thought that maybe the PS had failed in some way that produced an AC output, so I took it over to a scope and looked at the output. The scope showed a nice clean DC output. Of around 15V.

    So I took another (working) cheap voltmeter, and checked the output. This produced a reliable reading. Of 52V. And it didn't blow up.

    I am now confused. I suspect that the power supply is actually working properly, and that I am testing it incorrectly, but I can't think of what I can do to confirm or disprove this. I don't really want to connect it to my stepper motor circuit until I am reasonably sure that it is working. Has anyone got any idea what I can do to give it a decent test? Without blowing up too much test equipment?

    Failing that, can anyone explain what's been happening?
     
  2. crutschow

    Expert

    Mar 14, 2008
    22,825
    6,761
    Use an incandescent bulb as a load and monitor the output voltage.
     
  3. Dodgy Geezer

    Thread Starter Well-Known Member

    Nov 30, 2009
    84
    5
    I had wondered about that - presumably you are talking about a mains voltage light?

    Here the mains is 220-240V. I would need to match up the current draw? I can get a 100W 240V incandescent, which has a resistance of 44 ohm, but of course these change resistance considerably when they heat up, and I don't know if that's suitable. I don't have any 36v bulbs, of course...
     
  4. spinnaker

    AAC Fanatic!

    Oct 29, 2009
    7,815
    3,715

    A load is a load. You aren't looking to light the bulb just test the supply.
     
  5. Dodgy Geezer

    Thread Starter Well-Known Member

    Nov 30, 2009
    84
    5
    OK - tried that. The bulb glowed dimly. I still got 52v off the voltmeter, and 15v off the scope. Then I thought about the volt adjuster on the PS, and sure enough it was turned up full.

    So I adjusted it down to 36V (I was surprised to see it going up to 52v!) and got a 36V reading on the voltmeter. On the scope that became 10V. I tried several different probes, but ended up with the same signal - around 10v with a tiny bit of ripple on top.

    I guess that what happened was that I pumped 50+ volts into the 2596 board, which is rated at 40v. It worked for a while, and then shut down. During that process there were all sorts of spikes - which did the voltmeter in, and probably didn't do the 2596 any good either, which explains the later funny results I got.

    But I'm still wondering about the scope. I tried three probes, and got a reading of around half the volts each time. Finally, I just connected a pair of wires to the scope - and got around 36 V (or as near as I could tell with the jittering). Could all 3 of my probes be faulty....
     
  6. crutschow

    Expert

    Mar 14, 2008
    22,825
    6,761
    What is the probe attenuation factor and what is the oscilloscope input impedance set to?
     
  7. Dodgy Geezer

    Thread Starter Well-Known Member

    Nov 30, 2009
    84
    5

    If I knew what any of those things were I could probably answer you... :(

    A couple of the probes have a 1x - 10x switch on them - I was obviously using 1x. The scope is a Gould OS4020, and each input connector has a little note under it saying 1 Mohm 20pF 400V max. I don't know what effect that has on the input signal....

    Thanks for the comment about the light-bulb, by the way - that was the thing that led me to looking at the PS voltage adjustment. I thought that a 36V PS would put out 36V - + or - a few percent. This one turned out to be nearly capable of doubling its voltage...
     
  8. Dodgy Geezer

    Thread Starter Well-Known Member

    Nov 30, 2009
    84
    5
    Aha! I have just realised that I have two channels on this scope, and I had been only getting spurious voltage measurements on the one channel. I thought that which channel I used wasn't that important, since they were both set to the same settings. But I have just noticed that the Volts/Cm setting know has an unmarked inner knob, which was fully to the left on one channel, and fully to the right on the other. This must be some kind of fine scale, and having it fully over meant that one scale was practically half of the other....

    One of these days I'll learn how to use these things... :oops::oops::mad:

    So I think there is nothing wrong with the power supply. Now I just have to buy a new multi-meter and another 2596 board...
     
Loading...