Don't Miss

Strange Power Supply Tester Result

By on February 27, 2021
atx power supply tester

Yesterday my desktop power computer user friend brought over 2 of his now defect ‘bigger’ power supply units for repair. One was 650W, and the other one was 850W. I tested both smps units with my special made 75Watt safety bulb power box placed in between the 230VAC power line input to both supplies.

The 650W desktop supply worked at first but the 3.3V line was dead. And I noticed a bit of smoke coming from the middle of the board. I found why this unit was the cause of blowing out his main power line fuse at his home. He had placed the input switch to the 115VAC setting which was totally wrong of course! And I only had tested this unit when the switch was placed back to the correct 230VAC position.

how to fix atx power supply

Because I noticed a bit of smoke coming out of the case I took the power board out the case and checked the primary components for short circuits. And I checked the 4 pin PC817 compatible optocoupler, both primary HV transistors (T13009S), and the diode bridge. All checked out fine!

And none of the primary transformers failed when I checked them with the Blu Ring tester. Also both primary e-caps checked out fine! As also all diodes seemed to pass all checks. And none of the other smaller transistors had any short either when checked in the diode beep test of my DMM. Afterwards I soldered back the tested components I had taken out to test them with my Peak Atlas DCA75 semiconductor tester, like the primary cooler plate with both HV T13009S transistors on it (one had an Hfe of 21 and the other just about 18). I also tested the optocoupler which I tested with the small digital tester I also had used and shown in my previous repair. And a few other parts I checked before retesting the power supply that previously showed some smoke and light bulb

problems. But this was when tested with my second PC desktop output voltage tester that also had five 15 Watt load resistors on all voltages to resemble a connected PC mainboard. [ * I will come to this later on!]

Because it was fair to assume that his main power line fuse had protected his power supply from getting any real damage. I therefore started testing the output of this 650W Power supply with my first ‘black cased’ output tester (that only has a smaller internal load). And it showed that all voltages were there without any problem. And my 75W light bulb didn’t show any short currents either.

atx power supply fix

Above both primary npn HV transistors, and the PC817C that were fine! Below my first Power supply tester that approved all outputs, and the safety light bulb didn’t show any short current issues!

atx power supply tester china

atx tester

atx psu tester

Above photos showed on the right my special safety test Light bulb placed in a metal case of an old defect desktop power supply. On the left the perfectly working 650W desktop power supply when tested with above shown first Power Supply Tester.

So it looks like that this Power supply is just fine.

* And here I come to my conclusion about the testers used! Here apparently the first above shown Power Supply tester is wrong and not at all trustworthy!

Because when I tested the same 650W unit that tested good with the first tester again now with my second Power supply tester with several 15W resistor loads on it, I got completely different results! Now my light bulb is practically completely lit up! And not all voltages are there now! Next photo shows my lit up safety light bulb and the photo after that shows the second output tester that reveals a big problem!

light bulb test atx power supply

Next photo shows that second Power supply tester with the fewer active output voltages.

Now the previous good tested 3.3V output is gone and a real problem! So there must be some shorting circuit problem in the 3.3V circuit somehow.

resistor psu tester

If there is a problem in one of the non replaceable parts of the supply, like a partly bad transformer, then fixing this power supply is too expensive and all parts are only good for spare parts in another repair.

So don’t get fooled by the good results when using the above ‘black cased’ Power Supply Tester without the external 15 Watt resistor loads! To complete this article I show a few photos with the IC’s that were soldered in this unit.

controller unit psu

Above on the left the special SD6863 IC. On the right the AC input selector switch that was wrong placed into the 115VAC position. On next photo the 16pin DIL UTC 7500 controller is shown.

fixing atx psu

On next photo on the left the UTC339 is viewed. On the right photo the optocoupler and the SD6863.

not accurate atx power supply tester

Below on the left the 10D241k VDR that apparently caused the smoke before it blew up with BANG!

vdr burnt

It was removed from the now empty spot in between both 560uF 200V e-caps in the previous photo on the right. The red wire on the right of the diode bridge is one of the wires from the AC input voltage selector in that photo. So the VDR was definitely damaged by the wrong 115VAC switch position it was set in before. Next photo shows the 3 HF transformers of this 650W power supply.

how to fix an atx power supply

All 3 transformers look a bit brownish on the outside yellow tape.

brownish transformer

Maybe their transformer coils collapse only at higher current energy levels? Like when tested with 15 Watt loads on the output voltages? Next photo shows the solder board side. The red circle marks the spot where the now bad VDR is removed.

atx board

atx board repair

Both smps desktop units had some sort of short current issues, because the 75W 230V AC light bulb lit up in both cases but only when the second Power Supply Tester with the five 15W loads was used. However both the primary fuses were still intact! And the cause of the problem is not yet found and maybe also unfixable if it involves parts like coils or the special HF transformers that normally can’t easily be replaced. The transformers tested fine at their primary sides, but it still may be that the secondary coils are damaged and they are normally low impedance. The VDR blew later on with a loud big bang after I kept the power supply running while it was connected to the second Supply tester with the 15Watt load resistors. While the 75 Watt bulb kept indicating something was still wrong.
Maybe replacing the special SD6863 chip or the UTC 7500 controller will help but from previous experience I guess that replacing the 7500 controller probably will not make any difference here!

The 8pin dil SD6863 primary controller is according to its datasheet : “” , a max 11 Amp Drain Current Pulse current mode PWM + PFM controller with built in HV mosfet driver with a max constant current of 3A at max 28V. (10 of them cost about 1,82 euro incl. Shipping). I ordered those – just in case I may need them in future -but I doubt it will help to salvage this unit in the present condition it is in.

Conclusion : lesson learned here is… better use more different Power Supply Testers to rule out unexpected major shorting circuit currents! Especially use a tester with decent load resistors on the voltage outputs. It will certainly save you time not having to remove the afterwards still failing power supply from your computer again.

albert from netherlands

Albert van Bemmelen, Weert, The Netherlands.

Please give a support by clicking on the social buttons below. Your feedback on the post is welcome. Please leave it in the comments.

P.S- If you enjoyed reading this, click here to subscribe to my blog (free subscription). That way, you’ll never miss a post. You can also forward this website link to your friends and colleagues-thanks!

Note: You can read his previous repair article in the below link:



  1. Waleed Rishmawi

    February 27, 2021 at 1:57 pm

    Very informative article. Thanks for sharing. On the other hand I always stay away from such power supplies because they are complicated to fix and very cheap to buy. Have a blessed day.

  2. Mark

    February 27, 2021 at 5:12 pm

    Well done Albert.
    Shows the importance of testing circuits under a working load. Some of those testers don't put enough load on the circuit and give dodgy results.
    Thanks for the article.

    • Albert van Bemmelen

      February 28, 2021 at 12:31 am

      Indeed Mark. Sadly too many electronic test tools are still sold that after a closer examination fail in being useful. We probably often only detect this after using different testers on the same task.

  3. Albert van Bemmelen

    February 27, 2021 at 5:12 pm

    They indeed are rather complicated to fix Waleed. But these higher power units are not exactly cheap and 850 Watt units still cost around 142 Euro exclusive any additional EU import fees! So fixing these would certainly mean saving some money!

    • Albert van Bemmelen

      February 28, 2021 at 2:23 am

      PS: I later on also tried to find the culprit in the 650 Watt power supply by using my Variac which I slowly turned from 0 VAC to 220 VAC without the safety lightbulb inbetween. Hoping to find which component would overheat or show signs of stress. Sadly the 560 uF 200V DC e-cap on the left got cooked and burst open and I noticed that the second 10D241k VDR was a very low resistance (about 7 ohm). But even after I removed that second bad VDR and had replaced the 560uF 200V e-cap in the primary circuit, the 650 Watt power supply still had the same problem as before. So likely the main transformer is the culprit because that the 3.3V is still the problem at the large current of 38 Amps this desktop supply should give (and 5V at 45A. But that voltage is not the problem here)!

    • Waleed Rishmawi

      March 1, 2021 at 3:09 am

      Albert: they might be expensive there but here they are kind of cheap for such devices. They go around 80 dollars but the parts are hard to find and if I find them it would be too expensive and the customer would prefer to buy a new one. Have a blessed day.

  4. Parasuraman S

    February 27, 2021 at 11:25 pm

    Very informative and lengthy article packed with tips and experiences which are of unique in nature. A blessing for technicians like me! Many thanks for taking so much time and efforts to write the article! Well done!

  5. Lee

    February 28, 2021 at 6:41 am

    Nice one Albert.
    Where can you buy the outboard tester with big resistors from.

    • Albert van Bemmelen

      February 28, 2021 at 1:38 pm

      I bought my second PC desktop power supply tester with the 5 resistor loads a few years ago and couldn't find any seller anymore that still sells these now. Not on eBay and neither on Aliexpress or Banggood. Sorry Lee, if I do find one I'll let you know in a new comment here. But maybe you still could make one from websites like this one?:

    • Albert van Bemmelen

      February 28, 2021 at 2:19 pm

      PS: Here we can find an example of schematic of a power supply circuit with a 115VAC/230VAC input selector switch, like the one examined in the article:

  6. Lynn Blakely

    February 28, 2021 at 11:49 am

    Sir Albert: Thanks for a very informative report. Where can we purchase the tester with the 15watt load resistors in it.Thanks for the tip about the black tester, that gave a false result.

    • Albert van Bemmelen

      February 28, 2021 at 7:57 pm

      Sorry Lynn, apparently the 15W load Power supply tester board no longer can be bought. See my post to Lee that shows the seller that previously had those for sale. Maybe you can build one from the weblink I gave in the comments?

  7. Albert van Bemmelen

    February 28, 2021 at 1:47 pm

    Found one Lee! See: But they sadly sold out on them.

    • Albert van Bemmelen

      February 28, 2021 at 2:05 pm

      Here the schematic of the DIY PC desktop Power supply tester on the mentioned webpage that uses 9 (!) 10 Watt power resistors as load:
      That one even may be better than the power supply tester I used with the only five 15 Watt resistor loads? After all 9 x 10 Watt is more than 5 times 15W? Anyway, good luck in building or finding one Lee!

  8. Yogesh Panchal

    March 1, 2021 at 3:55 pm

    Good attempt!
    I am also working on one SMPS which gives loud Chick....Sound with visible arc/Spark between Driver TX and Main switching TX on primary side checked each components out of the Circuit on primary side didn't find any bad component having doubt about PWM 8 pin IC Marked with JD3313D and Controller 16 pin IC on secondary side M13M385 but not finding any data or detail on the web.may use parts for some another repair.

    • Albert van Bemmelen

      March 1, 2021 at 7:31 pm

      I am trying to improve my successrate in fixing the more troublesome high power supplies. And hoping to at least find the reason when they do not work if they fail. And now knowing which supply tester we can trust will hopefully help. I can't tell you what kind of chips you have that you mentioned but recently noticed that a UTC7500 (in this article) and the TL494 are completely pincompatible. Jestine confirmed that they may be (completely) interchangeable.

      • Albert van Bemmelen

        March 2, 2021 at 4:07 am

        But Yogesh look at this IC on Aliexpress with top marking 31AMS13 which in fact is just a CD/HEF4001 cmos chip. So maybe your marking is not the right chip marking?

  9. A van Bemmelen

    April 19, 2021 at 6:37 pm

    PS: Later on replacing the 8pin dil SD6863 primary controller didn't make any difference in helping to fix the strange smsp problem, as I already suspected!


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.