Computer SMPS has voltages from a moment and then disapear

[Francesc V.]

Hi:

I am trying to repair a 1000W cheap power supply, but so far, with no much success.

The ATX has stable 5V STBY.
When shortening PsuON to ground, it triggers all voltages for a fraction of a second and then disapear (I see the leds flashing for a moment from the ATX tester)


- I removed and tested the capacitors. They are OK.
- Removed and tested the Output rectifiers from the secondary.
- Removed and tested the Mosfets from the primary.
- Outputs are not shorted.
- Overakk review of the solder joins.


How you would try to troubleshoot the problem? Any idea how to isolate the problem? Not sure if it's the primary side or the secondary side that is faulty.
Just opening the post to gather some ideas.

I will post some pictures tomorrow.

Regards,
Francesc.

[hobostove]

What's the model of the power supply?

[Francesc V.]

It's model YL ATX-0200. You can get it in Alibaba :-)

[stj]

you got F'd
thats a traditional 250-350w psu design.

[Francesc V.]

you got F'd
thats a traditional 250-350w psu design.

Sure..........never rely on Chinese Watts :-)

I scanned the PCB, just for reference. On the secondary side, the channel supervisor IC it's a GR8313. On the primary side, the PFC IC chip it's the CM6800.


Vcc (pin13 on CM6800) has no voltage.....only the first moment it seems it has, but then the capacitor gets discharged, I guess.


Regards,
Francesc.

[CapLeaker]

In what state is the FPO pin on the supervisor? Check the main filter capacitors capacitance too.

[momaka]

Before anything else, first always try power supplies with a proper load. An old worthless Pentium 4 / Core 2 motherboard is usually good for this... though if you don't have that, a few 12V auto/car incandescent bulbs (20-40 Watt ratings) on the 12V and 5V rails are usually good enough too. The reason I suggest this is because I have seen this quite a few times where some PSUs won't start without a proper load. They simply "twitch the fans" and turn off.

So try the PSU with a proper load and let us know what happens then. If still the same symptoms, put your multimeter on the 12V rail and try powering on the PSU, with the load still connected. Does the 12V rail go up and to what value? Do the same test for the 5V, 3.3V, and -12V rails, then post the results here. We want to see if perhaps some rail is not outputting any voltage (perhaps due to a bad solder joint or open-circuit somewhere on the PCB - I've seen it happen usually at either the rectifier leads or the output toroids.

As for the PSU itself... indeed it is not a 1000W PSU, but by the looks of it, might be capable of providing 400-500 Watts if the Chinese designed it right. All in all, it's a relatively modern forward-converter design with APFC. Speaking of which, if this PSU came to you as *used*, remove and check the primary cap for proper capacitance and ESR. On PSUs with APFC, especially those that come with non-Japanese brands of primary caps and have been in use for a while, the primary cap should always be checked if you encounter issues with the PSU not wanting to power up or shutting down under higher loads.

[Francesc V.]

In what state is the FPO pin on the supervisor? Check the main filter capacitors capacitance too.

FPO pin goes high when powering on the SMPS (4.3V) and remains the same (maybe decreases some milivolts) when pressing PSON to switch on the SMPS.

Before anything else, first always try power supplies with a proper load. An old worthless Pentium 4 / Core 2 motherboard is usually good for this... though if you don't have that, a few 12V auto/car incandescent bulbs (20-40 Watt ratings) on the 12V and 5V rails are usually good enough too. The reason I suggest this is because I have seen this quite a few times where some PSUs won't start without a proper load. They simply "twitch the fans" and turn off.

So try the PSU with a proper load and let us know what happens then. If still the same symptoms, put your multimeter on the 12V rail and try powering on the PSU, with the load still connected. Does the 12V rail go up and to what value? Do the same test for the 5V, 3.3V, and -12V rails, then post the results here. We want to see if perhaps some rail is not outputting any voltage (perhaps due to a bad solder joint or open-circuit somewhere on the PCB - I've seen it happen usually at either the rectifier leads or the output toroids.

As for the PSU itself... indeed it is not a 1000W PSU, but by the looks of it, might be capable of providing 400-500 Watts if the Chinese designed it right. All in all, it's a relatively modern forward-converter design with APFC. Speaking of which, if this PSU came to you as *used*, remove and check the primary cap for proper capacitance and ESR. On PSUs with APFC, especially those that come with non-Japanese brands of primary caps and have been in use for a while, the primary cap should always be checked if you encounter issues with the PSU not wanting to power up or shutting down under higher loads.

I put loads in every rail (3.3, 5V, 12V). 40W bulb with a switch.

Momentary voltages when switching on are 4.3, 3 and 11............... so I guess that for some reason there is over-voltage in the 3.3 rail and the supervisor protects the SMPS (I measure 4.3V, then decreasing in pin 5 from GR8313)

Regards,
Francesc.

[CapLeaker]

O.k. So you are getting weird voltages on the secondary and the supervisor is shutting that PSU down. How many volts are you getting across the main filter cap in the primary during that short blip? Is the PFC working? You can disable that supervisor by grounding out the FPO pin.

[Francesc V.]

O.k. So you are getting weird voltages on the secondary and the supervisor is shutting that PSU down. How many volts are you getting across the main filter cap in the primary during that short blip? Is the PFC working? You can disable that supervisor by grounding out the FPO pin.

Thanks!!! That was the most valuable tip so far :-) I will write it somewhere........shorten FPO to turn on smps whatever it takes :-)


By grounding FPO pin to GND (shorted pins 2-3 in the supervisor) the SMPS turns on.

I had just time to measure secondary voltages before my wife started shouting, however I see secondary STABLE voltages (stable but wrong).

In the 3.3 rail I have a stable but WRONG 9.2 V
5V, -12V and 12V rails are stable and with the RIGHT voltage.

With a secondary stable voltages, I guess there is nothing wrong in the primary..........who knows.

Now I have more idea where to look at.........a faulty 431? wrong reference resistor in that 431? faulty optocoupler in the 3.3 rail?
(in the PCB I am counting up to 3 431s).



Regards,
Francesc.

[CapLeaker]

With that established, pay attention how the 3.3v, 5v and 12v is built on the secondary. Squiggle a rough schematic together. Again, make sure you are getting BOOSTED voltage from the PFC first, even so you are getting correct 12v.
Is the 3.3v made through a DC-DC converter maybe? Otherwise it doesn't make any sense to me right now how you get 9.2v out of a 3.3v rail. I couldn't find anything with that model number of the PSU.

[Francesc V.]

Finally!!!!, I discovered and fixed the problem.

I was getting a stable 370V in the primary cap, so PFC was working.

Then I started measuring the REF voltage for each of the 431 that are on the PCB. One of them had no voltage in pin1, others 2.5 stable volts.
Then following up neighbors I realized that pin1 was directly connected to a resistor that it should be 820 ohms, but I was getting K ohms............so I've replaced and auto-magically I started getting the expected 3.3V.

Joint between pin 2-3 removed and SMPS switches on normally.

Big thank you for all your feedback!!.

Regards,
Francesc.