Dear members,
I have this PSU which was used since 2010 without problems. Recently I started getting random sudden shutdown. I narrowed it to the PSU.
The unit itself has CM6800G on the hot side on a daughter board. Protections done by PS229 and WT7518D on the cold side on another daughter board which also has fan control.

The psu works fine for 10-15 min at load but shuts down either under load or as soon as it goes to idle, sometimes even without any load just shut down at idle. All caps on main board measured within specs (on and off board measured with ESR). I changed the caps on the protection board but they where all good just to bee sure. Still the same with shutdowns. Added a resistor in parallel to resistor in series with PGI of PS299 and raised the voltage from 1.16v to 1.32 volts. Still the same behavior. Any ideas where to look?

The build of the PS is exactly the same as corsair cx500, VX550w.

Are you sure of the quality of the contacts in the wall outlet and the power cables?
Try check PSU with nominal resistive loads corresponded value and power without motherboard.

replace p-n-p transistors in magnetic stabilizers of 3.3 volts and 5 volts

Try to to change optocouplers, may be it helps

@ mia
No problems with outlets or cables. I forgot to mention earlier that i tried with another power supply and everything worked fine.

@apach71
Which transistors u mean. Did u experience same problem?

May be problems in standby PS

he supervisor that does not like change PS229 on WT7510-TPS3510

will order the supervisor ic and report after changing

Got new ps229 and changed, still the same.
As for the WT ic the one in the board is wt7518d and not wt7510, still waiting for the new ic to arrive.

Why don't you say what changes have occurred hardware/software in your system, after which you have fallen into the present situation? Maybe there is an overload of one power supply, and the other is not. Compare the characteristics of power supplies.

Neither software nor hardware changes made recently. As i mentioned i tried the system with two other PSU, one 500W and the other 480W and both work very well even under torture load.
Ther defective supply when attached to same machine or Another one behave erratic with sudden shutdowns.

Why don't you monitor in time the voltage at the output of the power supply, including the standby voltage 5V stb ? Perhaps measurements will help you locate the cause of this behavior of the power supply.

Monitored the 5v stb,+3.3v, +5v and 12v.
4 cycles of random shutdowns, two at load and 2 at idle. Nothing strange with those voltages just zeroing suddenly at shutdown though solid 5v stb at 5.18 all the time.

The +5v at 5.19 - 5.20 at load - idle.
+12v at 12.17 - 12.11 att load - idle.
3.3v at 3.35-3.42 att idle - load.
Loaded using OCCT.

You use software tools to test the power supply. But they are indirect and cannot nominally load each of the voltage channels. The most correct, for example, is the load of each channel in the form of parallel-connected resistors of corresponding resistance and power.

The measurements were done using 4 Fluke multimeters connected to powersupply violet, red, orange and yellow, COMs parallelled with a black wires during operation. The loading done through software while the pc was on. I sure can load with parallell resistor too byut do u Think there will be any difference?
Same time it would be a mess to load all line at same time or should i load each line at a time.

I always check the power supply under load according to the specifications on the PSU cover label. In this case, the connectors of all channels are loaded simultaneously through separate resistive assemblies of the corresponding resistances and power. Never software can not achieve the load of the power supply according to the specification.
I hope that you have previously cleaned the power supply unit with a fan from dust, updated the lubricant of the fan, tightened the fastening of components to the radiators, carefully visually checked the soldering of printed circuit boards for the absence of microscopic cracks, checked the state of the contacts of the connectors.

Couldn't get hold of so many Power resistors to test all outputs at same time. Testing each powerline by itself show 12v line working fine up to 10 amps. 3.3 v and 5v lines can't stand load and the device turn off. It might be wrong to test each line by itself? or does this tells the 3.3 and 5v line are the problem?

In most cases, for normal operation of the power supply, the + 5V channel must always be loaded.
The label on the power supply case shows the maximum permissible combined load, for example + 5V and + 12V or + 5V + 12V + 3.3V, etc. When checking 3.3V, be sure to connect the load to the contact with the orange and brown wires, the voltage across the load is measured through the brown wire.