Resistor Value.

[HellasTechn]

To keep this low voltage / cold supply side, still all isolated and safe, get a 9V transistor type battery, and use 2 flying wire leads, to connect battery + to a 470 ohm 1/2 W resistor and a lead to the other - terminal of the battery.
Your free end of the 470 resistor is going to connect to the pad of old R18 that went over to the 817 optical isolator.
Your batteries free flying lead will be held in hand to be temporarily touched to the negative power output terminal of the power supply.
If all is well in the 817 and the HOT side of the power supplies regulatory circuitry that that 817 phototransistor feeds, there will be a decrease in power supply output voltage, that you had read before, while holding that connection, just after powering up, for a short time needed to monitor the output voltage again.
This time it could be down to 9? 7? volts, but rise just as you lift the wire for the battery supplied LED power.

Should i run this test now that the prev one faield ?

 

[(*steve*)]

I forgot to mention that while powered on the PSU makes a loud clicking noise.

That is generally a good indication that a power supply is overloaded.

Now with the resistor removed the DMM could not give a steady reading

Let's assume it's part of the feedback. When you remove it, the power supply keeps running because it thinks the output voltage is low.

Withing 5 seconds or so the output capasitors get very very hot and if powered on for about 10 sec they start to pop.

I've accidentally disabled the feedback on an amps and the caps vented. However all was silent and I'm pretty sure it took less than 10 seconds.

The clicking and the capacitors getting hot make me think about the rectifiers just before those caps. Are they ok?

It's possible that damage to something in the feedback circuit could lead to a very high output voltage which could in turn lead to failure of the rectifiers and then the capacitors. Even if the cascade didn't occur at the time of the fault, it may have happened later.

 

[(*steve*)]

I kinda like the idea of replacing that option, but breaking one connection on the led side and keeping it turned on with a 9V battery and an appropriate resistor. If all is well, putting on the power supply should cause it to be completely inactive. It should think the output is at or above the required voltage.

The next thing I would do is wire a red led in place of the opto led (easy if you still have the input of it isolated). And with the power OFF, inject power into the output of the power supply.

Firstly, (at a low voltage) the current should be low. If it's high, those rectifiers are again very suspect.

As you increase the voltage, the led should turn on at the output voltage of the supply you're testing. At this point I'd be expecting a current of no more than 50mA.

Increasing the voltage still more (but no more than the voltage ratings of the calls, diodes, or tl431 (whatever the weakest point is) should result in slightly increasing current and the led staying on.

This test should indicate if the other half of the feedback circuit is working.

 

[(*steve*)]

And I would think R18 should be a couple of hundred ohms if it's the one connected to the anode of the optocoupler.

 

[HellasTechn]

The clicking and the capacitors getting hot make me think about the rectifiers just before those caps. Are they ok?

If you mean the one (dual diode) mounted on the chassis then yes it is good. I have replaced it because the one originally on was shorted. Anyway right now it is good.

It's possible that damage to something in the feedback circuit could lead to a very high output voltage which could in turn lead to failure of the rectifiers and then the capacitors. Even if the cascade didn't occur at the time of the fault, it may have happened later.

I have checked all components i can see. all but the resistors around the tl431...

 

[HellasTechn]

The next thing I would do is wire a red led in place of the opto led (easy if you still have the input of it isolated). And with the power OFF, inject power into the output of the power supply.

If i understand correctly, i should remove the opto and solder the led biased just like the led inside the opto. then give 12V from my bench psu to the output to see if the led will glow.
If it does then that probably means that feedback circuit is ok. yes ?

I will replace R18 with a resistor probably like 480 Ohm or 560 Ohm like shown in the schematic in post #11

 

[(*steve*)]

If it does then that probably means that feedback circuit is ok. yes ?

I would use a variable power supply to slowly crank the voltage up, but yes, that's the idea.

That test will check the low voltage side of the feedback circuit.

If the rectifiers are ok and this test doesn't draw excessive current (which may point to some other fault like shorted caps) then any clicking sound when the smps operates must be caused by an overload somewhere else. A shorted turn in the transformer would have that effect, but then your output voltage would be low or zero.

I would also try the test of the high voltage side of the feedback circuit (externally powering the opto) and ensuring this totally disables the smps. This checks that the low voltage side of the feedback circuit can actually control the high voltage side.

 

[HellasTechn]

With a 560 Ohm resistor the led glows at 12.3V. That is normal i guess.
And the current drawn from my bench psu is 65ma. Funny thing is that when i connect the psu leads on the output of the psu board its sparking. probably due to the reverse biased diode on the output i think.

By the way what purpose does it serve ?

 

[HellasTechn]

And now thanks to you this case is solved !
I discovered a bad resistor (r12) as well as the opto itself was bad. R12 is connected to one lead of the opto's internal transistor.
I also now have an idea about how the feedback circuit works !

THANK YOU.

P.S.
I can now start troubleshooting the second bad psu i have that is identical to this one.

 

[(*steve*)]

With a 560 Ohm resistor the led glows at 12.3V. That is normal i guess.

Yep, that sounds fine

And the current drawn from my bench psu is 65ma.

That's a bit higher than I would have expected, but probably fine.

Funny thing is that when i connect the psu leads on the output of the psu board its sparking. probably due to the reverse biased diode on the output i think.

No, I think it's due to the low ESR capacitors that will look like an almost short circuit when discharged.

That reverse biased diode prevents a reverse polarity voltage appearing across the power supply. This could happen if two power supplies are connected in series and this one is not powered on.

And now thanks to you this case is solved !

Excellent. And good luck with the other power supply.