Help Please, Switching Power supply shutting stright off.

[JFerris321]

Hi, hopeing someone out there might be able to help me diagnose a problem i have with an old PSU.
When i first came to test it was completely dead and upon opening it up it was clear that the few of the capacitors had swollen and show signs of starting to leak, so i decieded to recap the board and replaced them all, before reassembling i test it and all apeared to be working. However once back together the supply was dead again, so i took it apart to further test but could not find the fault and when testing it all apeared to work again. All voltages present the only issue being that the 12v is more like 11.6 but thats not uncommen on psus form this era so i put it back together only for it to be dead again.
Upon futher investigation i have discovered that 99% of the time when out the case it starts up and works fine but as soon as it is screwed back in to the case it starts up very beifly then shuts straight off again. As you can see from the images attached the circuits at the 2 screw points are only connected by the case (the yellow wire has been added by me to create this connection for testing out of the case) and it is connecting these together that causes the supply to stop working.
To put it blunt connecting the capcitors at C2 onto the negative rail stops the supply working.
Has anyone any ideas as to what could be causing this?

[tautech]

Normally the first step in SMPS repair is to identify the controller IC and download its datasheet.
Very often the Typical Application circuit is copied quite closely by PSU manufacturers.

From your pic it seems the controller is on the daughter board.

Normal failure points
VCC flyback diode
VCC cap
HVDC dropper resistor or chain of.

Controller operation can be tested with no mains supply with just a DC supply to the IC and scope its output to the switching device.
Good luck.

[Harry_22]

Hi!
Check the resistance between the yellow wire and the negative terminal of C5.

[JFerris321]

Resitance between yellow wire and c5 negative terminal starts at 5 mega ohms and then climes up as the caps in ciruit charge. Also if i disconnect the yellow wire and test the negative terminal of c5 to the ground side of c2 i get an open ciruit.

[asis]

Hi,
Look at two similar diagrams.
Maybe you will find something useful for yourself.

[JFerris321]

Thanks for the schematics but i already have worked out a rough schematic in my head of the supply. The problem is although i understand what most sections of the supply are doing as a part time hobiest i do not fully understand how switching supplies such as this one work and can not figure why the 2 EMI suppression filter caps on the live side of the circuit are causeing the supply to stop working. I could simple remove there connection to ground and by proxy the negative rail, but this would be a cop out and i still would not know that was going wrong, plus i think there would be safty concernces with doing so.
Also i can see that although HUB ST3326 is a similar sort of design (especialy on the high voltage side) it does not have a connection from the low voltage ground to earth ground.
Ignoring the fact that the low voltage side is diffrent i have drawn the connection that i am referring to onto the schmatics.
The Blue line is the perminant connection to the negative rails of both the high and low voltage side, this is always present. The yellow line is the connection i have added for testing out of the case and this stops the supply working and the red line is the connection to earth that is not present when testing out of the case, how ever when in the case both the yellow and red line are connected by design. Hope that makes sence.

[Harry_22]

Check that all four diodes are exist in the rectifier assembly.
Just to clear some doubts.

[JFerris321]

I have tested the rectifier and there is nothing wrong with that, also if the rectifier was not working correctly i do not believe it would work at all

[Harry_22]

Could you clarify the capacitance of C2, C3 and C4.
Better take a photo from the side.

[JFerris321]

The capacitors at c2,c3 and c4 are all CS472M. Datasheet says they have a capacitance of 472, how ever i believe that it is 0.472uf.
Unfortunatly it is not possable to get pictures that show much details of values as most components have there values faceing other components.

[Harry_22]

Try starting the PSU and then connecting the yellow wire.
Also measure the voltage on C8 when the unit is working and not working.

[JFerris321]

C8 is 11.4 volts when the supply is working, as soon as i connect the yellow wire the supply shuts off and the voltage drops to 6.4 volt and remains there even when the yellow wire is disconnected again. infact with the yellow wire disconected the device will not work until the main hv cap has discharged completely.

[tooki]

The capacitors at c2,c3 and c4 are all CS472M. Datasheet says they have a capacitance of 472, how ever i believe that it is 0.472uf.
Unfortunatly it is not possable to get pictures that show much details of values as most components have there values faceing other components.

That is not how ceramic cap values work. The last digit is the multiplier, and the final value is in pF. So 472 = 47 followed by 2 zeroes: 4700pF, or 4.7nF.

[Harry_22]

C8 is 11.4 volts when the supply is working, as soon as i connect the yellow wire the supply shuts off and the voltage drops to 6.4 volt and remains there even when the yellow wire is disconnected again. infact with the yellow wire disconected the device will not work until the main hv cap has discharged completely.

Very interesting case. i like it!

Your controller supply voltage is underestimated even when it is working.
At 11.5 it can already turn off.
Also the voltage doesn't return to normal during restart.
This indicates the increased consumption.

Unsolder R12, connect external power supply to C8. Gradually increase the supply voltage and measure the starting and operating current.
You can control the status via 5 volts on pin 8. It is there when the controller is running and disappears when it stops.

Please also measure the reverse process of reducing current consumption when reducing the controller supply voltage.

[asis]

Hi,

@JFerris321
Important point.
Do you carry out all your manipulations with a connected load?
At least some.
Especially at the voltage output of the channel to which the optocoupler with the REF circuit is connected.

[JFerris321]

Testing has been done both with and with out a load attached, it does not seem to make an difference to the results.

[asis]

You said you replaced some ECAP's.
The most vulnerable is ECAP C8 (10mF x 25-50V?).
Check C8 & C9 carefully by disassembling them.
Or replace it.

[JFerris321]

C8 (10uf 50v) and C9 (0.68uf 50v) have both been replaced already, C8 was dried up and measured out of spec, C9 was okay but changed it anyway just in case. Have checked the new caps they seem fine, also tried the original C9 back in the supply it made no diffrence.

[tautech]

C8 (10uf 50v) and C9 (0.68uf 50v) have both been replaced already, C8 was dried up and measured out of spec, C9 was okay but changed it anyway just in case. Have checked the new caps they seem fine, also tried the original C9 back in the supply it made no diffrence.

C8 is on controller VCC rail, right ?
Sounds like the controller is kicking off but dropping out due to its UVLOCKOUT.
Than means VCC supply is not fully working and the flyback diode would then be your first suspect.

[Poroit]

Looking at your original post where you said the PSU was working out of the case at 11.6V.
Did you try adjusting the trimpot to see if it increased the output?

Also...when you put it back in the case and power it up...does it trip your home/workshop Earth Leakage Breaker

[JFerris321]

Looking at your original post where you said the PSU was working out of the case at 11.6V.
Did you try adjusting the trimpot to see if it increased the output?

Also...when you put it back in the case and power it up...does it trip your home/workshop Earth Leakage Breaker

Im afriad it has no trim pot and no it does not trip the RCD.


In regards to the diodes, i have tested all the diodes on the supply with my multimeter in diode test mode and they all seeem around the figure i would expect.

[JFerris321]

Out of intrest do anyone have any adea would i should expect to happen if the capacitor at c8 (i presume this is the vcc supply but not sure as i can not identify the chip, markings are 3882 with 701 under that) was changed from 10uf to 12uf?

[SeanB]

Have you tried resoldering the transformer pins, and the pins of the secondary rectifiers. They look suspiciously like some are dry jointed, which will do that.

[tautech]

Out of intrest do anyone have any adea would i should expect to happen if the capacitor at c8 (i presume this is the vcc supply but not sure as i can not identify the chip, markings are 3882 with 701 under that) was changed from 10uf to 12uf?

Short form datasheet:
https://www.mouser.com/datasheet/2/308/KA3882C-1295487.pdf

[asis]

Which chip is designated U3 and what is the orange wire next to it?
This TACHO fan?
I don't see the REF chip (TL431), or is it in the module next to Q8?
Q2,Q3,Q8,D6 types?
It is necessary to evaluate the condition of the thermal insulating gasket and look at the type of dual diode of the +5V channel.
We need a picture of the PWM module U1.
In principle, I can bring the scheme to the absolute level if you help me with determining the denominations