SMPS 48v POE power supply - Trying to repair low voltage problem !

[amin137398]

hi.
I'm trying to repair this 48v poe power supply and I was hoping you could help me figure out the problem.
It should have 48v out put voltage but it only has 4 up to 6 volts. I tried replacing a smd capacitor and the optocoupler , these two were bad components but the problem is still there. after looking around more I found that the high voltage side is working just fine and the problem is on low voltage side. specially the feedback circuit. because when I short-circuit the output of the optocoupler for 1-2 seconds I get my 48v back! also on the input of the optocoupler  I don't get any voltage. is that normal ?
my second suspect is the shunt regulator but I'm not sure and don't have spare parts right on hand to test it.
please help me if you know what I should do next .
here are the pictures of the board :

[PKTKS]

typical symptom when the internal protection is being tripped ...

it looks like some bad component is causing a partial protection trigger..

Paul

[PKTKS]

alas... check if some 431 just has gone for a wild walk in the park...  typical fault

[amin137398]

Hi
I replaced the regulator yesterday and the voltage is now fluctuating around 3v. Is there anything else to do or I'll have to modify this whole feedback circuit ? 

[PKTKS]

Just replacing usually does not help unless we have a very clear faulty.

In these cases there are some very common faults related to
aging or failed components..

To trace them you need a deep understanding of how the particular
SMPS is doing sampling and error correction and MAINLY HOW PROTECTIONS
ARE IMPLEMENTED.

Then.. start probing w/scope these critical points..

EXAMPLE:
- if a particular supervisor is REQUIRED TO HAVE a 3.3  voltage
a PIN NAMED FB/CS  - it **MUST** be this value.
the datasheet or the board already marks that point voltage.

FB is the optocoupler mid point generated voltage and very frequently
is protected using a BIAS circuit from the auxiliar winding..

very common faults the transistor of this BIAS or the zener itself..

Another faults are the sampling resistors of the UVP protection
they are cascaded using GLUE on the board..  aging the GLUE
starts to be conductive ... and SMD fails miserably .. the component
is always good... but the GLUE is causing parasitics..

removing the GLUE fixes the UVP protection triggers...

and so on...   all the  culprits are linked to PROTECTIONS
associated with the  FB  or the CS (current sensing) pins..

EACH should be carefully inspected..

Paul