FSP400-60GLN worth fixing?

[Behemot]

Re: FSP400-60GLN worth fixing?

I got a box full fo these PSUs an just found a HER103 high-efficiency rectifier as D41 in FSPXXX-60GLY(NEW2) MAIN-BOARD REV1, a 400W OEM version of this PSU. The area also looks slightly different. Seems to me they started cheaping in out lately with other revision. This one only had bulging first +5 V SB cap and it was making sounds but still starting up (no load). Recapping ATM.

[Behemot]

Re: FSP400-60GLN worth fixing?

Got some problem with stand-by rail. I've redid chanegs on it (not with the filtration output caps as they've been bugled OFC, checked their polarity 3 times so they are OK ) no change so it is something else.

Basically the rail is oscillating heavily between 3 V and 4,5 V, I can see that even on basic multimeter. If I ground PWR_ON than the PSU runs oscilated as well. All the diodes in there are OK, I think the PWM IC is good as well. I have very little voltage on the rail's optocoupler, so may it be bad? I have replaced TL431 with actually the same from another FSP supply, no change.

[andypandy]

Re: FSP400-60GLN worth fixing?

Be aware that coated copper wire comes in different colours!. When it first came out (at least 50 years ago) all of it was dark brown,and was called enameled copper wire. In the last 15 years ,better coatings have been found( higher voltage breakdown,more flexible when being wound). The colour has changed with it to a lighter brown.To anybody who does not repair electronics much,can be confused

[everell]

Re: FSP400-60GLN worth fixing?

Got some problem with stand-by rail. I've redid chanegs on it (not with the filtration output caps as they've been bugled OFC, checked their polarity 3 times so they are OK ) no change so it is something else.

Basically the rail is oscillating heavily between 3 V and 4,5 V, I can see that even on basic multimeter. If I ground PWR_ON than the PSU runs oscilated as well. All the diodes in there are OK, I think the PWM IC is good as well. I have very little voltage on the rail's optocoupler, so may it be bad? I have replaced TL431 with actually the same from another FSP supply, no change.

So what did you find?????

[Behemot]

Re: FSP400-60GLN worth fixing?

Last time I tried to push the replacement opto in, I ended installing it backwards OFC There was something like 18 V, on the other hand it was giving steady voltage, no drops

[Behemot]

Re: FSP400-60GLN worth fixing?

OK, turned it around, now it is somewhat better with 4,5 V but I still hear it is in protection mode and the PSU does not turn on anymore. So possibly some of the driving ICs is shorted overloading the stand-by rail as you suggested sometimes happens?

[Behemot]

Re: FSP400-60GLN worth fixing?

I forgot what I did last time So i removed the opto again, installed it back and same behaviour but it has slightly higher voltage, about 4,8-5 when PWR On is grounded and 5-5,20 when no load. Any idea where to look next?

ADD// just found bad zener Z41.

[Behemot]

Re: FSP400-60GLN worth fixing?

Replacing zener did not change anything.

everell, in older thread you sad that Z41 and the cap in there are for Vdd of the DM311 so they should have 9-20 V? I have unplugged the secondary coil to see if some circuitry there maybe is shorted, drawing too much power. The SB voltage is now 4,96-5 V, oscillates much less but still does, but that Vdd is only 0,15-0,4 V.

ADD// replaced both CM6800G board and PS23 board, no change.

ADD2// replaced DM311 for DM0265R. NOTHING! AARGH

ADD//3 Changed few more optocouplers, nothing, lifting J4 did not help - the problem is not in the main primary side (everything was tested or replaced). At least nothing bigger than SMD components. Gonna replace transformer.

[Behemot]

Re: FSP400-60GLN worth fixing?

YES! I noticed burned fusible resistor R42. Replaced and it went again, what? Than I checked more SMD components. D41 was shorted (must have swapped it either for bad one or it went bad somewhen during the process) while D40 was opened. Now the stand-by is rocks steady at 5,05 V.

However, the main unit was shutting down immediatelly. I found one +12 V pin was bent in the PS223 board conenctor, most likely for UVP. I have corrected that and now the unit is running! It is somewhat whinin with no load, the +3,3 V is higher. Gonna pack it all together, add some thermal greas as I rubbed it all away in time and also add varistor in there. Than I think I deserver some or maybe more like booze

[Behemot]

Re: FSP400-60GLN worth fixing?

And another one. I started doing one unit from my stockpile, another Blue Storm II 500 W. This time only +5 V SB was the problem but I went ahead and did complete recap.

The D42 diode has completelly baked everything around but it did not short itself. I replaced both D41 and D42 for ultrafasts and all the caps. The R42 resistor seemed burned but it was still showing 2 ohms so I kept not. Not sure if this was the problem or something else but I got another case of the oscilation and later a tiny bit of smoke came from it!

Now the D41 or parallel components seemed shorted but when I removed them, they all measure fine. Interesting.

[Behemot]

Re: FSP400-60GLN worth fixing?

So I dissassembled everything and bad several times, placed the D42 the wrong way several times All the components were OK but I've repalced the 2R0 resistor with 2R2 I got in hand jsut to be sure. Finally the stand-by came up and oh, I foudn the reason for the smoke. Tiny fire appered It was burning between two traces filled with carbon and remains of colophony and the <20 V were enough to set it on fire lol

Have to clean that properly and cover with instant glue or something.

[ben7]

Re: FSP400-60GLN worth fixing?

So I dissassembled everything and bad several times, placed the D42 the wrong way several times All the components were OK but I've repalced the 2R0 resistor with 2R2 I got in hand jsut to be sure. Finally the stand-by came up and oh, I foudn the reason for the smoke. Tiny fire appered It was burning between two traces filled with carbon and remains of colophony and the <20 V were enough to set it on fire lol

Have to clean that properly and cover with instant glue or something.

Ohh, I never knew of colophony. Looked it up, it's Rosin.
Yep, I've had that happen before, be sure to clean the excess flux, people!
Scared me because I thought I fried something at first. xD

-Ben

[Behemot]

Re: FSP400-60GLN worth fixing?

I cleaned it for sure but did not notice the board was burned so much between traces. I don't even know what was burning, if the board or the flux, possibly it got hot due to some short in the PCB and the colophony catched on fire because of that. Had to scratch the board between the copper, hopefully it will be OK.

Never seen this happen with proper ultrafast or even schottky diode in that position, only when they use the ordinary 1N4007 it burns so badly. The cap near it had ESR of 29 ohm and almost no capacitance.

[ben7]

Re: FSP400-60GLN worth fixing?

I cleaned it for sure but did not notice the board was burned so much between traces. I don't even know what was burning, if the board or the flux, possibly it got hot due to some short in the PCB and the colophony catched on fire because of that. Had to scratch the board between the copper, hopefully it will be OK.

Never seen this happen with proper ultrafast or even schottky diode in that position, only when they use the ordinary 1N4007 it burns so badly. The cap near it had ESR of 29 ohm and almost no capacitance.

That alternate current path from the carbonization could cause all sorts of problems. Also, too much current, reverse voltage, ambient temp, reverse recovery time, and/or leakage current could make a diode fry.

My guess is, when you installed D42 backwards, the primary was not being clamped any more (well, maybe at D42's reverse breakdown voltage), and so it caused the traces to arc over, and fried D41 because of too much reverse voltage.
Did any capacitors go BOOM? :P

-Ben

[Behemot]

Re: FSP400-60GLN worth fixing?

Nothing is bad, all diodes were OK and it worked for 30 seconds when everything wasOK. The D42 is the low-voltage diode to power driving ICs, transistors etc., the D41 is the clamp one and that was OK.

[ben7]

Re: FSP400-60GLN worth fixing?

Nothing is bad, all diodes were OK and it worked for 30 seconds when everything wasOK. The D42 is the low-voltage diode to power driving ICs, transistors etc., the D41 is the clamp one and that was OK.

What part number of diode did you use for D42? Might not have been suitable for this application.

-Ben

[Behemot]

Re: FSP400-60GLN worth fixing?

I think it was HER104? Ultrafast 1 A/400 or 600 V. While not the best choice, it is still about 1000times faster than ordinary 1N4007 diode. But as I said, for the first time, it was probably shorted and burning…than I swapped it around, soldered it a bit etc., the result is now it is all clean and plugged the right way…and I am waiting till the instant glue hardenes. As I have dropped significant amount in there and such hardenes for a day or so, nothing instant about it

[Behemot]

Re: FSP400-60GLN worth fixing?

DAMN this crap! Was burning again. I've disected it somewhat more, it is completelly carbonized to dust under the traces. Damn, I am so late with my work, two reviews should have been done already, another computer is stuborn to transfer XP from HDD to SSD so it needs complete reinstall while customer is waiting for this. Have to make another one tomorrow as he wants that on friday, I have one there which seems 100% fine, nothing is even bulging, most likely they just threw it away after seeing how good the other ones are. Will do express recap, test it for half an hour 100 % nominal and pack it for transport with case…

I am keeping this one for later (god knows when), the black shit needs to be touched with my Ferm tool and drilled away completelly, than I can work with that. I will probably just lift the traces, drill a hole in there, solder components on the traces and glue it together. Just for the record, can I use ordinary bathroom silicone?

[Behemot]

Re: FSP400-60GLN worth fixing?

OK, just learned experimentally (and by one of those 2R0 resistors - I need to buy hundred of them, got no left) that 60 V reverse is too low for D42. As there may be up to 20 V, it needs at least hundred or so. That disallows many schottky diodes I have But ultrafasts are OK, those usually start at 200 volts.

[Per Hansson]

Re: FSP400-60GLN worth fixing?

Old thread revival time
This PSU has not seen any love from me for some time.
I was anyway going to order some stuff for other projects so thought I might as well get some stuff for this old POS as well
TBH I don't even remember what I did with it last time I worked on it

But I think I'm going to replace the D41/D42 diodes, would these on Elfa be good replacements and also something decent to have on stock?
1A ultrafast: https://www.elfa.se/elfa3~se_en/elfa...0-024-80&toc=0
3A ultrafast: https://www.elfa.se/elfa3~se_en/elfa...0-005-61&toc=0
Not sure how high current rating I should be expecting here, I know someone said 2A in the thread, but maybe it was just what was on hand at the time?

I also tried to find fusible SMD resistors on Elfa but was not able to, but maybe I just don't know the eight search terms...
If you remember I incorrectly replaced the fusible resistor 2R0 at location R44 with a regular resistor which caused some nice fireworks when the DM311 chip went up in flames