Two dead ATX SMPS's - TL494 and FSP3528 based

[domas]

Re: Two dead ATX SMPS's - TL494 and FSP3528 based

Thank you momaka!

I will consider building a little bit more simple light bulb current limiting device if i will continue to work on shorted primary sides. now both of them are okay there (not shortened at least)

Regarding the first FSP psu,you were totally right. 2 dioded where shortened. It took me some time to find them as for some reason I was very sure that it should be some heatsinked to-22 design rectifier. Apparently for 0,5amps of current for -12V line it is enough to have 2 simple diodes. And this would kill my transistor.. Funny enough as OC protection should have kicked in, but it did not and would kill the one of primary BJTs. Same way as OC protection would not kick in as i overloaded it for the first time. Strange. Both of my atx do not look too bad - at least they have a lot of components, in comparison to some we all have seen in hall of shame thread.

Regarding the second supply, the Q-tek, i have been through the mentioned components already. There were 2 resistors that were fried after explosions, as far as I remember 2k7. I couldn't recognize them as the were nicely black, but i have taken them from the same place in another working FSP supply (it was the donating components for 2 killed supplies.) at the same time i have taken primary transistors from it. There were some diodes that were black but still tested to be OK, but i have changed them anyway. Now I am thinking to re-test the caps you mentioned, as I have only a DMM, but I have to read about that a bit. I am also concerned about my transistors... It seems I can't distinguish between BPJ transistor and a mosfet. Have to read some more. Maybe I have put wrong ones in the PS.

kudos to momaka

[momaka]

Re: Two dead ATX SMPS's - TL494 and FSP3528 based

I think you may want to start with building a current limiting device for the AC side. 40W to 100W incandescent lightbulb in series with the live AC line on the PSU is one easy way to do that. This is how to wire it up:
https://www.badcaps.net/forum/showpo...4&postcount=70

Here is what budm's device looks like:
http://s807.photobucket.com/user/bud...82195675176907

That way, if something goes wrong when you troubleshoot the PSUs, the light bulb will turn on and stay lit.

Now as far as the actual problems:
On the first PSU, check the -12V rectifier diode(s). Mostly likely they became shorted when you pulled 4A of current through them.

For the Q-tec PSU, check the base-drive circuitry for the BJTs on the primary side. Should be 2 diodes, 2 electrolytic caps (10uF, I believe), and a few resistors.

[domas]

Re: Two dead ATX SMPS's - TL494 and FSP3528 based

And some pictures of Q-tek 350W supply, the one i killed by trying to make an adjustable

[domas]

Re: Two dead ATX SMPS's - TL494 and FSP3528 based

So i have taken some pictures of my supplies. Not sure if this will help you to give me some advise, but it makes the thread more interesting to look at.

I know my work is a little bit messy. Still learning. How do you guys clean up the black residuals after soldering? I guess it is some soldering paste that stays there. I clean it off mechanically between the tracks on pcb, but is there an easy way to clean the whole pcb off? It looks sooo messy with this crap around the connections where i have soldered.

The one next to heatsink screw on primary side is getting killed everytime i put a new one in.

I have attached pictures of FSP supply in this post.

[domas]

Hello folks, my first post here.

I have killed 2 SMPS recently. Both of them are totally dead except from 5vSB. No other outputs are working.

First one:
FSP ATX-250PA, controlled by FSP3528 & DM 311 for 5vsb, fsp chip is described here: https://www.badcaps.net/forum/showthread.php?t=22171 Killed it by overloading -12V output. I charged a 12V li-on battery pack and i needed more than 12V - i used -12V rail and +3.3V rail. Was going at about 0.2-0.3Amps. Tried -12V + +5V rail as well. Started charging at 4 amps (while my -12V line is rated at 0.5amps) smoked after 4 secs. Stupid me.
Everything seemed fine inside, no signs of dead components. Found out that Q2 (please refer to generic schematic of HV side of SMPS here: http://www.geocities.ws/luizissao/im...DO_ATX-250.gif ) d304x mosfet was dead/shortened. Strangely no other components were killed, tested every diode and transistor on the circuit. After changing it with FJP13009 () the supply not recover. Only 5vSB was alive.

After testing the FJP13009 i found out that it got shortened as well, this time just over Base-Emitter junction. It did not make any sound or smoke, had pcb in front of me. After changing FJP mosfet with another one (have ordered 5 ) got the same result. Another brand new FJP mosfet dead. Then I changed both mosfets on primary side - the good d304x as well. I thought it was the reason, as I assumed that they have to be identical. Same result, q2 shortens on Base-emitter. q1 intact. I am soon running out of mosfet's.

Do you guys have any idea which component might be a reason for this?

Before the stupid mistake while charging battery it was working fine. I am using it as bench power supply.

NEXT: Q-TEC PSU ATX 350W LOW NOISE.

Used to be a nice atx as it is the only one i have that is controlled by TL494CN (pwm drver) & LM339n (comparator for protections)

killed it while trying to make an adjustable PS. by this hack: http://boginjr.com/electronics/lv/atx-mod/
I was nicely adjusting way from ~4-6v (can't remember) (on 12v rail) to 12.00V. It would simply shutdown at 12V, as protections had to be disabled. I had to remove the correct zener diode to eliminate the OV protection but to keep OC protection. I would love to crank this atx up untill 15V (and maybe even more if i decide to change the caps to more than 16V). After removing one diode it would let me go over 12V, but mains fuse exploded at 12.5V.
After this, the supply never worked again. I have reconnected feedback circuit to original configuration, changed fuse (exploded again), then changed bridge rectifier, both mosfets, 2 resistors under them, thermistor, and 3 diodes - they were black but worked fine - changed them anyway. The mentioned were changed over 3 times trying to turn it on and resulting in main fuse in my apartment switching off every time. Now I have primary side working (no shorts) and 5vSB rail functioning.

Trying to repair the main output I have changed the transformer for mosfet control, found one open on primary side (had to swap pins on secondary side as it was a little bit different), put a new TL494cn and a replacement LM339n, tested every diode and transistor on the pcb.


To sum up/applicable to both atx supplies: i test diodes and transistors by simple DMM (diode range), sometimes including a test of desoldered component.
The reason why I want to bring the 1st supply to life is that it already has nice outputs of voltages, led, switches and so on - it is already converted to easy to use bench top power supply.
The reason for recovering the second one, the q-tek is that it is the only supply controlled by TL494 chip, therefore can be converted to adjustable supply easily.

Any ideas of how to continue troubleshooting both power supplies? feeling desperate