Help needed to revive a blown power supply

[mikey5791]

Hi all,
Got a used working computer running Windows 7 about 3 months ago. Two days ago, the power supply blew up with loud spark and i immediately shut off the power. It was an unknown brand Tatu (output 450 w).
Upon opening, i found out the two transistors on heat sink 1are fried with burn mark. Replaced the two trans(E13007 type). Also checked if the surrounding components are affected. The diode seems to be working, i pulled one leg off and tested using diode test. One side got reading while swapping test probe shows 1(infinite).
I also pulled out the nearest 2 small caps at C5 and C6, they are both 10uf50v. Capacitance read 10.5 to 11.2 uf while esr from 0.46 to 0.5 (based on these reading, i soldered them back on board).
Later, i pulled the 2nd heat sink out, tested those rectifiers by diode test. Black probe on middle leg while red probe on left and right leg got some reading. From the readings, i guess those rectifiers are still good.
Tested the 16 pin(SDC 7500 voltage pwm),dc test with black probe on pin 7 and red probe on pin 12 shows 15vdc, at pin 14 it shows 5vdc.
Tested the 14 pin (SDP 339 voltage quad comparator), dc test with black probe on pin 12 and red probe on pin 3 shows 5vdc.
For testing this blown power supply, i replace the fuse with a series lamp bulb. Upon switched on power, the bulb lighted for 1 second then no more light. So there is no short. But the output power on the 24 pin still not appearing. Only the 5vdc standby power is detected from the green wire. Any help to revive this psu is greatly appreciated.

[Per Hansson]

Re: Help needed to revive a blown power supply

When you switch on the PSU with the green PS-on wire to ground you also need to provide a minimum load in many cases.
An old hard drive that you do not care about can for example act like a minimum load for the power supply.

[PeteS in CA]

Re: Help needed to revive a blown power supply

Realistically, those MJE13007 clones, the size of the main transformer, the wimpy heatsinks, and not very large output inductors mean the realistic maximum continuous output power is in the 200W-25oW range. Note also that the 12V and 5V rectifiers are both rated for less than the label currents, and the 3.3V label and rectifier ratings are equal.

I realize the system had been running for some time with this PSU, so those factors would mainly be relevant if you hope to expand the system.

[momaka]

Re: Help needed to revive a blown power supply

This is a very sad-looking PSU. Probably candidate for this thread, but digress. I don't suggest to repair it for use in a computer, as it's just very anemic. However, if you do want to repair it for learning purposes, these cheap PSUs are fairly straight-forward (well, I say that after going through a similar rodeo and struggling with one for quite a bit.)

That said, probably start by reading this thread:
https://www.badcaps.net/forum/showth...light=kdmpower

In particular, read these two posts in detail:
https://www.badcaps.net/forum/showpo...00&postcount=8
https://www.badcaps.net/forum/showpo...1&postcount=11
https://www.badcaps.net/forum/showpo...3&postcount=12

TL;DR: check the driver transistors (typically 2SC945 or equivalent) and surrounding diodes on the secondary side that drive the middle transformer (which drives the 13007/13009 BJTs on the primary.) See this schematic as a reference:
https://www.badcaps.net/forum/attach...6&d=1608943994
Your PSU may not look exactly the same, but it should be similar. This is just the driving circuit for the main PS. In my case, one of the diodes in front of Q8/Q9 was partially-shorted, and it gave me a good run-around until I found it. So check the driving circuitry and also for any open resistors in that circuit.

[mikey5791]

Re: Help needed to revive a blown power supply

I had used an old hard drive as load but the psu still not working as it should. Also read through the given threads and they are helpful. Since i spent quite an amount of hours without success, i think i will keep this psu aside. Maybe some of the parts will be useful for other project.
Anyway, thank you for all the tips and assistance.

[anikin73]

Re: Help needed to revive a blown power supply

normally when these type of PSU blow, you have to measure the resistors around it also. I make a close up look, your resistors seems to be gone.

[mikey5791]

Re: Help needed to revive a blown power supply

normally when these type of PSU blow, you have to measure the resistors around it also. I make a close up look, your resistors seems to be gone.

Those resistor near bjt13007 are at R8 (1ohm), R5(1ohm) and R6(2.7k ohm). R6 is open and i replace it with 6.8k ohm resistor from my stock.
Will replacing with a higher ohm resistor affect the psu power on?

I am trying to repair this old psu for learning purpose and not to use on computer or anything critical. Fyi, the two big main cap is 470uf 200v type. Upon power on, i tested it and dmm showed dc reading of 158v on each of the cap. Our ac main is 230v type.
The psu output 5vdc standby and no other output. The fan not moving a bit. I replaced all the small caps at c6,c5,c24 and c22 to 10uf 50v type and the 5v standby still remains and no other output. Any suggestions?

[mikey5791]

Hi to all forum members,

Found some free time to go back to this old post. Upon tips from momaka, i desoldered few of the driver transistors near the PG(power good) area. At Q5 is TL431, Q7 is STC945(npn) Q8 is STA733 while at area near the opto isolator is TL431. Apparently after testing by dmm diode test, i found both the trans TL431 are faulty.
Since TL431 is pnp type, on dmm diode test i place black probe on base, while red probe on collector, then on emitter. The red probe showed some reading meaning it is good. But when i place black probe on emitter, then red probe on collector the dmm showed reading...i guessed the TL431 is partly open.
Since both the TL431 are pnp type, i tried replace the faulty ones with A733P for testing since it is also pnp type. Upon power on ac, the 100watt light bulb lits brightly as current limiter. Is it meaning that something is shorted or the A733P is not compatible? Can anyone experience kindly advice what could be wrong here... Thanking in advance.

[stj]

TL431 is not a transistor

[mikey5791]

TL431 is not a transistor

Thank you for informing.
Fyi, i did a rough test on TL431 using input of dc 6v from dry batteries pack assuming the TL431 will switch on the red led upon reaching reference voltage of 2.5v. I had attached image of the test circuit and the breadboard with lighted led. Does this test confirm the pulled TL431 is still good?

Appreciate if anyone can correct me if my test method is not proper and kindly suggest if i can use a digital multimeter for further test.

[momaka]

I'm chiming in a bit late, but here goes...

Those resistor near bjt13007 are at R8 (1ohm), R5(1ohm) and R6(2.7k ohm). R6 is open and i replace it with 6.8k ohm resistor from my stock.
Will replacing with a higher ohm resistor affect the psu power on?

Yes, you need to use something closer to 2.7 kOhms. In fact, I don't recall seeing this resistor anything else but 2.7 kOhms in just about every half-bridge PSU I have seen. Maybe just once I saw 1.5 kOhms, but I don't remember.
So to be safe, use 2.7 kOhm resistor. I suspect 2.2k and 3.0 k would be "acceptable" too. Whatever you pick, make sure both resistors that were 2.7 kOhms originally are now replaced with the same value - you need to have the Base drive to the BJTs fairly symmetric.

Oh, and if your R6 was open, then the 13007 BJTs are probably bad too. Don't even check them, just replace with known good ones. Also change ALL of the components going to the Base of those BJTs - i.e. aside from the 2.7 kOhm resistors, also the one or two diodes (probably 1n4148's), and even both 1-Ohm resistors.

And since you did have a primary-side failure, also check the two TO-92 transistors near the small transformer in the middle. Should be something like 2sc945 or 2sc1815. These drive the middle transformer, which essentially drives the BJTs on the primary side. Also check their protection diodes, bias resistors, and their grounding diode(s).

Fyi, i did a rough test on TL431 using input of dc 6v from dry batteries pack assuming the TL431 will switch on the red led upon reaching reference voltage of 2.5v. I had attached image of the test circuit and the breadboard with lighted led. Does this test confirm the pulled TL431 is still good?

Yes, but add this to your test: if you disconnect the 1k resistor between the 6V battery and the reference pin of the TL431, does the LED go Off. If yes, then the TL431 is working OK.

That said, the TL431 likely isn't a cause for your PSU not working here. These TL431 shunt regulators are used only for the 5VSB rail and the 3.3V rail. Your 5VSB is working, so that leaves the 3.3V rail one. But if that one was damaged, you'd at least see the PSU try to start. In some cases, it may even stay up and running, with the 3.3V rail low/high out of spec. But again, you had failed components on the primary, which is almost certainly going to be a fault somewhere with the BJT drive circuit now, be it on the primary or secondary. So check the components I mentioned above and report back what you find.

[mikey5791]

I'm chiming in a bit late, but here goes...


Yes, you need to use something closer to 2.7 kOhms. In fact, I don't recall seeing this resistor anything else but 2.7 kOhms in just about every half-bridge PSU I have seen. Maybe just once I saw 1.5 kOhms, but I don't remember.
So to be safe, use 2.7 kOhm resistor. I suspect 2.2k and 3.0 k would be "acceptable" too. Whatever you pick, make sure both resistors that were 2.7 kOhms originally are now replaced with the same value - you need to have the Base drive to the BJTs fairly symmetric.

Oh, and if your R6 was open, then the 13007 BJTs are probably bad too. Don't even check them, just replace with known good ones. Also change ALL of the components going to the Base of those BJTs - i.e. aside from the 2.7 kOhm resistors, also the one or two diodes (probably 1n4148's), and even both 1-Ohm resistors.

And since you did have a primary-side failure, also check the two TO-92 transistors near the small transformer in the middle. Should be something like 2sc945 or 2sc1815. These drive the middle transformer, which essentially drives the BJTs on the primary side. Also check their protection diodes, bias resistors, and their grounding diode(s).


Yes, but add this to your test: if you disconnect the 1k resistor between the 6V battery and the reference pin of the TL431, does the LED go Off. If yes, then the TL431 is working OK.

That said, the TL431 likely isn't a cause for your PSU not working here. These TL431 shunt regulators are used only for the 5VSB rail and the 3.3V rail. Your 5VSB is working, so that leaves the 3.3V rail one. But if that one was damaged, you'd at least see the PSU try to start. In some cases, it may even stay up and running, with the 3.3V rail low/high out of spec. But again, you had failed components on the primary, which is almost certainly going to be a fault somewhere with the BJT drive circuit now, be it on the primary or secondary. So check the components I mentioned above and report back what you find.

As per your recommendation, i replaced R6 with a 2.2k ohm resistor. Those diodes 1n4148 connected to base of 13007bjt are tested good with one leg desoldered off board . Also both the 1 ohm resistors are good within spec as well.
In addition , the TO-92 transistor near the middle transformer are the c945 type. Off board, they tested good.
Upon power on ac with series light bulb still connected in place of the fuse. The light bulb lights a second then off, but there is still no dc voltage output on any rail. I wonder what other component could be the cause of no power, with 5v standby still available.

[momaka]

So after checking those two c945 BJTs near the middle transformer, did you also check the small diodes (should be 1n4148 or equivalent) across their C-E terminals? If not, do that as well, and out of circuit. Also, check all of the small resistors and diodes on the secondary near the 7500 PWM controller. For diodes, lift one leg and test for proper Vf (500-700 mV) and open-circuit in reverse polarity. Any diode that seems suspect should be replaced.