PT-A450W repair failure

[pfrcom]

This may cause some from the electronics gurus around here, but ...

Cheap & nasty power supply had one of its pair of STI 13007A blow apart - mounted on the primary heat sink if I understand correctly

Decided to try a repair as a learning exercise and scavenged a pair of "F E13009F" from some scrap I had lying around

They were associated with a "F 2N60B", versus the original STIs which had a "F 2N60C"

After getting data sheets for both 2N60s, I decided they were compatible, and subbed in the "F E13009F" as a pair, replacing both STI 13007A

Now the power supply won't even turn on

What have I done wrong ? - apart from not throwing it in the garbage in the first place

[stj]

Re: PT-A450W repair failure

does it have the standby supply, or 100% dead?

[selldoor]

Re: PT-A450W repair failure

Im sure you have these covered but:
Fuse OK
Load connected ( If Required)
Bench switch on black and green ?
Some power supplies also need a power good signal on pink or grey?

[stj]

Re: PT-A450W repair failure

i think the power-good is an output from the psu controller chip, not a feedback.

what's black&green?

[selldoor]

Re: PT-A450W repair failure

Hi - sorry assumed the number of posts the OP has made they would know

when testing a psu not connected to a motherboard (benchtest) it is usually switched on by a jumper (paperclip) on the socket connecting the green wire to any of the black ground wires.

The grey pink as I recall was on a dell psu and retiredcaps was having difficulty
starting one using just the black and green and the answer was

"For some reason, PWR_OK wasn't getting pulled up to 5V on this model. So I jumped 5V FP to PWR_OK in addition to PS_ON to GND. So with both jumpers in place, the 5161-7DS turns on."

[stj]

Re: PT-A450W repair failure

hmm.

well they dont always follow the same colur code for the signal cables.
specially if it's dell!

[pfrcom]

Re: PT-A450W repair failure

does it have the standby supply, or 100% dead?

5.03V measured between 5VSB & Ground on ATX connector

Im sure you have these covered but:
Fuse OK
Load connected ( If Required)
Bench switch on black and green ?
Some power supplies also need a power good signal on pink or grey?

Fuse is OK
HDD & ODD connected
Won't power on with a jumper between PSON & Ground, then switched on at wall
Haven't tried extra jumper between Power Good & Ground

I've tested each of the three components on that heat sink, namely one "F 2N60C" and two "F E13009F"
Done separately, in situ, using the Diode Test setting of my Fluke 77
None did a continuous beep between middle leg and either outside leg, or between outside legs
All did a quick beep between various pairs of legs, which in my limited experience usually means a good IC

When switched on at the wall, a high pitched "singing" starts, I think, from around the transformers? between the two heat sinks

[stj]

Re: PT-A450W repair failure

o.k. check the diodes on the other heatsink for shorts.
screaming usually means you have an overload.

[pfrcom]

Re: PT-A450W repair failure

check the diodes on the other heatsink for shorts.

Here's pics

For all three ICs, just a quick beep but no continuous beep from centre leg to either left or right leg, or from left to right leg, with my Fluke 77 meter in Diode Test / Continuity mode

But there is a continuous beep if I test equivalent solder points on the PCB

Does this mean the problem is further down the food chain ?

[budm]

Re: PT-A450W repair failure

The middle device is the Common Cathode (middle pin) Diode, the two anodes (two outer pins) are tied together so the two diodes are connected in parallel. Is that the one that shows continuity on the copper pads?

[stj]

Re: PT-A450W repair failure

ge rid of the brown glue,
and check those diodes around where the heatsink was.
then check the ceramic capacitors.
that's the flat disc's

[momaka]

Re: PT-A450W repair failure

Also measure resistance between ground and 3.3V, 5V, 12V, -5V (if present), and -12V rails. There appears to be loading resistors on the secondary, so you likely will get a low resistance reading on 3.3V and 5V rails. That said, do post what resistances you get.

The other part that commonly fails is the BJT driver transistors on the secondary side (usually a pair of TO-92 transistors connected to the middle transformer), the Base drive resistors and diodes for the main BJTs on the primary side (also connected to middle transformer).

Lastly, the 5VSB has a critical cap. Change it, along with the 5VSB output cap. The 5VSB may appear to spit out 5V properly, but the auxiliary (usually 10 to 20V) for the controller could be much higher. In fact, try loading the 5VSB with about 1A and see if it's stable at 5V or if it makes any weird noises. Bad output caps on the 5VSB and a bad critical cap can damage the controller IC.

[pfrcom]

Re: PT-A450W repair failure

The middle device is the Common Cathode (middle pin) Diode, the two anodes (two outer pins) are tied together so the two diodes are connected in parallel. Is that the one that shows continuity on the copper pads?

Using Diode Test/Continuity setting, with Ground probe on PCB pad for centre leg of middle IC, got continuous beep to all other pads but one
Exception was one side pad for middle IC, which gave just a quick beep

If these results don't make sense, I may not have been thorough enough - pursuing a wild goose chase instead
Found two of the 470uf 16v caps which were canted over, were doing so because they were bulging out from the bottom, not because they were assembled that way
So I replaced them, resoldered the heatsink & associated ICs to the PCB, reassembled it all and hoped for the best

Doubtless you won't be surprised to hear - nothing, not even a twitch from its fan

Although the "singing" noise after power is applied, now seems to be gone

I'm starting to feel like the person who's swum half-way across the pool, and decides they can't make it to the other side
Should have thrown the damn thing away in the first place

ge rid of the brown glue

Easier said than done, but I have removed what I could see/get at

and check those diodes around where the heatsink was

Found I had continuity in one direction, but none in other direction - sufficient check ?

then check the ceramic capacitors.
that's the flat disc's

Got a quick beep from each, by touching their legs on bottom side of PCB

Also measure resistance between ground and 3.3V, 5V, 12V, -5V (if present), and -12V rails

Measured at ATX connector:
Gnd to 3.3v - 8 ohm
Gnd to 5v - 45.8 ohm
Gnd to 12v - 213 ohm
Gnd to -5v - 509 ohm
Gnd to -12v - 511 ohm

Lastly, the 5VSB has a critical cap. Change it, along with the 5VSB output cap.

Please help me ascertain which ones they are

[stj]

Re: PT-A450W repair failure

find the 5v stanby wire soldered to the board - then follow the track.

[momaka]

Re: PT-A450W repair failure

Found two of the 470uf 16v caps which were canted over, were doing so because they were bulging out from the bottom, not because they were assembled that way
So I replaced them, resoldered the heatsink & associated ICs to the PCB, reassembled it all and hoped for the best

Doubtless you won't be surprised to hear - nothing, not even a twitch from its fan

Although the "singing" noise after power is applied, now seems to be gone

What rails were these caps connected to?

Measured at ATX connector:
Gnd to 3.3v - 8 ohm
Gnd to 5v - 45.8 ohm
Gnd to 12v - 213 ohm
Gnd to -5v - 509 ohm
Gnd to -12v - 511 ohm

Resistances seem reasonable if there is a load resistor on each rail. I would guess 6.8 Ohms to 9.1 Ohms for the 3.3V rail, 47 Ohms for the 5V rail, and possibly 220 Ohms for the 12V rail. Check the big resistors next to the output caps. See if they have similar values. This is just to make sure some of these low resistances aren't due to something else.

Please help me ascertain which ones they are

See this picture:
https://www.badcaps.net/forum/attach...4&d=1403929078
There are three small electrolytic caps next to the primary heatsink. Two of them are for the BJT base drive circuitry - usually 1 uF to 10 uF and rated for 50V. They should both have the same value. The last one, closest to the 5VSB transformer (the small transformer on the bottom of the three), is the critical cap. Usually 10 uF to 47 uF, and 25V to 50V.

I'm guessing you still haven't checked the BJT drive circuitry too. Above those small electrolytic caps on the primary side (particularly, the two identical ones), you'll see some diodes and resistors. Make sure those are okay too - that is, check the resistors for open-circuit and the diodes for short-circuit.
As for the BJT drive transistors on the secondary, my guess is (based on that picture above, again) that they are located in front of that 3.3V torroid. Not sure, however, so either you will need to provide a good picture of the bottom side of the PSU or trace them yourself. They are normally a pair of C945's. At least one of their pins should be connected to the middle transformer. Probably will have to remove them out of circuit to check them. Make sure they don't have any short-circuit between any of their pins out of circuit.

[PeteS in CA]

Re: PT-A450W repair failure

If you get this fixed, I would not use it for much more than 200 or 230 Watts. There is no way in Hades those heatsinks or main transformer are good for 450W! For more than a few seconds, at least.

[pfrcom]

Re: PT-A450W repair failure

What rails were these caps connected to?

The 12V, I think

There are three small electrolytic caps next to the primary heatsink. Two of them are for the BJT base drive circuitry - usually 1 uF to 10 uF and rated for 50V. They should both have the same value. The last one, closest to the 5VSB transformer (the small transformer on the bottom of the three), is the critical cap. Usually 10 uF to 47 uF, and 25V to 50V.

There were two identical 22uf 50V, and one 47uf 50V

Replaced the 22uf with ones recycled (probably from a deceased motherboard) - the replacements measured ~1.2 on Dick Smith ESR meter, compared to both originals ~60

The original 47uf measured ~1.6 so I put it back in

Not good practice, but I begrudge spending a cent on this P.O.S.

Reassembled yet again, switched on power, and nothing

I'm guessing you still haven't checked the BJT drive circuitry too. Above those small electrolytic caps on the primary side (particularly, the two identical ones), you'll see some diodes and resistors. Make sure those are okay too - that is, check the resistors for open-circuit and the diodes for short-circuit.
As for the BJT drive transistors on the secondary, my guess is (based on that picture above, again) that they are located in front of that 3.3V torroid. Not sure, however, so either you will need to provide a good picture of the bottom side of the PSU or trace them yourself. They are normally a pair of C945's. At least one of their pins should be connected to the middle transformer. Probably will have to remove them out of circuit to check them. Make sure they don't have any short-circuit between any of their pins out of circuit.

Found one C945 with continuity beep between E & C, so I replaced it with a recycled H945

With all transistors back in place, continuity tests gave quick beeps, so high hopes ...

... dashed - still as dead as a doornail

This thing's getting dangerously close to going where it should have gone in the first place

[momaka]

Re: PT-A450W repair failure

The 12V, I think

Okay, just checking, in case they came out of the 5VSB.
That said, DO CHANGE the output caps on the 5VSB output with known good ones, even if the original ones do look fine. 2x 680 uF to 1000 uF should do the trick.
You may be getting 5V at the 5VSB output, but the secondary-side auxiliary rail for the PWM controller could still be haywire and all over the place.

There were two identical 22uf 50V, and one 47uf 50V

The two identical 22 uF caps are then the ones for the BJT drive on primary. Good you changed them. Their ESR was quite high. I wouldn't be surprised if something else is bad in that circuit. So like I said, check the resistors and diodes in that same circuit - it's the circuit that contains those 22 uF caps and also connects to the bases of the two 13007/13009 transistors and the middle tranformer. Report back what you get.

Not good practice, but I begrudge spending a cent on this P.O.S.

I totally understand you on that - I do the same quite often as well. If I get the device to work and if it's worthwhile repairing, then I put good caps in it. Otherwise, I just repair it with crappy parts just for the fun of the repair itself.

Found one C945 with continuity beep between E & C, so I replaced it with a recycled H945

Replace both 945's then, along with the protection diodes across their C-E junctions (usually a pair of 1N4148). If one of those transistors has shorted, there's a high chance the other is bad too. They do very occasionally fail open-circuit.
Other items to check:
- that the 494 PWM controller is getting a voltage on its VCC pin (and if yes, what voltage is it?)
- there should be two diodes in series. First diode should have it's cathode connected to ground and anode connected to the cathode of second diode. The second diode's anode should connect between the path that joins C on one 945 and E on the other 945. Check that these series diodes are good, or replace if in doubt. Usually, they are standard 1N4148 as well.
- one of the 945 transistors should have its base pulled high. Probably almost as high as the VCC on the 494 controller.

This thing's getting dangerously close to going where it should have gone in the first place

Maybe it will go there regardless. But better learn from it first, no?

[kc8adu]

Re: PT-A450W repair failure

yes its a cheap education.
practice on it and you will know what to do when one comes along thats worthwhile to save.