Tiny Tech ATX480W output power ticking

[mikey5791]

Hi to all forum members,

This psu (Tiny Tech model ATX 480W) had been installed in a very old Windows XP PC but recently i pulled it out as the pc cannot start. To test it, i jumped the green wire and black wire on the 24 pin connector. The psu fan jerk a bit but not spinning. I used my dmm to check dc voltage on output connector wire, the 5v red wire keeps ticking a bit at around dc1v....then drops and ticking again around 1v. It is the same situation with the 12v yellow wire.

Fyi, i desoldered the secondary side electrolytic cap those over 1000uf type, found 2 caps out of spec and replaced with new ones. Btw, i also replaced those small cap next to 5vsb diode and caps above the +5vsb purple line . The small cap next to green ps on wire tested good.
Tested the 5vsb n channel mosfet CM2N60 at primary heat sink off circuit. On diode mode, red dmm probe to source, black probe to drain. Dmm reads OL. Then black probe to gate, then to drain. Dmm reads continuity.
Voltage test on D9(standby diode) shows dc5.43v while the aux rail is dc 14.28v
In addition, i also desoldered all the resistors around the +5vsb purple wire, the 20pin pwm SD6109, nearby green ps on wire and near grey pg wire. The resistors all tested good within spec.
All the diodes also tested good within spec.
Tested voltage reading on pwm SD6109, pin 20 vcc reads dc5v. Is this sufficient to turn on the psu since the schematic of SD6109 says maximum voltage on pin 20 is dc12v.

On visual inspection, there is nothing burn or any dark spots on the pcb board and i had done most of the preliminary testing. But i might have overlook something else and i look forward to your kind suggestions and advice on how to revive and get this psu to work again. Any comments and tips on this will be greatly appreciated.

[stj]

a 300w psu on a good day - dont try to pull more than 250 from it unless the fan is good.
usually a fan twitch is an output overload, check the output diodes for shorts now you know the caps are atleast useable.

[mikey5791]

Before power on ac, i connect the psu power cable to a series bulb limiter and the bulb lights a second then complete lights off.
Is this confirming there is no shorts inside the psu or do i need further test to confirm the output diodes for shorts? Kindly advise. Thanks for the tip.

[mikey5791]

On secondary side, the dummy resistor at 5v rail is green,brown,black,gold(51 ohm) while at 12v rail is red,violet,brown,gold(270 ohm) and at 3.3v rail is brown,green,black,gold(15 ohm).
Did resistance test to ground, at red wire 5v rail resistance is 48 ohm while at yellow wire 12v rail resistance is 157 ohm and at orange wire 3.3v rail resistance is 10 ohm.

[stj]

probably o.k. then.
maybe a small cap is bad on the feedback so it thinks there is an overvoltage condition.

[CapLeaker]

a 300w psu on a good day - dont try to pull more than 250 from it unless the fan is good.
usually a fan twitch is an output overload, check the output diodes for shorts now you know the caps are atleast useable.

480W out of this? No way! Even the 250w is pushing it.

[stj]

the label does say 200w, but they are using a deceptive part number!

[Per Hansson]

the label does say 200w, but they are using a deceptive part number!

That's for the AC side.
The label claims:
12v x 17A = 204W
5v x 28A = 140W
3.3v x 16A = 53W
-12v x 0.5A = 6W
That totals 403W and the model number tries to make it sound like 480W, so it's a over unity free energy device!
This power supply is extremely unlikely to produce half of what the label claims within ATX spec.
It belongs in a scrap pile, not a computer.
Anyway, for test you usually need a load, so attach an couple old hard drives you don't care about and try again...

[momaka]

It belongs in a scrap pile, not a computer.

Agreed!

No AC input filter, so enjoy all of that HF noise go back on your AC line!
The output cap setup is probably only good for ~100W of load before the noise gets out of spec... maybe 120-130W with some luck.

As for the label, it's a lie, as we already know. Gutless PSUs like this usually have something like 10 Amp rectifier on the 3.3V rail, 16 Amp on the 5V rail, and a 10 or 12 Amp on the 12V rail - basically just enough to power a simple low-power 12V -based PC (i.e. Athlon 64 or low[er]-power P4 or low-end i3) with no additional discreet GPU and not more than a HDD or two. On the "better" units, 3.3V and 5V may get something like a 20 Amp rectifier, while 12V may get 16 Amp... but that's more rare in my experience with collecting these.

Anyway, for test you usually need a load, so attach an couple old hard drives you don't care about and try again...

Yes, always test ATX PSUs with a load of some type - preferably something that can put at least about 0.5 Amps on the 5V and 12V rails.
Usually, crap PSUs like this will turn On and run "OK" even without a load... but it's not guaranteed. So for proper testing, I agree with Per here - attach a few HDDs or whatever else you have that you don't care much about.
I usually go with 12V incandescent auto bulbs that are in the range of 20-50 Watts: one on the 5V rail and one on the 12V rail. 3.3V rail is usually independently regulated, so no need to load it. If PSU still doesn't work with a load, time to dig in deeper (and hopefully just as a learning experience rather than with the intention to use this PSU again.)

Also, VERT IMPORTANT: when testing an ATX PSU with a load (and for some PSUs even without a load), sometimes the series incandescent lamp on the input can interfere with proper operation... in which case, you either have to not use it, or use several incandescent bulbs all in parallel to each other to act as a higher wattage bulb. 2x or 3x 100W will usually allow an ATX PSU to sustain about 20-40 Watts worth of load.

Fyi, i desoldered the secondary side electrolytic cap those over 1000uf type, found 2 caps out of spec and replaced with new ones. Btw, i also replaced those small cap next to 5vsb diode and caps above the +5vsb purple line . The small cap next to green ps on wire tested good.

If you had any bad caps anywhere, usually the rest are not far behind.... so best to check every single cap in the PSU, except the big 200V ones, as those rarely go bad.

Voltage test on D9(standby diode) shows dc5.43v while the aux rail is dc 14.28v

Check the battery in your multimeter first. If it's good, then your 5VSB circuit is going out of spec with that voltage here. Allowable values for 5VSB are 4.75 to 5.25V (i.e. 5% tolerance.) 5.4xV is dangerously close to SIO/LPC and TTL circuitry -frying territory. Check all of the small caps on and around the 5VSB circuit. The small one close to the 5VSB transformer on the primary is a critical cap, usually. When it goes high ESR, 5VSB tends to drift up.

That said, since you do get 5VSB to output voltage, there's no point to remove the 5VSB MOSFET to check it. If it was dead, you wouldn't get 5VSB. In fact, if it was dead, the PSU wouldn't even try to start or do anything when your short green wire (PS_ON signal) to ground.

Tested voltage reading on pwm SD6109, pin 20 vcc reads dc5v. Is this sufficient to turn on the psu since the schematic of SD6109 says maximum voltage on pin 20 is dc12v.

Yes, 5V is fine.
Operating voltage for this IC is 4.5 to 5.5V.
12V is the absolute maximum withstanding (for short periods of time - i.e. voltage peaks), which means IC damage will very likely occur if you go beyond that value.

On visual inspection, there is nothing burn or any dark spots on the pcb board and i had done most of the preliminary testing. But i might have overlook something else and i look forward to your kind suggestions and advice on how to revive and get this psu to work again. Any comments and tips on this will be greatly appreciated.

Check all of the small/tiny electrolytic caps on the secondary side. The one on the aux. rail is critical for the main PS to work.

You already performed a check for short-circuit on all of the output rails and found they had low-value load resistors, which is normal. Unfortunately, those low-value resistors can sometimes mask if there is another issue hiding. I usually remove them before checking the rectifiers, so that way they don't interfere with my measurements. Also worth checking for short-circuit between the various rails - i.e. 3.3V to 5V, 12V, -5V, and -12V; 5V to 12V, -5V, and -12V, and etc... just to make sure you don't have any windings shorted on the output toroid.

You can also try the voltage-injection method I suggested in this thread (go to post #16 and read onwards):
https://www.badcaps.net/forum/troubl...a-split-second
.

[PeteS in CA]

No AC input filter, so enjoy all of that HF noise go back on your AC line!

On the plus side, users would not have to worry about Y-caps that are not safety agency approved.

[Dan81]

Before power on ac, i connect the psu power cable to a series bulb limiter and the bulb lights a second then complete lights off.
Is this confirming there is no shorts inside the psu or do i need further test to confirm the output diodes for shorts? Kindly advise. Thanks for the tip.

I doubt a single watt in spec, so 0W PSU. 100W if not counting ripple.

[momaka]

On the plus side, users would not have to worry about Y-caps that are not safety agency approved.

Heh, I'm personally not worried about those anymore either - most of the time anyways.

For the "Y2" caps between Ground and Neutral in particular, you can leave those even if they are non-approved caps, because there's no high voltage to be expected between these two (after all, Neutral and Ground are tied together at the breaker box.) In EU land with the Schuko plug being non-polarized, though, that's a different story, because you never know which side / AC input pin the live AC line will come from. So in that case, both caps need to be Y2 safety-approved, just in case.

As for Y2 caps between primary-side negative rectified bus rail and ground - those should ALWAYS be safety-approved, especially on non-grounded power adapters and equipment.

I doubt a single watt in spec, so 0W PSU. 100W if not counting ripple.

Nah, with a single cap on each rail, you might be able to pull about 3-4 Amps from each with ripple just still barely passing ATX spec.
Of course, considering that 4 Amps will net you a maximum of 48 Watts (from the 12V rail)... well, that pretty much tells you the kind of hardware you can run on this PSU. Atom CPU and on-board everything? - Yup, that's about it.

[mikey5791]

That's for the AC side.
The label claims:
12v x 17A = 204W
5v x 28A = 140W
3.3v x 16A = 53W
-12v x 0.5A = 6W
That totals 403W and the model number tries to make it sound like 480W, so it's a over unity free energy device!
This power supply is extremely unlikely to produce half of what the label claims within ATX spec.
It belongs in a scrap pile, not a computer.
Anyway, for test you usually need a load, so attach an couple old hard drives you don't care about and try again...

Fyi, i attached two old ide hard drives as load for testing but the psu fan remains still and there is no dc output .

[mikey5791]

Agreed!

No AC input filter, so enjoy all of that HF noise go back on your AC line!
The output cap setup is probably only good for ~100W of load before the noise gets out of spec... maybe 120-130W with some luck.

As for the label, it's a lie, as we already know. Gutless PSUs like this usually have something like 10 Amp rectifier on the 3.3V rail, 16 Amp on the 5V rail, and a 10 or 12 Amp on the 12V rail - basically just enough to power a simple low-power 12V -based PC (i.e. Athlon 64 or low[er]-power P4 or low-end i3) with no additional discreet GPU and not more than a HDD or two. On the "better" units, 3.3V and 5V may get something like a 20 Amp rectifier, while 12V may get 16 Amp... but that's more rare in my experience with collecting these.


Yes, always test ATX PSUs with a load of some type - preferably something that can put at least about 0.5 Amps on the 5V and 12V rails.
Usually, crap PSUs like this will turn On and run "OK" even without a load... but it's not guaranteed. So for proper testing, I agree with Per here - attach a few HDDs or whatever else you have that you don't care much about.
I usually go with 12V incandescent auto bulbs that are in the range of 20-50 Watts: one on the 5V rail and one on the 12V rail. 3.3V rail is usually independently regulated, so no need to load it. If PSU still doesn't work with a load, time to dig in deeper (and hopefully just as a learning experience rather than with the intention to use this PSU again.)

Also, VERT IMPORTANT: when testing an ATX PSU with a load (and for some PSUs even without a load), sometimes the series incandescent lamp on the input can interfere with proper operation... in which case, you either have to not use it, or use several incandescent bulbs all in parallel to each other to act as a higher wattage bulb. 2x or 3x 100W will usually allow an ATX PSU to sustain about 20-40 Watts worth of load.


If you had any bad caps anywhere, usually the rest are not far behind.... so best to check every single cap in the PSU, except the big 200V ones, as those rarely go bad.


Check the battery in your multimeter first. If it's good, then your 5VSB circuit is going out of spec with that voltage here. Allowable values for 5VSB are 4.75 to 5.25V (i.e. 5% tolerance.) 5.4xV is dangerously close to SIO/LPC and TTL circuitry -frying territory. Check all of the small caps on and around the 5VSB circuit. The small one close to the 5VSB transformer on the primary is a critical cap, usually. When it goes high ESR, 5VSB tends to drift up.

That said, since you do get 5VSB to output voltage, there's no point to remove the 5VSB MOSFET to check it. If it was dead, you wouldn't get 5VSB. In fact, if it was dead, the PSU wouldn't even try to start or do anything when your short green wire (PS_ON signal) to ground.


Yes, 5V is fine.
Operating voltage for this IC is 4.5 to 5.5V.
12V is the absolute maximum withstanding (for short periods of time - i.e. voltage peaks), which means IC damage will very likely occur if you go beyond that value.


Check all of the small/tiny electrolytic caps on the secondary side. The one on the aux. rail is critical for the main PS to work.

You already performed a check for short-circuit on all of the output rails and found they had low-value load resistors, which is normal. Unfortunately, those low-value resistors can sometimes mask if there is another issue hiding. I usually remove them before checking the rectifiers, so that way they don't interfere with my measurements. Also worth checking for short-circuit between the various rails - i.e. 3.3V to 5V, 12V, -5V, and -12V; 5V to 12V, -5V, and -12V, and etc... just to make sure you don't have any windings shorted on the output toroid.

You can also try the voltage-injection method I suggested in this thread (go to post #16 and read onwards):
https://www.badcaps.net/forum/troubl...a-split-second
.

Fyi, i did checking for short circuit between the various rails by setting my dmm to 2k resistance mode. Here are the results:
3.3V to 5V is 0.039 ohm
3.3V to 12V is 0.163 ohm
3.3V to -12V is 0.508 ohm
5V to 12V is 0.186 ohm
5V to -12V is 0.531 ohm
12V to -12V is 0.655 ohm
(For info, this psu does not have -5V rail)

In addition, i tried the voltage injection method as suggested. I connect Ground wire of a good psu to ground of the bad psu. My red dmm(on 10A socket) to 3.3v rail of good psu while my black dmm to 3.3v rail of the bad psu, while my dmm set to 10A current mode. By power ac on the good psu, short green to black wire to ps on, my dmm reads 0.17A. Then i power ac on the bad psu, short green to black but my dmm still reads 0.17A. Still no dc output on bad psu.
I did the same to the 5v rail, my dmm reads 0.08A.
Also no dc output on bad psu.
Did the same to the 12v rail, my dmm reads 0.13A.
Also no dc output on bad psu.
Any further advice on checking to revive this psu will be much appreciated.

[mikey5791]

Btw, am i doing the voltage injection test correctly as per my post no.14? Or i need to inject all the three rails namely 3.3v, 5v, and 12v simultaneously...

[momaka]

Fyi, i did checking for short circuit between the various rails by setting my dmm to 2k resistance mode. Here are the results:
3.3V to 5V is 0.039 ohm
3.3V to 12V is 0.163 ohm
3.3V to -12V is 0.508 ohm
5V to 12V is 0.186 ohm
5V to -12V is 0.531 ohm
12V to -12V is 0.655 ohm
(For info, this psu does not have -5V rail)

I think those readings above should be in the Kilo-Ohm range?? Doesn't make sense otherwise. Please inform.

And in any case, that's still not showing anything useful here, because you have the "dummy" minimum load resistors you mentioned in post # 4 above (i.e. 51 Ohms for 5V rail, 15 Ohms for 3.3V rail, and 270 Ohms for 12V rail.) In particular, the ones on the 3.3V rail and 5V rail are quite low in value and you may not see a partially-shorted rectifier. That said, partially-shorted rectifiers are rare, so low chance this would be the issue. Nevertheless, it is something to keep in mind and further check upon if nothing else pops out. (To do that, remove the dummy load resistors and test each rail's resistance to ground as well as to each other... and make sure to leave your meter connected for a few second so that the readings stabilise - that should make it clearer to see what is going on.)

In addition, i tried the voltage injection method as suggested. I connect Ground wire of a good psu to ground of the bad psu. My red dmm(on 10A socket) to 3.3v rail of good psu while my black dmm to 3.3v rail of the bad psu, while my dmm set to 10A current mode. By power ac on the good psu, short green to black wire to ps on, my dmm reads 0.17A. Then i power ac on the bad psu, short green to black but my dmm still reads 0.17A. Still no dc output on bad psu.
I did the same to the 5v rail, my dmm reads 0.08A.
Also no dc output on bad psu.
Did the same to the 12v rail, my dmm reads 0.13A.
Also no dc output on bad psu.

OK, at least that confirms no component in the DC filter circuit is failing under load.

Btw, am i doing the voltage injection test correctly as per my post no.14? Or i need to inject all the three rails namely 3.3v, 5v, and 12v simultaneously...

You should do both.

The first test where you inject voltages one by one is to see if any DC rail has problems (drawing too much current)... which we see is not the issue here.

The 2nd test where you inject voltages simultaneously is useful to see if the PSU can stay up --- that is, PG (power good) line goes high to indicate the PWM IC thinks the output voltages are OK. If it does, then you can remove the injected voltage rails one by one, to see when the PG goes low. On the other hand if PG stays low, then that means the PWM IC is still detecting an issue somewhere.

[mikey5791]

Sorry for late reply. Got some matters to attend lately.

Yes, my previous readings are in kilo ohm range.
Fyi, i did the voltage injection where i injected all 3.3v, 5v, and 12v simultaneously from a good psu.
The bad psu shows purple line standby votage of 4.54v and grey PG line of 5.26v. Then i removed the orange 3.3v line, the bad psu grey PG line goes up to 5.53v.
Next i removed the red 5v line, the bad psu grey PG line drops to 0.26v. May i know what all these tests are for? Is it saying there is still an issue at 5v line ? Kindly advice and guide if i need to do further test. Thank you.

[mikey5791]

Hi momaka,
As per your comment on post #16,

"To do that, remove the dummy load resistors and test each rail's resistance to ground as well as to each other... and make sure to leave your meter connected for a few second so that the readings stabilise - that should make it clearer to see what is going on.)"
Fyi, i removed all the 3 dummy load resistors 15ohm, 51 ohm and 270 ohm for all the 3 rails. The resistance test on 2k scale are as follows:
3.3v to ground="1"(this means infinite on my dmm)
5v to ground="1"
12v to ground=1.932k ohm
-12v to ground=501 ohm
3.3v to 5v="1"
3.3v to 12v="1"
3.3v to -12v=1.05k ohm
5v to 12v="1"
5v to -12v=1.308k ohm
12v to -12v="1"

In addition, i had also desoldered three nearby driver transistor at Q8(STA733 pnp trans), Q6&Q5(both are 2N2222 npn trans) and test their specs on a transistor tester. They all seemed tested good. As attached images.
Kindly inform if i need to perform any other test.
Will appreciate for your further comment and suggestions to revive this low quality psu for future small electronic project use instead of throwing into the landfill. Thank you in advance.