diagnosing an old ATX PSU help identifying resistors (PowerComputing TCX-20D)

[tony359]

Re: diagnosing an old ATX PSU help identifying resistors

I don't see why it should have changed

[CapLeaker]

Re: diagnosing an old ATX PSU help identifying resistors

Strange… what does the meter show with is set on AC and measure the AC side of the rectifier?

[stj]

Re: diagnosing an old ATX PSU help identifying resistors

put a new battery in the meter

[Per Hansson]

Re: diagnosing an old ATX PSU help identifying resistors

I'm befuddled!
I used your picture here and here to try to make some sense out of this.
This power supply looks to use a linear (50Hz) transformer right at the input? (the one marked 990-00084-00) I have never ever seen that.
This transformer seems to connect to the small vertical PCB right above it, what is on that PCB, just passive components or something active?
Traditionally this transformer would have been the common mode rejection (EMI purposes).

Then we get to another small transformer (913-20002-01) and it is hard to make out, but it looks like the normal common mode rejection transformer.

So far I have not been able to identify anything that could boost the voltage, that would be something that looks just like this second transformer but with just a single wire, and controlled by a MOSFET on the primary side.
Without that I see no way to get the readings you do so you will need to trace the circuit better, or confirm that your line voltage isn't really high?
(It would require ca 280VAC on the primary side to get the secondary voltage you see).
So it's either that or the first transformer mentioned is actually an inductor too and the associated vertical PCB contains the active circuit to drive it?

[stj]

Re: diagnosing an old ATX PSU help identifying resistors (PowerComputing TCX-20D)

looking at that,
the transformer is driven directly from mains, it feeds the small board.
the board looks like it switches the main supply and may also be the regulator for the standby.

if so the standby current must be something like 500mA!!!

[CapLeaker]

Re: diagnosing an old ATX PSU help identifying resistors (PowerComputing TCX-20D)

looking at that,
the transformer is driven directly from mains, it feeds the small board.
the board looks like it switches the main supply and may also be the regulator for the standby.

if so the standby current must be something like 500mA!!!

Thats why I was wondering what the voltage of on the AC side on the bridge rectifier is.

[tony359]

Re: diagnosing an old ATX PSU help identifying resistors (PowerComputing TCX-20D)

You’re right, the small vertical board has a voltage regulator for 5VSB and I do get 5VSB out of the whole PSU.

Voltage on bridge is 240V AC.

[CapLeaker]

Re: diagnosing an old ATX PSU help identifying resistors (PowerComputing TCX-20D)

I went and reread that whole thread and looked at these pictures again. There are 2 things I noticed:
There is brown glue on the primary. Remove that brown stuff on sight as it gets conductive. I see it on the bridge rectifier for example.
Next I noticed that you get like 16V peak to peak ripple on the main filter caps. Well that can't be any good either.
Something smells fishy to me. There is no way you should get that increased voltage what you are reporting.

[tony359]

Re: diagnosing an old ATX PSU help identifying resistors (PowerComputing TCX-20D)

First, thanks for taking the time to re-read the whole thread!

You're right about that conductive glue. It didn't go under the caps but you're right about the bridge, I think there is some ON it. I'll check and remove.

The 16V P/P ripple sounds weird indeed but I was also using my differential probe set to 500x which is not super-accurate. Still, 16V p/p is quite a lot. Maybe I can zoom in a bit and see what's going on there.

[stj]

Re: diagnosing an old ATX PSU help identifying resistors (PowerComputing TCX-20D)

use your multimeter and set it to AC volts.

[tony359]

Re: diagnosing an old ATX PSU help identifying resistors (PowerComputing TCX-20D)

246VAC in, 395VDC out. 0.122VAC ripple.

That said, I somehow measured 50VAC ripple for a moment (I have that on video!) then it measured 0.122VAC...

I removed the conductive stuff, the one around the bridge was still "soft" and not conductive. The stuff around the caps was dried and conductive.

[CapLeaker]

Re: diagnosing an old ATX PSU help identifying resistors (PowerComputing TCX-20D)

I don’t understand how you get more voltage out of a simple voltage doubler than mathmatical possible. Your AC and the ripple measured with the fluke is believable. The DC measured with the fluke and also the 16V p to p with your scope, not so much. With 16V peak to peak ripple, the caps certainly would be toast.

But how you get 395VDC out of something where there should be 348VDC beats me. I got to admit, I never came across something like this.

If it be mine, I’d be very tempted breaking the 2 high voltage DC traces right after the main filter caps so the high voltage DC can’t get any further and measure again. Then I also take the bridge rectifier out and solder on a 400V or better cap and check it again. That’s just wild.

[tony359]

Re: diagnosing an old ATX PSU help identifying resistors (PowerComputing TCX-20D)

I would need to check again with the scope, it’s a cheap MicSig differential probe and at 500x it gets pretty noisy. I’d trust the Fluke more for the time being.

I can try what you say, I’ve already ‘gutted’ the enclosure and replaced the psu with another one so this is my box of mystery parts ��

I understand what you mean, it doesn’t make sense at all.

[tony359]

Re: diagnosing an old ATX PSU help identifying resistors (PowerComputing TCX-20D)

Could it be that I was measuring 490V because I was measuring with the bleeding resistors removed - and then I measured 395V with them back on the PCB?

[natic]

Re: diagnosing an old ATX PSU help identifying resistors (PowerComputing TCX-20D)

These "bleeding resistors" have actually two functions in this case:
1st) Discharge the main filter capacitors when the power supply is disconnected from the mains power.
2nd) They distribute the output voltage from the bridge rectifier symmetrically to the both main filter capacitors.

I don't recommend to connect this power supply to the mains with these resistors removed, because the voltage on one of the main filter capacitors may exceed the capacitor rated voltage, which may lead to the damage to the capacitor.

First of all remeasure the voltages with another meter to make sure the measurements are correct.

[tony359]

Re: diagnosing an old ATX PSU help identifying resistors (PowerComputing TCX-20D)

Thanks.
Yes, someone explained me they also balance the voltage.

However when I measured, both capacitors were getting more or less the same voltage.

You're not the first one asking me to re-measure, I've done that already multiple times, with the Fluke, with an Extech and with the oscilloscope
Yes, I know it doesn't make any sense

[petehall347]

Re: diagnosing an old ATX PSU help identifying resistors (PowerComputing TCX-20D)

try a new rectifier .
all i can think is some sort of Joule thief is at play here .

[momaka]

Re: diagnosing an old ATX PSU help identifying resistors (PowerComputing TCX-20D)

On initial though, it crossed my mind that this PSU might have one of them "booster" circuits, as outlined in this thread:
https://www.badcaps.net/forum/showth...Jye#post103490

Detailed circuit from there can be seen here:
https://www.badcaps.net/forum/attach...0&d=1268166583

However, I don't see the 2 caps, diodes, and resistors that would be associated with that circuit anywhere on this PSU. So that's not it.

The only other thing I can think of is that somehow the main PS switch components on the primary are causing some kind of an interference and thus fooling your meter(s) to show a higher voltage. But that doesn't make much sense either.
I don't think the little line-connected transformer used for the 5VSB could be the culprit either.
And the common-mode choke... again, I don't see how that can be boosting the voltage.

Perhaps do as was suggested earlier and disconnect everything after the primary caps. For the main PS on the primary side, there should only be a low-value metal oxide resistor that would need to be disconnected. It's the one connected directly between (-) primary bus and the MOSFET's Source pin on the primary. If that doesn't do it, remove the high-value resistor that feeds the PWM IC on the primary side with standby / startup power (typically a 500k or higher resistor.)

The switch must have been marginal, it tests 100% in resistance mode. I guess it leaks some voltage when higher voltage is applied. That might explain the "in between" voltage of 490V I measured?

Shouldn't be the case. If it "leaks" any voltage, you'd definitely hear arcing noise.
But maybe remove that switch altogether, just in case. It's also annoying if that switch is set incorrectly and you live in a country with 220/230/240V AC. The usual result is blown primary fuse and possibly caps too, if there were no MOVs to protect them.

I did not test the capacitors but I know they were making sparks and lighting up my lightbulb when I went to discharge them - so they are not dead. I can test the capacity no problem.

Yeah, they are probably OK then.
Still nice to confirm their capacity and all... but I doubt they would be the cause of this.
.
.
.
Anyways, all of that above is just to see why there is such a high DC voltage on the primary side. In terms of the original issue here (PSU not switching ON when PS_ON is pulled to ground), I'd say that's caused by something else.

My suggestion for now is to feed the PSU directly with 240V and not bother with the ISO transformer, just in case somehow that's the culprit of the HV (I don't see how, but regardless.) To make sure nothing blows up in case a more serious issue develops, use a 100-300 Watt incandescent/halogen bulb or heating element in series with the AC input going to the PSU (on the Live.) Then continue troubleshooting the rest of the PSU.

How are the caps on the output side, BTW? Did you check their capacitance and ESR yet? With the PSU being at this age, I suggest checking all of the caps on the output side - even the small ones. Actually, same goes for the small caps on the primary side too. Would be too easy (but not uncommon) if the whole issue of the PSU not working is caused by a bad "startup" cap on the primary. I see two small electrolytics next to the main traffo. Best to pull them out and check them.