Sorting out old PC power supplies?

[roadrash]

Originally Posted by PeteS in CA

The color coding looks like standard ATX, except I don't see a white -5V wire. I don't know if that had become common back in 2007. Look at the location on the ATX connector where the -5V wire normally would be. If it's -5V or there is no wire to the location, well and good. However in the late 90s and early 2000s I know (former Delta employee, here) Apple used the -5V location for an extra 3.3V socket. So if there is a wire and it's orange, you have an Apple power supply.

Your right I cannot see a -5v wire but there is a orange 3.3v one. Its not a apple.

[roadrash]

Originally Posted by PeteS in CA

The point at which I suggest starting is whether or not you have 5VSB, the purple wire. If no 5VSB then that's where you start looking.

If you have 5VSB and grounding the PS-ON (green) wire doesn't turn on the main outputs then the PWM circuit is your starting point. Check the Vcc voltage on the 384x. If it's too low then there's probably a ~47uF cap that might be bad, or the resistor that charges it might be bad. Those provide the voltage to start up the 384x. One the PSU starts up the 384x Vcc usually is supplied by a bootstrap winding.

I dont get any DC voltage on 5VSB. I have a ATX PSU tester plugged into power cables so should be turned on. What precisely is the 384x you mentioned. Your dealing with a hard trying newb here remember.

[stj]

8pin chip

[roadrash]

Originally Posted by stj

i suspect it was 150v, although that may be too low for a 240v input.
maybe that's why it blew up.

total rectified voltage is going to be about 340v so divided by 2 it's 170v
so using parts rated between 200 and 250v is probably wise

btw, the psu will run without them - so if it's dead it probably blew the fuse or killed an input filter

Just out of interest which ones would you pick from the links below:

VISHAY-EPCOS-METAL-OXIDE-VARISTOR-150v-QTY-5/123787501997

3-Metal-Oxide-Varistor-MOV-20D431K-275VAC-120/121769194537

Varistor-76-0J-300V-ac/391288103662?

Varistor-38-0J-250V-ac/391741394396?

[roadrash]

Originally Posted by stj

8pin chip

There two 8 pin chips in this thing but only one with 384 in its number. The one in this photo I hope.
I was unsure where to put the negative probe so put it on the "-" of the rectifier.
Doing this I didnt get any voltage reading.

[stj]

250v is my choice, pay attention to the diameter.
bigger is better - if you have room for it!

[R_J]

Originally Posted by roadrash

I dont get any DC voltage on 5VSB. I have a ATX PSU tester plugged into power cables so should be turned on. What precisely is the 384x you mentioned. Your dealing with a hard trying newb here remember.

The standby +5v is likely supplied by the smaller transformer and it's primary is driven by a mosfet / single transistor circuit not by the 8 pin ic's
One pin of the transformer primary will be connected to the main cap +. The other pin of the primary connects to the mosfet Drain, and there should be a low ohm resistor from the fets source pin to hot ground (main cap -)

[roadrash]

Originally Posted by R_J

The standby +5v is likely supplied by the smaller transformer and it's primary is driven by a mosfet / single transistor circuit not by the 8 pin ic's
One pin of the transformer primary will be connected to the main cap +. The other pin of the primary connects to the mosfet Drain, and there should be a low ohm resistor from the fets source pin to hot ground (main cap -)

Thanks for that RJ.
I did try to see if I could get a connection between the Mosfet (drain and one of the connections on that small transformer but couldn't get anything. Anyway I removed the Mosfet (WN9K90Z) for testing and testing it with a multi-meter in diode mode, I got a open circuit between source and drain but if I touched the positive lead to gate to charge it the source and gate still remained open circuit which is not right. So I put my component tester on it and it correctly identified it as a Mosfet (see photo) so that must mean its OK I think.
Did anyone find a schematic for this PSU, I couldnt?

[PeteS in CA]

Originally Posted by roadrash

Your right I cannot see a -5v wire but there is a orange 3.3v one. Its not a apple.

If it has an orange wire in the -5V location in the 20- or 24-pin connector it is incompatible with MBs that use the standard ATX pin-out. The only computer company I've seen use that location for an extra 3.3V wire is Apple. Like I posted, I used to work for Delta Products. I was in tech support here in Silicon Valley, and Apple was one of the customers I supported, back in the late 1990s. That is when I saw Apple using the -5V location for an extra 3.3V wire. My current employer is also a computer company, and our 20- and 24-pin ATX connectors either have -5V in that location or the location is empty.

Originally Posted by roadrash

I dont get any DC voltage on 5VSB. I have a ATX PSU tester plugged into power cables so should be turned on. What precisely is the 384x you mentioned. Your dealing with a hard trying newb here remember.

The UC3842, UC3843, UC3844, and UC3845 are a series of 8-pin current-mode pulse width modulator ICs. Which specific part a company uses depends on the specific design. While I remember that Delta used parts from that series, I don't remember which part Delta most commonly used (possibly the 3845), so I substituted an "x" for the digit I could not recall.

Returning to my comments above, once you figure out the problem with the 5VSB circuit, if you have a 3.3V wire in the -5V location, the power supply is not compatible with your ATX PSU tester.

I'm pretty sure Delta used the Power Integrations regulators for their 5VSB circuit, possibly from this series, https://www.power.com/sites/default/...tny274-280.pdf or this series, https://ac-dc.power.com/sites/defaul..._datasheet.pdf . If you look at the sample circuit on page 8 of the first data sheet, look for components equivalent to R5, R3 & R8, C6, and VR2 and check them out. Those have to do with the the start-up and operation of the IC.

[roadrash]

Thanks for that Pete, I think this is standard atx as I just compared its atx connector pinouts with a picture online and it shows one position empty which is the -5v line.
Here is a photo of thhe 20 pin connector. I have 6 of these desktops to fix the domed caps and ive done 4 of them and they are all working. This is one of the remaining 2 psu's' to fix.

[roadrash]

This redistor doesn't make sense.
Its colours are:

Brown
Green
Silver
Gold
White

I measure it at 0.76 ohms
But this combination of colours doesnt make sense on any 5 band calculator.

[stj]

it's wirewound - the fact that it meters low ohms means it's o.k.

[roadrash]

Originally Posted by stj

it's wirewound - the fact that it meters low ohms means it's o.k.

I had a feeling it was that as its was mention a bit earlier about a low value resistor in that area. Why doesnt it match a resistor colour chart or was i reading wrong?

[petehall347]

Brown
Green
Silver
Gold
0.15 ohms

[PeteS in CA]

Originally Posted by roadrash

Thanks for that Pete, I think this is standard atx as I just compared its atx connector pinouts with a picture online and it shows one position empty which is the -5v line.
Here is a photo of thhe 20 pin connector. I have 6 of these desktops to fix the domed caps and ive done 4 of them and they are all working. This is one of the remaining 2 psu's' to fix.

The -5V location is empty, so it should be OK with a MB that doesn't use -5V and with your tester.

[R_J]

I believe the mosfet you removed is for the main supply not the standby, the transistor for the standby circuit is near R905

[goodpsusearch]

Originally Posted by momaka

It's not a cap, it's an MOV - completely different component.

As stj mention, sometimes they can also fail with time as they absorb high voltage spikes on the line.

If each of these is placed in parallel with the two 200V caps, then the new MOVs should have a rating of 180-220V. 180V may be a bit too low, though.

But again, these are not necessary,

chinese psu manufacturers would stop reading right here and press the like button

[goodpsusearch]

The incandescent lamp in series would have saved one of the psus from blowing. You can google it and always connect it in series with the psu when you test psu like this in the future.

[roadrash]

Originally Posted by R_J

I believe the mosfet you removed is for the main supply not the standby, the transistor for the standby circuit is near R905

Thanks RJ for that. That transistor is a smd one? I haven't replaced one of those yet but should manage it if I now what type it is and can find one.
Can i test it in place before changing it?

[R_J]

The small smd transistor drives a larger mosfet that has it's Drain pin connected to the small standby transformer, I made a Red square around the standby mosfet and a yellow one around the drive transistor.
On the small standby transformer, pin 1 should have 160vdc, and so should pin3 which is connected to the mosfet drain., this is the primary. There is likely a low ohm resistor connected to THIS mosfet's source pin to hot ground.