Power supply broke

[pfrcom]

This was my reference power supply for testing P4 motherboards

Now it will not even start, even if I load its Molexes with a pair of old disks, and short PS-ON to GND with a paperclip

No bulging capacitors, no burnt stuff, fuse has continuity, 5VSB and PS-ON both read 4.95V on DMM

Can an electronics amateur troubleshoot and possibly fix it ?

[linuxguru]

Re: Power supply broke

If it has good +5Vsb, good primary switching transistors, and apparently good secondary caps: Check all the secondary rectifiers for shorts/opens.

[ratdude747]

Re: Power supply broke

hmm... teapos are usually ok in power supplies, but they might be in bad shape- teapos like to dry out without bulging.

is you fan ok?

[etnietering]

Re: Power supply broke

Does it make a quiet chirp or squeak sound when you try to turn it on? I had one that did, and it was the OCP shutting it down because of a bad output rectifier. Also, go over the solder side very carefully to check for bad solder joints.

Oh and I LOVE that engrish on the label! "As sealed stick was removed, lost or damaged, it shall be out of warranty validity!"

[pfrcom]

Re: Power supply broke

Fan was OK last time it ran (not that long ago)

No noise at all when I try turning it on, and no twitch from the fan either

[370forlife]

Re: Power supply broke

Agreed on the secondary schottky's. The delta in our dell had one short on the secondary and refused to power up.

[pfrcom]

Re: Power supply broke

Hope I've removed the correct components - does this heatsink contain the secondary rectifiers ?

If not, guess I've provided some amusement

On pic labelled EX CM RS-350, the components left to right are marked:

1. ON B20100 CQ344 AKA

2. ST STPS3045CW CC1CR VW MRC 423

3. ST STPS3045CW CC070 VW MRC 424

On other pic, L to R are:

1. F D0420 FDP 7030BL

2. ST STPS3045CW CC070 VW MRC 424

3. N (National Semiconductor logo ?) JM45AN LM7912CT


If I'm on the right track, what do I test, and how ?

I have an (original) Fluke 77 DMM that has a Diode Test function

[Toasty]

Re: Power supply broke

Dual Schottky's are essentially 2 high speed diodes (typically) in a cathode-to-cathode configuration at the center leg, the anodes being the outer 2 legs. You should read (about) the same from outer legs to center on all -3- of the *3045 devices in both directions. Black to center, red to outer(s) should be low. Reverse the leads and it should show open. The B20100 device values will be different, but should still give the same indications.

A zero or very low reading in both directions indicates a shorted unit. Same the other way with a infinity reading indicating an open unit.

B20100 - 20A/100v dual Schottky in a TO-220 package (datasheets only show D2PAK which has the tab and center leg chopped off to make it a surface mount)
https://cdn.badcaps-static.com/pdfs/...739ff38292.pdf

STPS3045CW - 30A/45v dual Schottky's.


FDP7030BL - MOSFET


LM7912CT - -12v regulator

[Toasty]

Re: Power supply broke

While you've got the heat sinks off, get some pics of the board on top. A top-down with some good light. Then a couple from some side angles. Something may jump out at one of us.

[PCBONEZ]

Re: Power supply broke

Something may jump out at one of us.

HEY MAN! - Halloween is over....

[pfrcom]

Re: Power supply broke

Have only removed one heatsink

Hope these fill the bill

[Toasty]

Re: Power supply broke

Did you get those devices on the heat sink tested yet?

Thanks for the PM!

Toast

[pfrcom]

Re: Power supply broke

Quoting from the Diode Test instructions for my Fluke 77:

"Forward bias: The meter displays the forward voltage drop (Vf) in volts up to 2V, and beeps briefly for one diode drop (Vf < 0.7V).

Reverse bias or open circuit: The meter displays 0L.

Short circuit: The meter emits a continuous tone."


I have tested the OnB20100, and the three STPS3045CW:

With the Red+ probe on the middle leg, there is no tone from touching the Black- probe to either the left or right leg


With the Black- probe on the middle leg, all give a beep when touching the Red+ probe to the left and right leg

For the OnB20100, the meter settles on a reading of .260, the reading is the same for both left and right legs

For the first STPS3045CW, the meter settles on a reading of .209, the reading is the same for both left and right legs

Same for the second STPS3045CW, the reading is .203

Same for the third STPS3045CW, the reading is .206


As a Go / No Go test, I assume that means all the secondary rectifiers are Go ?

[Toasty]

Re: Power supply broke

Yes, it seems that way. Your trouble(s) lie elsewhere.

No testing on the 7030BL?

Here's one way: http://www.4qdtec.com/mostest.html

[pfrcom]

Re: Power supply broke

Thanks again for your patience and persistence Toasty

With the meter set to Diode Test, I touched its Black- probe to the right lead of the 7030BL

Then touched the Red+ probe to the 7030BL's left lead

Then changed the Red+ probe to the 7030BL's centre lead, meter gives a continuous beep and reads .000

If I repeat the procedure with the probes the other way around, when I change the Black- probe to the 7030BL's centre lead, meter gives a single beep and reads .537

Could this mean the 7030BL is the culprit ?

[Toasty]

Re: Power supply broke

You missed the finger-touch step. Does it go high when you do that? If so, then it's probably okay.

Check the device(s) on the other heatsink. That's the hot side so make sure those 2 big mains caps are at zero volts before you go working there.

Just because the caps -look- good, does not mean they are.
Do you have a supply of caps to work with?

What brand are the output caps?
Do you have an ESR meter? Capacitance meter?

[pfrcom]

Re: Power supply broke

You're right about missing the finger touch - I thought the beeping might indicate the end of the quest

And dammit, when I touch my finger to the leads, the beeping stops after a second or so, and the meter reads various values approaching 2.000

The output caps are Teapo (560uf, 200V)

I do have a Dick Smith ESR meter - guess it's time to start playing with that

[Toasty]

Re: Power supply broke

Okay. That reading is 2 ohms? That doesn't sound correct by the test procedure link. It says the "meter reading should go high". 2 ohms doesn't sound high. Did you mean to type 2,000 ohms?

>>The output caps are Teapo (560uf, 200V)<<
No, the OUTPUT caps. The ones down where all the wires come out. Those 200v caps are the MAINS caps.

Again -
Do you have a supply of caps to work with?

Now you tell us you have an ESR meter...

Should have used that First on ALL those OUTPUT caps. Check everyone that has the score marks "X" or "+" on the top.

Readings of caps of 1000uF and up should be very low - 20 milliohms or .02 ohms or less. I don't have any experience with that meter. I think you can test them without removing them. Check the instructions.

Get checking

[PCBONEZ]

Re: Power supply broke

You can check them without removing if they are not in parallel with other caps.

[pfrcom]

Re: Power supply broke

Thanks everyone who contributed to this, especially Toasty for your patience which I wore a bit thin

Majority of the output caps are Ltec, with some of the smaller ones being Teapo

For the 7030BL test I recall the readings jumping around 1.9xx , cannot re-test because I've soldered it back together

I cannot find any abnormal readings with the Dick Smith ESR meter, but I have had the same experience on motherboards with bulging caps, so maybe these ones are in parallel

Although I do have a limited supply of caps, I think I've probably wasted enough of my life on this P.O.S. which will probably end up in the garbage where it should have gone when it failed without any prior abnormalities

I've bought an oem FSP ATX-400PN to replace the CoolerMaster RS-350-AMSR

From what I can see through its vents, the FSP is all Teapo - all its smaller caps have the typical three legged vent scoring, with an additional hyphen in each of the three segments