OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

[Agent24]

Was given this PSU from someone when it stopped running their machine.

Symptom was that their Asrock motherboard would not POST (onboard status LEDs said FF) but worked with another PSU. Weird thing was, the PSU ran fine in a machine I had, but not another. All voltages tested normal.

Found that the Power_Good\PS_OK signal was not present. Guess one of my PCs doesn't care... FF error made sense now, his board was being held in reset state. Anyway...


Check inside, found monitor IC a PS224 (see datasheet).
No signal direct from Pin 16. Checked all other pins. All seemed OK or I was unsure but PGI Pin 1 obviously wrong at 0v, should be 5v going by example circuit.


Traced back Pin 1, ends up at 12v rectifier output. Found burnt diode and resistor D308 and R315.

Checked all other components between R315 and Pin 1 for shorts but found nothing.


Any ideas why D308 and R315 burned? Any ideas on replacements? R315 I'm guessing is 10 Ohms, don't know about D308.

[stj]

Re: OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

doesnt that resistor lead to c314?

maybe a shorted cap?

and the resistor may be readable if you wip the soot off with some ipa.

diode - no idea, you could use an smd version of a 1n4148 (LS4148)

[Agent24]

Re: OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

That was my first thought, but I tested the resistance across C314 and got several kOhm, I expect from resistors R906 and R902 in parallel with C314, so it does not appear to be shorted.

Unless perhaps it shorted, then burnt out?

[stj]

Re: OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

doesnt look burned.
not straight either, poor assembly!

[stj]

Re: OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

btw, it looks a bit familiar.
does the pcb say "GP" on it?
if so,it should have the fuse type and psu wattage maked - be interesting to know if it's really 600.
i have a similar board in an "i-cute" unit labelled 700w.
the pcb is clearly marked 420w!!!
(model "REX-420A")

[Agent24]

Re: OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

I found the same problem here: http://radiokot.ru/forum/viewtopic.php?f=34&t=68343

Seems his C314 shorted and replacing the three parts fixed it. I will try C314 out of circuit tomorrow, check capacitance or see if it has visible damage anywhere (maybe on underside?)

Pity half his images on a 3rd party site have broken links...


Can't see "GP" obviously on it. Where is it likely to be? I think it looks like a popular design though since it appears to be also used in that Chieftec APS-650C.

I saw what I think says "NO-636" though.

Fuse silkscreen says T10A-250V - 10 Amp? WTF? That's the same as the fuse in the meter box...

[Agent24]

Re: OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

Checked C314, it's not shorted, and reads as 200nF on the meter, so I guess it's OK. But it looked like there was a slight dent\chip or something in the side, so I decided to replace it anyway.

I realised that at 12v with a 10 Ohm resistor the current to charge C314 is going to be 1.2 Amps - perhaps this thing blows because of a high inrush current? Then again the capacitor is quite small so probably not...

Replaced D308 with a 1N4148 and tried a 100 Ohm resistor for R315 as that was what I had lying around, but I think it may be too high since I only get 1.20v on Pin 1 of the PS224 chip and I still don't get Power_Good. According to that Russian site it needs at least 1.25v to give the Power_Good signal.

I'll try to find a smaller value resistor and see what happens.

[Agent24]

Re: OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

It works!

Put another 100 Ohm resistor in parallel and now I get 1.3v at Pin 1 and a nice 5v Power_Good signal.

Had no SMD parts so I just bodged in some normal ones...

If it keeps working I'll get rid of the Teapo infestation.

[stj]

Re: OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

lol
nice trick with the cap - i think that one has a higher ripple current!!!

[momaka]

Re: OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

Nice fix!
Yeah, I hate SMD parts in a power supply. But what PSU doesn't have them these days?

May I also suggest replacing all of the caps for 5VSB?

[Agent24]

Re: OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

For testing yeah they're a bit annoying, can't just lift one leg, have to desolder the whole thing and then try to get the probes on a tiny little thing that tries to run away... and then it's a pain putting them back since with the leftover solder they don't lie flat.

Would have used SMD parts if I had any though, but that's what I had, that's what I used.

Seems OK. Ran it for about 3 hours this afternoon, no problems so far.

[Heihachi_73]

Re: OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

Another show pony for the uber leet community, with cheap-ass Teapos inside. ModXStream indeed.

[Rulycat]

Re: OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

Didn't the OEM of this use TRec caps before they used Teapo? If so, Teapo is an improvement.

[cheeky2]

Re: OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

Just had a OCZ700MXSP with the same pcb inside with exact same problem.

Resolved with replacing R315 100ohm (mine open cct) and replacing D308 with a 1N4148. I have attached the diagram from the russian site linked earlier, just so that it makes it easier to understand the schematic operation.

I originally replaced the PS224 IC with no joy however after checking the pins found the PGI low where should have been around 1.35V. I thought that pin16 had blown its output when all the psu outputs were within specs.

So thanks for this info I had to take a close up of PCB to see resistor blown and diode looked ok until removed from cct to test!

[Agent24]

Re: OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

Good to hear! Seems a common problem with this design but who knows what exactly happens or why...?

So far mine is still working after the repair, though.

[Behemot]

Re: OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

I got 700W version of this, it is slightly different as it mostly has non-SMD parts on this feedback so the layout it also changed. It is Rev-A6 from 7. 7. 2010.

Anyway, this one had bulging primary cap, I've replaced it but still no joy. Fan only twists. All the components you got problems with are OK. Any idea? Stand-by seems to work OK, ligh bulb does not indicate any short on primary. The C314 in my case is film cap 224J100, I guess that's 220 nF for 100 V.

[Agent24]

Re: OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

If the primary capacitor bulged maybe there is still some problem in the primary side - either causing the bulged capacitor, or caused by it.

Perhaps the APFC has some issue? Enough power gets through to 5vSB circuit but it won't start normally?

Have you checked for shorts on main outputs?

[Behemot]

Re: OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

Yep all the rectifiers measure OK. There was also a cold joint for the -12V diode but resoldering that did not help. PFC section also measures OK (or, at least not shorted). Only if the ICs were bad…

The cap may have been bulging just because it was crap, Teapo or Crapxon, don't remember…

[Agent24]

Re: OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

Yeah but as you know, if the capacitor dies, the rest of the circuitry has to make up for it. If the capacitance dropped the PFC circuit might have overworked itself to death trying to keep the voltage up.

[Behemot]

Re: OCZ OCZ600MXSP ModXStream-Pro 600W - Runs but no Power_Good, burnt components.

In such case it usually shorts, at least the power silicon, which I think is exluded both by no problem according to my DMM, and according to the lighbulb in series. At least the power diodes there are OK, those I can measure easily.

I'll also measure the small caps later, just to be sure.