Xigmatek 500W ATX PSU went POP and stopped working

[dicky96]

This ATX PSU is off a customers computer. Originally it had a faulty motherboard GA-X58-USB3 that had a VRM fault. I fixed that and returned it working, but when he plugged it in something went 'pop' and it went dead.

I have the PC working again using my test ATX PSU so I decided to look at his PSU

Its a Xigmatek 500W ATX

When I opened it up the whole thing was full of dust, dirt and fluff.

The in-circuit fuse on the 220V input has blown

I cleaned it all out, then removed the heatsink with the Output MOSFETS that drive the main SMPS transformer. They test OK, but I noticed a small wire (looks like a component lead) down between the heatsink and the SMPS transformer and it is just sticking up in the air and doesn't do anything (see pics) I can't figure out what this is supposed to do.

The Output MOSFETS tested OK so I then removed the other large heatsink with the PFC MOSFET(s) and the main PFC circuit Diode - at least that is what I think it most likely is.

Now here is the interesting thing, one of the PFC MOSFETs looks like it has blown, in so much as the Drain and Source legs look partially melted where they enter the body, and there is some evidence of blast 'deposit' on the heatsink behind it - but the odd thing is it tests OK as an N-Ch MosFET same as it's comrade in arms next it. They are both 18N50.... But it definitely looks like it's blown even though it test just fine.

See pics.

And any suggestions? My guess is the dust/dirt/fluff went conductive around the device legs and that is what went pop, not the divice itself - and if I put it all back together it might even work - but if that is the case I have to say I've never seen that before in over 40 years messing with electronics.

Then there is that wire sticking up doing nothing useful that I can think of

Suggestions?

[stj]

Re: Xigmatek 500W ATX PSU went POP and stopped working

they look like salvaged mosfets!

[dicky96]

Re: Xigmatek 500W ATX PSU went POP and stopped working

That may be how it looks but I'm 100% sure they are the originals as there was no sign of any previous rework

[R_J]

Re: Xigmatek 500W ATX PSU went POP and stopped working

That short jumper wire looks like it was soldered into that hole closest to it and one end was unsoldered for some reason. Does the solder side of the hole look like it was soldered?

[dicky96]

Re: Xigmatek 500W ATX PSU went POP and stopped working

No. it's just a hole, no pad. It looks like it was/is/could be used to short out a small 10R SMD resistor (R11) on the other side, because the end of the wire that is soldered actually also passes through a hole with no pad, and is then soldered to one end of the resistor. it's not 100% obvious whether the other end was ever soldered or not. It certainly hasn't broken off but it could have been deliberately unsoldered.

I still don't see the point of it, why not just fit a 0R resistor in place of the 10R resistor if you wanted it shorted for some reason?

[dicky96]

Re: Xigmatek 500W ATX PSU went POP and stopped working

Hmm no it doesn't 'just work' once I put it back together (minus all the dust and fluff)

With a light 220V bulb in series, the lamp lights up quite brightly - I would expect it to light then go dim/off with the PSU in standby

I only get around 0.3V on the SB supply and about 25V on the main electrolytic after the Bridge/PFC so I am guessing that the dodgy looking 18N50 in the PFC is breaking down under load. The two 18N50s look to be in parallel, with a 20R resistor on the gate of each, driven from the same pin on the controller

These 18N50 are very cheap on Aliexpress so I think I will get some to try to fix it, just out of interest really.

[MHTSOS]

Re: Xigmatek 500W ATX PSU went POP and stopped working

Those 18N50 on Aliexpress are 100% fake. If you measure their RDSon you'll see its much higher than normal. It's better to salvage the mosfets from another PSU or buy from a legit supplier
.

[Sparkey55]

Re: Xigmatek 500W ATX PSU went POP and stopped working

This ATX PSU is off a customers computer. Originally it had a faulty motherboard GA-X58-USB3 that had a VRM fault. I fixed that and returned it working, but when he plugged it in something went 'pop' and it went dead.

I have the PC working again using my test ATX PSU so I decided to look at his PSU

Its a Xigmatek 500W ATX

When I opened it up the whole thing was full of dust, dirt and fluff.

The in-circuit fuse on the 220V input has blown

I cleaned it all out, then removed the heatsink with the Output MOSFETS that drive the main SMPS transformer. They test OK, but I noticed a small wire (looks like a component lead) down between the heatsink and the SMPS transformer and it is just sticking up in the air and doesn't do anything (see pics) I can't figure out what this is supposed to do.

The Output MOSFETS tested OK so I then removed the other large heatsink with the PFC MOSFET(s) and the main PFC circuit Diode - at least that is what I think it most likely is.

Now here is the interesting thing, one of the PFC MOSFETs looks like it has blown, in so much as the Drain and Source legs look partially melted where they enter the body, and there is some evidence of blast 'deposit' on the heatsink behind it - but the odd thing is it tests OK as an N-Ch MosFET same as it's comrade in arms next it. They are both 18N50.... But it definitely looks like it's blown even though it test just fine.

See pics.

And any suggestions? My guess is the dust/dirt/fluff went conductive around the device legs and that is what went pop, not the divice itself - and if I put it all back together it might even work - but if that is the case I have to say I've never seen that before in over 40 years messing with electronics.

Then there is that wire sticking up doing nothing useful that I can think of

Suggestions?

I have seen that same green crystaline glue dawbed all over the leads of PSU fets and rectifier packs before. Usually to hold down ferrite beads. i have not as yet seen any conductivity issues with it. But you never know. BEWARE of that yellow glue that turns brown/black, that is conductive.

[momaka]

Re: Xigmatek 500W ATX PSU went POP and stopped working

BEWARE of that yellow glue that turns brown/black, that is conductive.

^^ THIS ^^
Clean out ALL of the tan/beige glue from the primary side - especially from the APFC coil and on any traces (especially high-voltage ones.)

Also pull the big mains 400V CapXon cap and check its ESR and capacitance. Better yet, just change it. APFC does NOT like cheap cap brands and tends to roast them after a while. Heck, even good Japanese brands tend to get roasted in APFC circuits. A lot of that has to do with cost-cutting too. 400V caps, especially GP 85C ones really aren't adequate for APFC circuits. Use a 420 or 450V cap rated for 105C and preferably something that has a higher ripple current rating. The combination of high frequency ripple and high voltage is really what roasts most PSUs with 400V-rated caps. Adding a 630V 0.1 to 0.47 uF ceramic or 1-3 uF film cap may help ease some of the high-frequency ripple from the mains cap (and will also provide a safe place for the APFC coil to dump its load, should the primary cap fail open-circuit.)

[dicky96]

Re: Xigmatek 500W ATX PSU went POP and stopped working

I like to tidy up my threads so...

Sorry I didn't see what MHTSOS put until after I had ordered some el cheapo 18N50s
Sorry I didn't see what Momaka put until after the 18N50s arrived from AliExpress

So anyway I fitted the new MosFETS in the PCF circuit, left my light bulb in series and the PSU powered up and I have 5V Standby with the light bulb flashing once them dimming out to nothing. So I connected a load (old faulty HDD and a 12V fan), forced the ATX PSU to 'on', bypassed the light bulb.... and it powered up and ran just fine.

So it seems that the doubts about cheap fake 18N50s and capacitors in this case were not applicable

Actually about the cheap FETs from Aliexpress - the ones in the pic on the listing look like some generic make, but the ones I received were Fairchild branded.

https://www.aliexpress.com/item/4000...3da24c4dku4X7Z

They all test OK on my DCA55 - they all have the same Gate threshold 4.39V (where as the original Magnachip read 4.42v) and they seem to be good.

Rich

[sam_sam_sam]

Re: Xigmatek 500W ATX PSU went POP and stopped working

I would recommend that you keep a close eye on it for a while just in case it takes a crap again

[dicky96]

Re: Xigmatek 500W ATX PSU went POP and stopped working

Yeah I will. I already fitted a replacement repaired PSU to the customers computer so this one is back on my shelf of parts. I'll use it as a test PSU for a while and see how it goes.

[momaka]

Re: Xigmatek 500W ATX PSU went POP and stopped working

So it seems that the doubts about cheap fake 18N50s and capacitors in this case were not applicable

Well, not so fast!

You will know if those 18N50 are up to snuff only once you give the PSU a more serious load, like a PC with a power-hungry GPU and CPU... or more ideally, a full-rated load according to the label. The problem with cheap MOSFETs is not that they don't work (they usually do), but rather that they tend to blow at a much smaller load than reputable-made ones.

And as for what I said about the primary Capxon cap - that still stands true. Even if it isn't bad now, it just won't last very long. I know those big caps are expensive and make the repair almost not worthwhile... but when (not *if*) that cap blows, it will take out the APFC MOSFETs, and possibly a good deal of other primary components as well (APFC controller, typically, among other things.) And that will make the PSU certainly not worth a repair again. Not to mention there is also a risk of damage to any connected PC components.

[sam_sam_sam]

Re: Xigmatek 500W ATX PSU went POP and stopped working

If you want to be on the safe side and you have the money to invest in a electronic load tester would be the ideal way to test this power supply

I have tested a few switching power supply’s with this electronic load tester and I have been impressed with the results that I gotten with this tester

I have one power supply that I want to know if it would put out the amperage that it has on the data plate it does not come even close and the switching power supply case got real warm in a hurry with half of the rate power that the power output would stay on ( what crap )

I would highly recommend one if you are going to repair switching power supply on a regular basis

I just bought one a few months ago and tested a switching power supply at full load and now it will not power on the only power I have is the stand by power of 12 volts this not a ATX power it one that only has 12 volt output

As of yet I have not really had time to troubleshoot this problem but I have modified this switching power supply so the power output connectors are not on there any more so I have to take another power supply like this one and remove the power connectors to see if the fan controller that reports to the computer board has something to do with it not powering up because it has done this before

I have a ATX switching power supply that I going to test in couple of days and see if it put out what the data plate says it should put out

[dicky96]

Re: Xigmatek 500W ATX PSU went POP and stopped working

Hi Sam, Momaka

OK, I'm up for some fun.... let's try to do the worse to it

One thing puzzles me from your post Sam, is there supposed to be a url link to the electronic load tester that you bought?

Cheers

[sam_sam_sam]

Re: Xigmatek 500W ATX PSU went POP and stopped working

Hi Sam
One thing puzzles me from your post Sam, is there supposed to be a url link to the electronic load tester that you bought?

Cheers

Here is the link to the eBay sellers listing

https://www.ebay.com/itm/30A-Dual-Ch...53.m2749.l2649

It has gone up in price scenes I bought it back in September of last year

One note I did not find the CV mode to be very useful to me or I really do not understand how to setup it correctly or how to use it correctly ( it might be useful for battery charging but I have not tried it )

I have only use the CC mode for testing switching power supply

[dicky96]

Re: Xigmatek 500W ATX PSU went POP and stopped working

Thanks. That's quite expensive from the USA and does not ship to Spain, but I can find the same model on Aliexpress with good feedback and a much more attractive price in Euros including postage.

[sam_sam_sam]

Re: Xigmatek 500W ATX PSU went POP and stopped working

I know that AliExpress has them but I have the opposite problem a lot of the time the wait time is a lot longer and cost more most of a of the time but a bigger problem is a lot of sellers do not take PayPal as payment

[dicky96]

Re: Xigmatek 500W ATX PSU went POP and stopped working

At the moment Aliexpress is being really fast delivery wise for me for some unknown reason - generally about 10 days.

I normally pay Aliexpress orders by credit card. They do take paypal but more so on Alibaba where they have this 'trade assurance' thing.

[momaka]

Re: Xigmatek 500W ATX PSU went POP and stopped working

I know that AliExpress has them but I have the opposite problem a lot of the time the wait time is a lot longer and cost more most of a of the time but a bigger problem is a lot of sellers do not take PayPal as payment

I have not ordered anything from AliExp, but I've heard a lot of bad stories about orders - at least here in the US. The main problem seems to be buyer protection, should you get a non-working item or no item at all. eBay makes it easier in that regard (and sometimes even sides a bit too much with the buyer than they should.)
So looks like in the USA, eBay is a better option, and in EU, AliExp might be better... i.e. pick your poison.
On that note, I avoid Amazon. Returns may be the easiest, but a lot of items are overpriced.

At the moment Aliexpress is being really fast delivery wise for me for some unknown reason - generally about 10 days.

Same here with eBay.
I guess people are done sending their gifts in the mail, so it seems like postal services are back up and running quick again. I just ordered a pair of USB 2.0 to 3.0 header cables from China, and they arrived in a little over a week. Wasn't from a US warehouse, pretty sure - those items usually arrive in 5 days or less.