Question So what blew here ? *** Updated, replacement online ***

[Markfw]

/
The system turned off for some reason when I started it up, I saw smoke/sparks and I pulled the plug. The PSU tester says the PSU is good, but what is this part that blew ?

This is a X399 Taichi motherboard with a 2970wx at stock, and 4 x 16gig DDR4 3200

 

[Markfw]

Its on the underside of the motherboard, right under the VRMs for the CPU.. That black screw you see holds the large heatsink on the motherboard, with a heatpipe to the other heatsink .

 

[mindless1]

That's a solid capacitor, probably tantalum polymer, probably developed an internal short. It's in parallel with the others next to it, so if the board isn't under warranty, you could desolder it and see if it works again. I wouldn't put it under extreme load missing the cap but at lower power states to check for life, should be fine without it.

If it's still alive, here's a potential replacement if it has the same dimensions: https://www.digikey.com/en/products/detail/panasonic-electronic-components/16TDC150MYF/10644021?s=N4IgjCBcoEwAwBYCcVQGMoDMCGAbAzgKYA0IA9lANrgCscABAK0BiIAuqQA4AuUIAqgDsAltwDymALKFs+AK4AnQiAC+peAHYU0EBkg4CJclXAA2egDV2XXpAEjxUmfKWq1IGCYAK2QbLIiaPQAKgAiAML0aNic2GiiZArsKkA

However, if there are multiple circuit layers, the failure could have caused damage that shorts those out too, so I'd definitely test the board w/o the cap on, before getting a replacement.

 

[Markfw]

That's a solid capacitor, probably developed an internal short. It's in parallel with the others next to it, so if the card isn't under warranty, you could desolder it and see if the card works again. I wouldn't put it under extreme load missing the cap but at lower power states to check for life, should be fine without it.

Except it will be running 24/7/365 at 100% load on that 2970wx. I ordered a new motherboard, just hoping that nothing else blew. I did not have my cochlear implant on at the time, so no idea on noise, but the sparks and the smell, very bad.

 

[mindless1]

I doubt it was very loud, regular, non-polymer tantalum caps can explode and there's nothing left, but the casing on that is intact, for the most part.

 

[Markfw]

I doubt it was very loud, regular, non-polymer tantalum caps can explode and there's nothing left, but the casing on that is intact, for the most part.

Were those good quality ? It was about 3-4 years old. The replacement motherboard was $400, and others were $600. I only paid $300 new, but I guess its the same story as video cards.

 

[mindless1]

-duplicate post-

 

[mindless1]

I guess the one that blew up wasn't such great quality but video cards have been using them for around 20 years now. I recall first seeing them on Geforce 3 TI 500 cards.

Failures aren't common but when they short, can be a very sudden, low resistance short when used in a high current VRM circuit, otherwise they can self-heal to some extent. The industry seems to favor them as an upgrade over aluminum polymer solid caps, generally reserving them for higher end gear. They have better high frequency performance and lifespan, though aren't really necessary to achieve most PC market design goals, including those of the enthusiast overclocker/etc, it's more of a luxury item and a bit crazy that motherboards have come to this, but if you want to o'c, all bets are off.

 

[Markfw]

OK, the new motherboard (x399M Taichi) is here, and everything is working, thank god ! So I am out $400, otherwise no worse for wear, except I lost 36 hours of 48 WCG threads.

 

[Assimilator1]

Good to hear the rest of the gear is ok , and I'm amazed it is too!

 

[Markfw]

That's a solid capacitor, probably tantalum polymer, probably developed an internal short. It's in parallel with the others next to it, so if the board isn't under warranty, you could desolder it and see if it works again. I wouldn't put it under extreme load missing the cap but at lower power states to check for life, should be fine without it.

If it's still alive, here's a potential replacement if it has the same dimensions: https://www.digikey.com/en/products/detail/panasonic-electronic-components/16TDC150MYF/10644021?s=N4IgjCBcoEwAwBYCcVQGMoDMCGAbAzgKYA0IA9lANrgCscABAK0BiIAuqQA4AuUIAqgDsAltwDymALKFs+AK4AnQiAC+peAHYU0EBkg4CJclXAA2egDV2XXpAEjxUmfKWq1IGCYAK2QbLIiaPQAKgAiAML0aNic2GiiZArsKkA

However, if there are multiple circuit layers, the failure could have caused damage that shorts those out too, so I'd definitely test the board w/o the cap on, before getting a replacement.

OK, I can get a 1920x for $135, and then remove the bad cap, and test ?? What do you think ?

 

[mindless1]

^ Seems like a plan, though I hate for you to keep buying more parts in case the board is shorted where the cap burnt. I'd wait on the purchase until you see how the PCB held up.

I would first minimially clean the area, then desolder the capacitor, then thoroughly clean and examine the area, something that would have been easier if it wasn't black. Top black coating gone isn't a problem but molten or delaminated copper could determine if there should even be a new cap put back on (if the copper is lifting off then I'd either leave the cap off or super-glue the copper back down to put a new cap on), and ultimately any areas with bare copper would have lacquer, or epoxy, or even nail polish put over it instead of staying exposed, after soldering on the new cap if it gets one.

That area is a lot of thick copper, will need a good soldering iron. It'll probably be easier to start on the burnt end, may be able to just bend the other lead, the whole thing up while it is molten. If not, can work it up a little from each end, or a hot air station could heat both ends at once, or tweezers (type of iron).

 

[Soulkeeper]

I doubt the capacitor was the initial cause of the failure.
Without seeing a detailed picture of the otherside of the board(heatsinks removed). I suspect it could be the mosfet or other vrm component on the otherside.
It's also possible that a case screw got dropped on the motherboard tray and shorted out to that + leg, That positioning is horrible. That overheating around it looks like the pcb melted, some traces could be blown or compromised internally.

 

[mindless1]

Other component failures are not very likely to blow that cap, severe overheating might but a sudden shorted fet wouldn't.

 

[Soulkeeper]

You're probably right, but everything should be inspected.
Caps are rated at a certain Z, if the VRM driver fails to operate in that range it can blow a cap like that(ie: it'll overheat, current runaway).
The hotter it gets, the higher the resistance/heat, which in turn causes it to get hotter.
With +/-10% ratings on the caps, the weakest one of the set is likely to go first(it can have a higher Z/MHz than the others). The cap then exerts increased load on the mosfets/driver/traces and potentially takes another component with it.
Yes, probably unlikely but anything could have caused it.

This is why I like to turn the frequency settings for the vrm down in motherboard menu's ie: "normal" "medium" or "regular".
No point in pushing the mosfets and caps to the edge. If you push everything up to the "extreme" or "high" settings in the bios, then run 24/7 load this stuff is not so unlikely.

Definitely worth looking at the other side of the board imo. And at the very least doing resistance/continuity tests on the traces and such.