Burning smell form PC, how to check PSU/other parts?

[vanGenne]

Hey all,

 

During gaming, my rig suddenly died followed by a smell of burned electronics. Always a great sign. 

 

I've identified that my GPU caused the smell, but I got my rig to boot properly using the integrated graphics after a BIOS reset. My question is, what the heck caused it to short/burn like that? Is my PSU at fault here? Can a dust buildup cause this? Is there any way to test this at home? I don't want to replace the GPU before I know it won't happen again. 

 

Specs: 

MoBo - AsRock Z87 Extreme4
GPU - EVGA GeForce GTX780 SC (stock)
CPU - i5-4670K (OC to 3.7)
PSU - Corsair CS750M
RAM - 4x4GB Corsair vengeance DDR3
Storage - 1x Barracuda HDD 2TB, 2x SSD (1 for windows, 1 for games)

The entire build is 5.5 years old, one of the SSDs and 2 sticks of RAM are newer. Not that that matters probably. 

[xARACHN1D]

I looked at what PSU you are using. Something that reliable should not have caused it unless you got a faulty PSU. Now here is my question... I see no fans on this list. Did you actually not put fans or did you just accidentally leave them out. Could you also let me know what case you are using and what fans you are using if you are using any.

[Fasauceome]

35 minutes ago, xARACHN1D said:

I looked at what PSU you are using. Something that reliable should not have caused it unless you got a faulty PSU.

well, they're on the old side, it's possible that it's just expired

[noyo87]

I'd personally replace the PSU. Even if you trust a specific model, shit can happen. No forum member can guarantee you the state of your PSU.

 

Personal again, I just had bad experience with Corsair PSUs (RMx 750 and dont remember the other one). I was into seasonic but with good reviews and forum feedbacks I gave Corsair two chances... which failed. So I'm back to Seasonic with no regret (12 years total using and recommending them on a lot of computers for my, family and friends).

 

Good luck!

Writing from my PC that run almost 24/24 7/7 with its good old Seasonic SS-560KM (i5-3570K + GTX 1080)

Edit: I should probably check my PSU and its capacitors now that I think about it..

[vanGenne]

2 hours ago, xARACHN1D said:

I looked at what PSU you are using. Something that reliable should not have caused it unless you got a faulty PSU. Now here is my question... I see no fans on this list. Did you actually not put fans or did you just accidentally leave them out. Could you also let me know what case you are using and what fans you are using if you are using any.

Well my case is the Antec Nine Hundred Two, and I'm using the stock fans supplied with it. I forgot to mention that, thanks for reminding me! So that's one large fan on top blowing air out, and 2 smaller ones in the front blowing air in. The GPU has its own cooling fans of course, and the PSU also has one. It's not like I replaced those or did anything special, so I didn't mention anything about those fans. 

I added a Hyper 212 EVO to the CPU to accomodate the slight overclock (and because CPU stock fans are a little wonky sometimes). 

 

The PSU now works well in my GPU-less system, but like I said before, I'm unsure what caused the GPU to selfdestruct in the first place. If the PSU is still fine I'll happily keep using it, I just don't know how to check that before I put a new 500 euro card in...

[vanGenne]

1 hour ago, noyo87 said:

I'd personally replace the PSU. Even if you trust a specific model, shit can happen. No forum member can guarantee you the state of your PSU.

This makes a lot of sense, honestly. So it's not strange to assume the PSU could have caused my GPU to fry? Do you know of any way to check the integrity of my PSU so I can be sure?

[noyo87]

59 minutes ago, vanGenne said:

Do you know of any way to check the integrity of my PSU so I can be sure?

I don't know if there are good ways except with good lab electronic gear.

A basic visual inspection of the internal component is a first step but opening a PSU is also something dangerous (high current shocks).

[vanGenne]

2 hours ago, noyo87 said:

I don't know if there are good ways except with good lab electronic gear.

A basic visual inspection of the internal component is a first step but opening a PSU is also something dangerous (high current shocks).

Right.. And the fact that the PSU is currently in my system quietly powering everything but the GPU shouldn't be taken as evidence that it's still fine? 

[noyo87]

I would personally try to put the system under sustained stress. But without GPU, the power draw is way lower ... so I don't know how you can achieve this.

From there, I don't know much more

[Mitico]

Download OCCT and thest the whole system (ofc no gpu, singe you would use the integrated one) and see if the system will be stable, if yes, psu is fine.

[Kilrah]

PSU may be OK with only CPU load but start doing crap when both the CPU and GPU are drawing lots.

[vanGenne]

1 hour ago, Mitico said:

Download OCCT and thest the whole system (ofc no gpu, singe you would use the integrated one) and see if the system will be stable, if yes, psu is fine.

Thanks for the tip! Will OCCT really be able to stress test my PSU without the GPU though? I don't think my other components have the capacity to draw enough power to really test it

[vanGenne]

Update: OCCT did bugger all. Also, I took my GPU apart for the sake of satisfying my curiosity, and it turns out that one of the R33 voltage regulators is completely blackened. I could possibly replace it, but I don't mind buying a new GPU to be honest. Besides, who knows that damage that blown regularor has done to the card, and what caused that part to blow in the first place. 

 

So I guess I'll have to decide if I gamble on not buying a new PSU to save the money. Essentially, I'm wagering the cost of my new GPU (~€500) to save maybe €100. I think I'll just play it safe guys.. Thanks for all the help! A new PSU and GPU it is! Who needs to eat anyway?