Bad GPU (Caught Fire) Troubleshooting

koolerkii

Prominent
Sep 29, 2017
1
0
510
Hello! I have my personal system that has been in use for about 2 years now. It has been running mostly 24/7 being used for gaming, video editing, and running game servers. I have done no overclocking to the system. It is my first (and only) build.

I came home from work to find my computer turned off, which was odd. So I turned it back on and a large flame erupted from the right side of the GPU. I quickly turned the power off, slid the side panel off, and blew the flame out. I took out the GPU, unscrewed the back plate, and found a burnt out resistor. I decided to try my second open PCIe slot on the mobo just for giggles and the same thing happened with more fire. I took it out and set it aside. Probably not the smartest thing to do but oh well. I think its safe to say that the GPU is toast (literally).

Unfortunately my computer wouldn't start after this happened. With power applied, the motherboard and AIO cooler lights would come on. Upon pressing start the computer would flicker to life for about 1 second along with the fans and turn off immediately. So I redid all of the connections for everything. The computer then finally booted up. Posted just fine, and I have a nice computer still running off of the integrated graphics chip.

I have ran the computer this way for about 7 days now without any performance/stability issues.

My question is if it would be safe to order a new GPU (I'm intend on ordering a 1080Ti) and putting it into those PCIe slots, or if I should play it even safer and replace the mobo and PSU.

I'm not sure if its the PCIe slots or the PSU that could have/be the issue. I don't have a random crappy GPU I can throw in there temporarily. Though if recommended, I could buy one. I have about $1000-1100 to use. I searched the forum and found a few threads with people burning out their GPUs, but none really helped my situation.

Current System Build:
  • - NZXT H440 Black/Red
    - Intel i7 4790k 4.0GHz
    - Corsair H100i Liquid CPU Cooler
    - Asus MAXIMUS VII HERO
    - Corsair Vengeance Pro 32GB
    - x1 Samsung 850 EVO 1TB SSD
    - x1 Samsung 850 EVO 250GB SSD
    - EVGA GeForce GTX 980 Ti SC+ (dead :/)
    - Corsair RM850 80 Gold Modular PSU
    - Asus ROG SWIFT PG278Q Monitor
    - Corsair K70 RGB Keyboard
 
Solution
Obviously the GPU is done, but not sure if it failed because of itself, or perhaps the Power Supply.
Be a shame to replace the GPU to have it eventually fail again if it wasn't the main cause of the problem.
I used TOM'S HARDWARE PSU TIER LIST to check how your Power Supply ranked, and I think it may possibly be a Tier 3 piece of hardware. I'd definitely try to RMA/Warranty claim the GPU if you can, but beware that the PSU may be a culprit in all this.
Man, that's tough. I had a friend who's GPU (GTX 460) went pop gave a smokey burning smell. Didn't catch fire, but the GPU was dead either way. He replaced the GPU and hasn't had any issues.

I'd see about RMA'ing that 980 Ti with EVGA. Is it still under warranty?
 
Obviously the GPU is done, but not sure if it failed because of itself, or perhaps the Power Supply.
Be a shame to replace the GPU to have it eventually fail again if it wasn't the main cause of the problem.
I used TOM'S HARDWARE PSU TIER LIST to check how your Power Supply ranked, and I think it may possibly be a Tier 3 piece of hardware. I'd definitely try to RMA/Warranty claim the GPU if you can, but beware that the PSU may be a culprit in all this.
 
Solution

iamacow

Admirable
I hate to say thing but trying the video card in a different slot was dumb after you clearly saw a poof a fire.

What brand is you PSU? If you have a good one than all the safeties should have kicked in and that is why it wouldn't boot until everything was disconnected.

You are lucky..'

Edit:
While that Corsair RM850 could be a issue, my guess is your VRMs blew up. Happens often when things run 24/7 and not properly cooled.