Hello everyone, Few days ago I was gaming a little when my computer just shut down with no indication of possible failure and then the min. after there was this burning smell coming out of computer case. Pulled out the plug, opened side door to see that my card was burned for no good reason. Warranty expired a year ago I believe. My computer still works fine once I took it out (IGPU, now I've tried Pci-e back up card and it works just fine), so I wonder why did it burn out with light gaming? I clean dust regularly, it wasn't dusty and it wasn't overheating, temps were around 70 to 80 c depending on games I play, haven't overclocked or overvolted the card, it has been ruining on stock from day one. So, with my limited knowledge of electronics I have no clue if this card could be saved somehow, what do you think, is it possible to fix even though the PCB has burned deep from the looks of it? My setup: PSU: Cooler Master rs-850-esba MoBo: Gigabyte Z68X-UD3h-B3 CPU: Intel 2500K @4.0Ghz GPU: Gigabyte Radeon HD 7970 3GB MEM: 12GB Kingson DDR3 1600Mhz HDD: Samsung Evo940 256mb UPS: APC SmartUPS 2200 Pictures: I was searching about this online and found out that I'm not the only one with this particular brand and card to catch fire...
Well thanks Airbud for uploading those pictures, I was hoping of fixing this one if possible at all instead of buying a new card. You know, the sad thing is that I've bought it for gaming, but since I lack time to game I pretty much game on average about 1.5-2h/day, imo that is really light usage for few years and I can't believe it crapped out on me. I have an old GTX 280 as a back up card, I've been torturing it for hours now just to see if my PSU killed my GPU but so far everything is fine and voltages are stable, will leave it for another 5 or 6 hours to run but its most likely just 7970 decided to die on me. I found out that someone else with the same card had the same happen to him/her: https://www.reddit.com/r/pcmasterrace/comments/30wqe8/7970_died_spectacularly_on_me_this_morning_fire/ It's possible that Gigabyte messed something up in design or something of that sort, has anyone else experienced same or similar problem?
I agree with Sykosis, too much looks damaged to repair (without a huge amount of effort and time, sourcing the right parts, which may or may not work), time to upgrade IMO
Yeah, its probably not worth bothering with it, its most likely beyond repair... A friend of mine is offering me Gainward GTX 670 Phantom for $100, though I haven't used nVidia since mx440 lol, and looking at benchmarks its about the same as 7970 if not little stronger. Maybe its wrong forum to ask about nvidia card in Radeon forums, but is it ok card, no problems, defects, etc.? In any case, thanks for responding, its much appreciated.
the part i circled in is the voltage reference ic, which is 99% the cause of the issue here. it stabilizes voltage/current independently from temps and power input/+draw. the components that are integrated downstream of this IC are affected of course. so either: - internally related - 1) the IC itself was faulty - externally related - 2) you had voltage(=current) spikes in the supplying main circuit (PSU) that strained the IC with out of spec numbers (most probable) 3) your card had a high power limit set and "asked" for more current than the IC could handle. (did you have the power limit manually adjusted?) of course it can be a combination of 1) + 2) or 1) + 3) as well. its a little sad to lose a card over that component. but repairing it still will need you to identify the damaged capacitors/resistors, remove them and the voltage reference, search for exact replacements, clean the circuit board of residue, resolder the new components. way too much effort especially considering not everyone has the tools... if you're an etech guy... sure go for it. although keep in mind the possibility that not all damage may be visible. leaving you with endless trial and error which parts have been affected additionally to the burnt ones. what you SHOULD do in any case though would be measuring your 8pin pcie rails of your PSU under medium gaming load (just as when the problem occurred). because an old faulty PSU (if that was the reason) will render your next new GPU useless again, just a question of time. the coolermaster is by the way known to have a relatively high standard deviation in its line regulation functionality. just as a sidenote.
Vase, good stuff man, much obliged. I still have my GTX 280 torture testing (though the card is messed up, shows some little artifacts in 3d, but I use it for testing purpose and as just in case card because in 2D it works fine, dont care if it burns or dies). Anyways, I managed to find multimeter and I measured 2x6 pins for PCI-E, it shows aprox. 12.12v and 12.11 (it fluctuates by 0.01-0.02v up and down) under full load, that should be great I guess, not sure how stronger 7970 is but I dont think its pulling that many more amps then 280. In HW monitor for 12V it says 12.122V value and 12.120V as minimum (12.172V as maximum). Ive bought this PSU because I've read few reviews on it and found out that it was pretty solid unit with CWT OEM stuff inside, I think there was review on JonnyGuru where he tested it heavily and it passed pretty much everything, and it was on sale so I said why not. But yeah, CM PSU's arent that great mostly. About power limit, if you mean through Afterburner manually adjusting it then yeah, that and Fan speed I manually adjusted, fan speed to more aggressive settings to keep GPU cooler and Power Limit to 125 because it kept dropping to 85 which affected my FPS (instant drop in FPS when Power Limit drops). I had no clue that raising it can kill something, damn... I also have APC UPS unit rated at 2000W, even if there was current spike that should've protected my components, but who knows what the hell happened here.
yeah! it's not a bad psu (generally spoken), i don't want to convey a wrong image! and yeah... who the hell knows what exactly happened there... :nerd: if everything is in order a raised power limit should be fine. just... when push comes to shove... it can be a deciding factor, amongst other preconditions that have to be present of course. again in general, raising power limit (inside the software boundaries, i guess thats +25% - 50% depending on tool ?!) in a perfectly stable system with healthy components is not a problem. if you raised it higher than your software slider offers to (for example through BIOS mods) then it's another story though...