Is my 280x dead? (pics)

[cooper69]

my 280x randomly started crashing on me lately (after a few minutes or seconds into 3d application) and showed that familiar looking colored pattern screen. didn't really look like traditional artifacts wih squares and dots everywhere, but more of a gradient pattern, so to speak. sometimes it's an even/solid color across the screen. anyway, this requires me to hard restart the computer

so at this point i have no idea what's wrong with it. i took pictures of the hardware hoping someone on here can help me out i saw that post about the dead gtx 570 and thought i'd try my luck too

 

[nurturedhate]

Any way to get better quality pics. Low res and blur makes it a little hard. What really stands out is the thermal paste on the gpu itself. If you didn't attempt to clean it at all I would say it looks completely dried out. Do you know what sort of temps you were getting before you took it apart?

The caps look ok, though the lighting and photos make it hard to tell. Anyone with a better eye wanna take a shot?

 

[Hi-Fi Man]

With artifacts there's only four things really plausible and that would be VRAM going bad (not visible), GPU going bad (not visible), VRMs going bad (usually visible) or capacitors going bad (usually visible). Since there is a heatsink covering the VRMs you need to take that off to rule them out. Capacitors look fine.

What environment was this card run in? Any dust blockage? Temperatures? All this info is needed because the card could have just been too hot.

 

[SithSolo1]

I don't see any leaking caps or scorch marks anywhere. Do you know if the fan was running?

 

[RussianSensation]

R9 280X came out October 8, 2013. Check asap when you actually purchased it per the invoice date and apply for RMA.

Do this ^ first as some 280X cards have 2 year warranties and you don't want to miss your window.

 

[cooper69]

Any way to get better quality pics. Low res and blur makes it a little hard. What really stands out is the thermal paste on the gpu itself. If you didn't attempt to clean it at all I would say it looks completely dried out. Do you know what sort of temps you were getting before you took it apart?

The caps look ok, though the lighting and photos make it hard to tell. Anyone with a better eye wanna take a shot?

I don't see any leaking caps or scorch marks anywhere. Do you know if the fan was running?

Yeah, the cooling shouldn't have been an issue. since I actually replaced the thermal paste shortly after taking the pictures. Fans are running fine, didn't get past 55 degrees celsius before it crashed on me.

R9 280X came out October 8, 2013. Check asap when you actually purchased it per the invoice date and apply for RMA.

Do this ^ first as some 280X cards have 2 year warranties and you don't want to miss your window.

unfortunately, the card isn't under warranty since I lost my receipt/proof of purchase

With artifacts there's only four things really plausible and that would be VRAM going bad (not visible), GPU going bad (not visible), VRMs going bad (usually visible) or capacitors going bad (usually visible). Since there is a heatsink covering the VRMs you need to take that off to rule them out. Capacitors look fine.

What environment was this card run in? Any dust blockage? Temperatures? All this info is needed because the card could have just been too hot.

no dust or blockage or absurd temperatures, but it was used for long periods (24/7 actually) at 100% load for computing.

i'm 90% sure this is the reason why, but i don't know if this is fixable

 

[supaidaaman]

Bake it. Dead card, artifacting, evil spirits. I bought a $5 toaster oven on Craigslist and have cooked all my cards instead of throwing them in the trash. My 580 is still going strong after the past few months and it was completely dead. No signal.

 

[MongGrel]

Register things when you buy them.

Should have been under warranty, I would imagine.

 

[RussianSensation]

unfortunately, the card isn't under warranty since I lost my receipt/proof of purchase

What brand's card is it? Some brands like Gigabyte trace warranty based on the serial number, not invoice. So if it's Gigabyte, the warranty could be intact.

If not, the only thing I can think of is baking it.

 

[spat55]

I'd just take the heatsink off the VRM's have a look it that seems okay I'd take off all the TIM replace that and test. If you still getting issues try lowering clocks with MSI Afterburner if that works you know it is bad hardware. Could even be a bad PSU.

 

[psolord]

I saw you referring to a GTX 570 thread and I guess that was my thread.

Unfortunately I haven't managed to fix it yet, but I still have some hope, because I definitely had some capacitors that were dead or close to dead.

I did mention in another broken card thread however, that I did manage to fix my 5850 by baking it, soon after I had my first experience with my 570.

I baked the 5850 for twice as long and for 10C more, but I cover the capacitors with tin foil, as Rus suggested. The plastics as well.

You should really check for the warranty validity though. Contact the manufacturer and see what they think. Also if you ordered online, maybe you can find the email of the purchase. They could accept that instead of a receipt.

 

[XFXSupport]

Yeah, normally if a card is within 24 months of purchase date, most manufacturers don't seem to mind replacing them. Then again, in my experience, this can change from agent to agent.

I don't believe that's an XFX card, single DVI, no XFX SN sticker, but i'm not absolutely certain. If its a 280x from XFX, it should still be covered under the standard 2 year warranty. Let me know.

Mark at XFX

 

[cooper69]

It's a Visiontek card. Will try to get pics of VRMs asap

 

[tweakboy]

Could have been worse. Be happy now youll buy a more current video card. gl