post #1 of 1
Thread Starter 
Some time back, my EVGA GeForce GTX 560 Ti started having issues. I have a EVGA GeForce GTX 970 on the way, and for most of the time between then and now, I've been using a PNY GeForce GTX 650.

The nature of the failure was that shortly after it was put under load (any game, Windows Experience Index putting it through the benchmark, etc.), it would have one or two flashes in spots on the screen (not the entire screen flickering, and not polygonal flashing; rendering was seemingly otherwise happening correctly), and then the picture would just freeze and sound would loop (if in a game, when freezing during the WEI it would just freeze). Temperatures of the GPU were pretty low (low to mid 50C range) as this occurred, so it didn't appear to be overheating. It worked fine one day, and then I turned my PC off to clean the dust off the filter for the side fan, and when I powered it back on after that, it started having issues very quickly in games.

Anyway...

I was looking at my old one to see if I could find anything, and when taking the heatsink off, I noticed this.



Along the right of the row of those twelve small chips, there's four smaller ones. The middle of the four, and most prominently, the bottom of the two middle ones, seems to appear to have some discoloration above them. Are these burn marks? I couldn't smell anything (not now, and not when the failure occurred), and these marks are actually somewhat easy to miss unless you look at it under the right lighting. They're not very apparent.

Are those part of power regulation/delivery? Maybe when it draws above a certain amount of power, it fails? That makes sense because it failed when at low temperatures, and actual scene rendering seemed more or less normal up until the flash and then freeze. It was fine when at the desktop, with a web browser, or watching videos (flash).

The heatsink (where it met the GPU and around where it met the VRAM, some worse than others) and I/O bracket (around where the air was vented out) also had discoloration (heat stains are they called?) on them, but it wasn't that excessive. I'm not sure if that is normal. I did notice my old CPU heatsink's heatpipes did a similar thing over a few years.

Also, as one last question, to anyone who knows a lot about power and such (I don't), is it possible the fan speed selector on my case (Fractal Arc Midi) could have caused this to the graphics card? Between it working and having this issue, all I did was clean the side fan dust filter, and the other thing I did was I changed the fan speed selector (this changes the voltage they get). The reason I did this was when I was setting my tower back upright, I must have switched the speed to it's highest, so when I turned my PC on, it was loud. During Windows loading, i set it from the top speed (12V) to the lowest (5V). Something like this can't hurt it, right? I just want to know for future reference.