Is it just me, or are video cards RMAed...A LOT!

[RedTiger]

I just got a new GPU in the mail yesterday, a EVGA gtx 275. Installed it, played some LOMAC, ran the Crysis benchmarks, so far so good.

 

So I sit down to actually play some Crysis just to see the improvement in-game. After about 30 minutes, the screen flickers with some static and then freezes. The graphics are a little corrupt with some strange lines that look like static diagonally across the screen. The music is still going but the PC is unresposive. "Oh well", I think, "gotta reboot".

 

I switch the computer back on and I'm greeted with a post screen that looks like its written in an alien language. The whole thing is scrambled. The Win XP loading logo is corrupt. Windows won't load! I make sure the card is re-seated, power's plugged in...still get the scrambled post screen. I swap it out for the old card (8800 GTS) and surprise! everything loads fine.

 

I begin to hunt around the web to see what caused it. The culprit sounds like failed VRAM. The only thing is that what can cause the VRAM to fail doesn't match my situation. The PSU's good (Corsair 650tx, never had any problems) and the card wasn't overheating. I'm forced to believe that the card was bad. I was overclocking a tiny bit, like 200 mhz over stock, but I can't understand why that would cause VRAM to fail. Instability, maybe, but a fried video card?

 

What I noticed in my research is that a crapload of cards seemed to be RMAed all the time. The internet isn't a good judge of the general population, but I'm noticing that people talk about returning video cards the way they talk about getting another beer from the fridge. No big whoop, happens all the time. Is that normal!?

[CyBerkut]

Well, as you essentially say, the internet isn't a good representation of the general population. After all, the people who have their cards work just fine are probably less likely to write about that trouble-free experience... it's not as interesting in most cases.

 

I couldn't give you a good feel for how 'normal' it is to have video card failures. I know it's not something that I hear a lot about from my online buddies. I didn't see much of it back when I had an ISP business that was also doing computer repair... but then, most of those machines were not on the higher end of the performance spectrum.

 

Considering the various things that can wrong with the technology, I don't think it is all that surprising to see failures... but it would be a 'big whoop' to me if it happened to *my* machine. :(

[EtherealN]

Question, when you overclocked the card by 200MHz, was that the vRAM you clocked?

 

My own experience of overclocking seems to indicate that the GeForce cards appear to be especially sensitive to vRAM overclocking, whereas I have a lot of leeway with the core and shaders. So after a few first experiments I have pretty much settled to leave my GFX card's memory where it's at (although for comparison I was able to run it stable at a 200MHz vRAM overclock, but this can of course vary from one individual card to another).

 

Note of course that since this is me, sample size is only one and thus not conclusive. But on the other hand it is a fact that testing for consumer products is less rigorous than testing for commercial products (with military and space being the next tiers), and while I don't recall the exact numbers failure rates for some consumer products were pretty shocking when I read an article on it some years ago.

 

Sadly, I have very little recent experience of it all, since until I bought the present computer I was (due to work) pretty much forced to just use a gaming laptop.

[RedTiger]

No, I was referring to my actual CPU. Its and E6850 that runs at 3.0 ghz. I built my system with a Zalman cooler in case I ever wanted to experiment with overclocking. I was able to overclock to 3.2 with complete stability with the 8800 GTS. I planned to go try for a bit higher after the GPU upgrade. I left it that way when installing the new video card, which was probably not the most careful way to do it.

 

Once I get the replacement, I'm going to run it with the CPU at stock and really pay attention to the temperatures and fan settings. To be more clear, it wasn't overheating when I last checked but I'm not certain of what it was doing when I was playing. For all I know the GPU fan could have failed. I didn't get the opportunity to check this. I'm going to be much more careful and observant with the next one. I'm interested to see what happens.

 

EDIT: before someone asks; YES I did have the PCI-e setting locked down at 100 in the bios. I know that increasing the FSB for an overclock will effect other components, but I didn't think it affected the video card if you locked down the PCI-e setting.

Edited August 20, 2009 by RedTiger

[EtherealN]

Oh, okey.

 

Then I do think it really is a case of a faulty card. As for how common, I don't know, but while it hasn't happened to me the statistics say it could be "relatively" common.

[RedTiger]

I really hope it was the card. Nothing else jumps out at me as being the cause. If it isn't, I wonder how many I'm going to RMA until I get the idea that it doesn't work with my system. :D

[Ramstein]

It is very hard to tell if it is the card or motherboard.. or software..

 

If you feel you hyave done everything possible, including re-installing the OS, swapping out video cards to test everything, swapping out the PSU, and it still has problems, it only leaves you with RMA.

 

:joystick:

[RedTiger]

It is very hard to tell if it is the card or motherboard.. or software..

 

If you feel you hyave done everything possible, including re-installing the OS, swapping out video cards to test everything, swapping out the PSU, and it still has problems, it only leaves you with RMA.

 

:joystick:

 

Well, no actually you can tell pretty easily if its the software (if you mean drivers and such). If your post screen is scrambled then the software probably isn't the problem since it wouldn't even have loaded yet.

 

Swapping out the PSU isn't an option and I'm not in the mood to reinstall the OS. I'm pretty much stuck with the card, be it a working GPU or a $230 brick since it can't be returned, only RMA for a new one. Based on what I've researched, past experiences, and some recent experimenting with absolutely no problems and total stability with the old GPU and PSU, I'm making an educated guess that it was a faulty card. This wasn't your run-of-the-mill stability problem, this appears to be the case of the card bricking itself in the middle of playing a game. It's a video card now only in terms of producing a video signal that can be interpreted as garbage by the monitor.

 

Moot point too since I've already dropped it off at the UPS store. :D

[Vault]

RT I had exactly the same problems as you. But for me it was my mainboard that was the culprit and I'm now on my 2nd RMA. The fonts during post/boot was jumbled up and I had red lines all over the screen, the only way I could boot the system was in safe mode.

[RedTiger]

Ouch! Luckily for me, everything was fixed once put my old video card back in. Everything ran fine. I would be really upset if it was the motherboard. :(