Is my graphics card about to die? (Update: RMA woes, excess solder on back of card?)

[Techhog]

Your opening post sound like a variant of GSOD, what drivers are you using? Flash videos was giving me a similar problem with 13.12, but I use Firefox, switched to 13.9, no problems so far.

I'm using 13.12, but I've never had this problem before and I only get screen jumpiness/tearing/artifacts (not sure which to call it) if I've overclocked since turning on the computer or waking it from sleep. (It stays if I return to stock, but goes away when I restart.) Also, the screen only stayed messed up like that for about a minute.

^^^Also try that.

I did. I still can't overclock without the jumping/tearing/artifacting issue when the screen is showing 2D or transitioning to 3D (it doesn't happen during any gameplay or benchmarks)

 

[Techhog]

Okay so, luckily, the freezing I was talking about earlier was due to a shoddy PC port. Other than slightly increased temperatures, I haven't had any issues at stock for a week.

 

[DigDog]

That is fraud, and highly frowned upon by this forum.

While I cannot say whether or not the company could "detect" overclocking, it's not true that it's just a "PC settings change", it does directly affect the card, when you change the voltage slider.

And he mentioned he enabled "Unofficial Overclocking", which means he went beyond the BIOS / mfg-supported overclock ranges, meaning he intentionally violated the warranty, and possibly damaged the card.

All of that said - perhaps not all is lost. Chrome's hardware acceleration is notoriously buggy. Try turning it off, and see if your PC is stable otherwise.

1) i know, it's hard to write with emphasys. i didnt mean to suggest he should do it; however, it has been done before, and by people here as well.

2) the OC settings on something like afterburner do in fact change voltages on the card, but trough windows. The same card plugged into a different system would not keep those settings because at best they are at driver level.

as for the warranty, we need to use a bit of common sense. Asking the card to run at a higher frequency is hardly gonna break it, it might at best result in slightly higher temperatures. Then she either does, or doesn't. What really damages the card is overvolting it.

Ofc warranties are worded so that the company can unload responsability under any possible circumstance, but in the real world, using afterburner and higher clocks, i wouldn't feel bad returning it if it failed.

BTW: i was serious with the baking thing. it has been done and it has fixed cards before. microfractures in the solder get repaired by simple baking. last option really but one worth considering.

 

[Techhog]

1) i know, it's hard to write with emphasys. i didnt mean to suggest he should do it; however, it has been done before, and by people here as well.

2) the OC settings on something like afterburner do in fact change voltages on the card, but trough windows. The same card plugged into a different system would not keep those settings because at best they are at driver level.

as for the warranty, we need to use a bit of common sense. Asking the card to run at a higher frequency is hardly gonna break it, it might at best result in slightly higher temperatures. Then she either does, or doesn't. What really damages the card is overvolting it.

Ofc warranties are worded so that the company can unload responsability under any possible circumstance, but in the real world, using afterburner and higher clocks, i wouldn't feel bad returning it if it failed.

BTW: i was serious with the baking thing. it has been done and it has fixed cards before. microfractures in the solder get repaired by simple baking. last option really but one worth considering.

I know you were serious. The thing is that I've been 100% stable at stock for a week... I'm really thinking that I just pushed it too far and that it just needs to be kept at stock.

That said, I'm still getting this weird glitch where GPU-Z says that it overvolted itself to 1.498V for a second. I really have no idea what's up with that.

 

[DigDog]

vrm blown ?

 

[Techhog]

vrm blown ?

Maybe. The fact that it only does it for a second, on top of the fact that my VRMs have never exceeded 90C, makes me think that it's just another GPU-Z glitch. I think I'll download another program to be sure.

 

[Techhog]

Yeah, the spike doesn't appear in MSI Afterburner or in HWMonitor. Either I'm moosking at the wrong thing or it's a glitch

 

[Techhog]

Well, it was fine for about 3 weeks. Now I'm getting artifacts everywhere in the Cold, Cold Heart DLC. >_> This is such a broken game that it wouldn't shock me if the game is at fault, but it's probably safe to assume that my card will die today.

 

[Face2Face]

How much did you OC your card and how much voltage did you give it? If the card dies, call up Powercolor and tell them what's going on with the card. Tell them the truth about your overclocking and they may just swap it out for you. There is no shame in trying.

 

[Techhog]

How much did you OC your card and how much voltage did you give it? If the card dies, call up Powercolor and tell them what's going on with the card. Tell them the truth about your overclocking and they may just swap it out for you. There is no shame in trying.

It's been at stock for 3 weeks :/ *sigh*

If it dies, I'll just sell the rest of my computer and stick to my laptop.

 

[superkoopatroop]

So, I started up a flash video, and suddenly my screen went with colored lines everywhere. It's been really jumpy due to changing clocks, and Afterburner confirms that it was not at idle clocks when it happened. It looked like an extended version of one of the jumps. The strange thing, though, is that I was on stock clocks at the time. I only have it overclocked when I actually need the extra power. It's really strange. Any ideas?

Sorry for the errors ahead of time my phone is like a box of chocolate. However I feel your pain. I really really do. I to bought a 7950. I got it for 200$ before the bitcoin craze and I couldn't of been happier. It was the HIS iceq and it ran about 80c under max load. The card was undervokted by his by default and I had it slightly overclocked with a voltage of 1.150 with 1025mhz. I had this running fine for 3 months and one day we lost power at my house. When I bottled my pc back up it slowly drawn an image on the screen from top to bottom and it was trying it's hardest. I reset my pc and overclock and it still slowly drawn an image from top to bottom. I reinstalled the drivers and everything seemed smooth for another day or so. Then it started showing blue pixels everywhere. They would go away but if u looked close enough at the pixels it was evedent they still where there. Giving me the finger. Then last week I was getting crazy studdering as if I was running 2gb of ram and then poor blank screen. From then on it would not post anymore. So in the course of 4 months my gpu died. I'm not sure if amd is to blame or HIS but regarless AMD lost a customer

 

[Techhog]

Either rebooting or turning on vsync fixed the issue... I should probably test with actually stable games to know for sure.

 

[Techhog]

I can't reproduce the artifacting in anything... I want to say that Arkham Orgins is the same game that crashed my entire computer before (red screen, then reboot). This isn't a game that I trust to not have issues... I'm gonna say that I'm fine.

 

[lavaheadache]

Sorry for the errors ahead of time my phone is like a box of chocolate. However I feel your pain. I really really do. I to bought a 7950. I got it for 200$ before the bitcoin craze and I couldn't of been happier. It was the HIS iceq and it ran about 80c under max load. The card was undervokted by his by default and I had it slightly overclocked with a voltage of 1.150 with 1025mhz. I had this running fine for 3 months and one day we lost power at my house. When I bottled my pc back up it slowly drawn an image on the screen from top to bottom and it was trying it's hardest. I reset my pc and overclock and it still slowly drawn an image from top to bottom. I reinstalled the drivers and everything seemed smooth for another day or so. Then it started showing blue pixels everywhere. They would go away but if u looked close enough at the pixels it was evedent they still where there. Giving me the finger. Then last week I was getting crazy studdering as if I was running 2gb of ram and then poor blank screen. From then on it would not post anymore. So in the course of 4 months my gpu died. I'm not sure if amd is to blame or HIS but regarless AMD lost a customer

AMD lost a customer because of that? Your problem could have happened no matter what brand card you had.

Where in Mass you from btw? We are a rare breed on these forums.

 

[Techhog]

AMD lost a customer because of that? Your problem could have happened no matter what brand card you had.

Where in Mass you from btw? We are a rare breed on these forums.

The fact that so many people consider Nvidia the premium brand and AMD the cheap brand causes them to automatically think that any problem is caused by AMD being low-quality. It's a perception thing and AMD will never lose it.

 

[superkoopatroop]

AMD lost a customer because of that? Your problem could have happened no matter what brand card you had.

Where in Mass you from btw? We are a rare breed on these forums.

I'm from fallv river and we so are lol. I've converted a few friends and built gaming pc's for them. It's true that could of happened with any card but in ny unbiased oppinion. Nvidia is walking at normal speed and not breaking a sweat. While amd hit the ground running and giving it everything they got and yet they both travel the same speed. I got to believe this has a predispotion On the lifetime of the card. like how an overclocked cpu is bound to die sooner. Nothing else failed on my pc. Or in the house in general. It may have been a combination of power surge and unstable ram. It used the same copper block to cool the gpu and ram meaning that the ram would also be running 80c+ and they may have failed seeing how it wasn't hynx ram but of lower quality. Either way in sadly now stuck with a 660 not ti for now. On the plus side I did discover shadowplay. So the glass is half full I guess. :-(

 

[superkoopatroop]

I can't reproduce the artifacting in anything... I want to say that Arkham Orgins is the same game that crashed my entire computer before (red screen, then reboot). This isn't a game that I trust to not have issues... I'm gonna say that I'm fine.

I thought I was fine but it came back with vengeance. Looking back on it now. I wish I would have downlocked everything down 100mhz and maybe It would still be alive. Ahhhhhhhhhh rip my sweet.

 

[Gloomy]

Could be the power supply.

 

[Techhog]

I thought I was fine but it came back with vengeance. Looking back on it now. I wish I would have downlocked everything down 100mhz and maybe It would still be alive. Ahhhhhhhhhh rip my sweet.

I just never know with this particular game though, so no idea. :/

 

[Techhog]

Could be the power supply.

Do you mean for me?

Thinking about it, I do need to clean the filter. It's been a few weeks

 

[Gloomy]

Do you mean for me?

Thinking about it, I do need to clean the filter. It's been a few weeks

Yeah. If it's the power supply, cleaning the filter won't fix it.

How old is your PSU?

 

[Atreidin]

AMD lost a customer because of that? Your problem could have happened no matter what brand card you had.

Where in Mass you from btw? We are a rare breed on these forums.

No kidding, next we'll see a post from someone who had their house burn down from a lightning strike and complain that AMD should have prevented the surge that fried all the electronics in the house, the subsequent fire, and the anger of Zeus that caused the bolt to exact mighty vengeance upon the house's inhabitants.

 

[Techhog]

Yeah. If it's the power supply, cleaning the filter won't fix it.

How old is your PSU?

Brand new. I got it in October, just like everything else.

It might be overheating from bad airflow though. I really need to move it somewhere better

 

[Gloomy]

Brand new. I got it in October, just like everything else.

It might be overheating from bad airflow though. I really need to move it somewhere better

Ah, if it's brand new, don't worry about it.

I've had issues with artifacting and poor behavior in the past, and it turned out the power supply was just too old to drive the GPU.

For a while I thought it was the GPU also, the artifacts would only happen once in a while and it was impossible to reproduce reliably. I didn't even consider the PSU until I had the GPU replaced and still had artifacts. What it turned out was, sometimes power consumption would peak while I was gaming and the PSU couldn't handle it... so just for that split-second, everything went bananas. The PSU was just too old, and became unreliable.

 

[superkoopatroop]

No kidding, next we'll see a post from someone who had their house burn down from a lightning strike and complain that AMD should have prevented the surge that fried all the electronics in the house, the subsequent fire, and the anger of Zeus that caused the bolt to exact mighty vengeance upon the house's inhabitants.

No need for the condescending attitude. I further explained my reasoning a few posts above. Perhaps the pcb was of lower quality, perhaps the solder. when my power blacked out my card wasn't a paper weight. it kept going for another month without much issue. Further more, I really have no evidence that the power surge actually did the damage, the card could of been damaged from something else entirely. The only reason I question the quality of the 7950 is because no other part of my PC failed. My corsair gs750 psu didn't fail, my 3570k didn't fail my asrock z77 extreme 4 didn't fail, my monitor didn't fail, my fridge didn't fail but my 7950 did. slowly and painfully. I mean, you would think it would of been the PSU failing considering it was plugged into the wall. I usually go amd because price/performance but now I'm jaded.

All I'm basically saying is I'm not going AMD again because my card died within 4 months. argue that logic please.


"""I'm from fallv river and we so are lol. I've converted a few friends and built gaming pc's for them. It's true that could of happened with any card but in ny unbiased oppinion. Nvidia is walking at normal speed and not breaking a sweat. While amd hit the ground running and giving it everything they got and yet they both travel the same speed. I got to believe this has a predispotion On the lifetime of the card. like how an overclocked cpu is bound to die sooner. Nothing else failed on my pc. Or in the house in general. It may have been a combination of power surge and unstable ram. It used the same copper block to cool the gpu and ram meaning that the ram would also be running 80c+ and they may have failed seeing how it wasn't hynx ram but of lower quality. Either way in sadly now stuck with a 660 not ti for now. On the plus side I did discover shadowplay. So the glass is half full I guess. :-("""


I've been doing some research and I found out that some 7950s vrm overheat badly. In my case, the vrm didnt have a temperature probe so the GPU couldn't actively monitor vrm temps and adjust the fan accordingly. HIS has stated they felt the vrm temps are cool enough that they didn't need it. Some sapphire cards hard the same issue heres the link. And this guy Belial seems to know a ton about the AMD 7950 and its a good read. link
"Recently companies have been using a cheapo 7950 pcb made by a chinese PCB company known for making cheaper versions of cards called Yeston, that has a weaker power phase with LFPAKs, is more prone to VRM blow-outs and coil whine, and has a much lower ASIC. However the worst, is that these PCBs use a no-name PWM chip instead of CHiL, meaning you won't get any VRM temperatures, amperages, or wattage readings."