repeated videocard corruption

[gypsiesun81]

Okay I have been having serious issues with my video cards for the last year or two. The problem began with my radeon 9700 pro, and most recently occured (repeatedly) with my Geforce 6800 GT. The problem is video corruption and artifacts. It goes like this...I get a new card...it works for a few weeks or months and then out of the blue this display corruption starts happening , forcing me to RMA the thing . I get the card back,... it works for a while and then it happens again. This only seems to happen with cards that require an additional power connection, I have a old geforce 2 gts and a radeon 9600 both of which dont require extra power...and have NEVER given me any problems. I need help here please! Im getting sick of sending my cards back for RMA.


My specs.

Athlon XP 3000+
Asus A7V 333 mobo
1 GB pc3200
Geforce 6800 GT ( which is on the fritz ....again...had to stick my 9600 back in ....its a bitch cause I was really getting into Oblivion ...now I gotta go back to 800x600 aarghhh )
SB Audigy 2

 

[steckman]

Heat can damage your cards over time, so maybe you have a temperature problem in your case. The cards that require an additional power connection probably generate a lot more heat than the others, so that could be a connection.
I suppose it could also be a power problem, but I think that would show up a lot faster than a few weeks or months.

 

[gypsiesun81]

Well I had suspected heat as a cause, but I also read in reviews that the card I was using ( Winfast A400GT ...a Geforce 6800 GT 256mb) runs hot as its norm ...(40C idle ,, abt 60-70C under load ...and those were the temps I was getting on my GPU. ) I have a 450 watt power supply..which as far as I know is sufficient. When I got my 6800 initially it ran flawlessly for about 6 months. Then after RMAing it ( twice ) each time it only worked for a few weeks or a month. Now it looks like I have to send it back a 3rd time.

 

[steckman]

Those temperatures sound perfectly normal so I'm reaching now. Maybe you have some kind of corrosive agent in your AGP slot that is degrading one of the contacts with the card over time. I've never seen that, but it's theoretically possible.

 

[gypsiesun81]

Well this has happened to the same card on 2 different motherboards! So the corrosive agent thing is...i agree.... a bit of a strech. My temps never got anywhere near the 120c that causes GPU damage. This is whats driving me crazy.... sure it ran a little hot ....but not out of the ordinary operating range for my card. I just cant figure out why this keeps happening.

 

[cleeve]

Get a voltmeter and see what your PSU is outputting on the molex connection for the cards that require a power cable.

If it's out of spec it could be killing them...

 

[perz]

Just get an Artic cooling vga silenser. They are the best.
Normal VGA-fans are crap, i belive they are actually made to break down....

 

[mdd1963]

"I have a 450 watt power supply..which as far as I know is sufficient. "

Rather, you supsect it is sufficient, going purely by it's wattage rating..

not all PSUs are created equally....

As you have seen this prob with two mainboards and two video cards, I'd start with a new, quality/namebrand PSU...

 

[soeddy]

Hello I already had the same problem, I posted him here and they advised me to add more fan in the pc and it resulted try it...maybe help somehow

 

[Djhotchips]

I think I may be having the same problem..

I'm running a 6800GT (AGP), running in a Thermaltake Tsunami Dream case (480W Thermaltake TWV480 PSU), with 2x 120mm fans, and an 80mm side fan, my GPU core temp sits at 62 degrees Celsius under load, and when things go pear-shaped, I've noticed that 1 of 3 things tends to occur:

Most frequently, the screen will freeze, and after about a minute, I'll get a blue screen claiming that the nv4disp.dll driver file threw an exception. This starts a memory dump and I've got to reboot.

The next most frequent reaction is that I'll get horizontal artifacting (generally only 1 pixel in vertical depth), that also tends to hang the machine for a minute or so, then goes back to normal.. if this happens, it tends to do so repeatedly every few minutes, before it finally crashes on me.

The most infrequent reaction (in fact, only happens when trying to run 3DMark05), is that the entire system freezes indefinitely.

I've tried 2 different graphics cards, on two different mainboards, with 9 different versions of the NVidia drivers (versions tried: 53.03, 71.84, 71.89, 77.72, 77.77, 78.01, 81.95, 81.98 and 84.21), and still can't figure it out.

I'm rapidly running out of ideas for what could be causing the problem, but luckily, am still within a few months of my warranty, so if it's a hardware problem, I can still return the defective part, the only thing is, I can't isolate the fault

Any assistance would be extremely appreciated.

 

[steckman]

I have seen the nv4disp.dll cause a blue screen crash before. This consistently happened with a PNY 6800 Ultra (PCIe) hooked up to a 24" Dell flat panel monitor while running an OpenGL application. We actually had two of those boards and they both suffered from the same problem.
The interesting thing was that we never saw the issue when hooked up to two CRTs that combined had a higher resolution. And we only saw it when an OpenGL application was running. Anyway, we narrowed it down by process of elimination to either a design flaw in the hardware or the driver. And by "we", I mean me, our client, and PNY technical support.
Once we narrowed down the issue so that it was easily reproduceable, we were able to demonstrate that it only happened on our PNY 6800 Ultra PCIe boards ... not on X300s, 6600s, vanilla 6800s or a 6800GS. At that point, PNY threw up their hands and gave our client free upgrades to 7900s.
If you have the time and access to extra equipment, I suggest you jump through all of those hoops. Maybe you'll get a free upgrade out of it!

Incidentally, the nv4disp blue screen crash has to do with the video card getting stuck in an infinite loop waiting for the display to respond. I don't think anybody has a good answer as to why this happens. Not that I could find anyway.

 

[Djhotchips]

hey guys, thanks for the quick response.. I'm not sure that I've found a permanent fix, but I found (after several extra hours trawling the web looking for an answer) that by switching from the standard Nvidia drivers to the latest forceware drivers, I eliminated the problem totally.. Could not reproduce it. Ran 3DMark05 overnight at max settings. It certainly chunked at points, but that was due to my CPU not being able to handle it, and had absolutely nothing to do with the card.

Am still tempted to jump through all the hoops so I can try and get a 7900, but for the moment, I'm happy now that my precious 6800GT is running properly again

Hope this helps someone else out there having the same problems!