Nvidia 6800GT Pc Lockup problem

SirLemonhead

I'm running a 2800+ XP Athlon processor overclocked to 3200+ spec on an Nvidia Nforce 2 motherboard by Abit. 1gig of 3200 ram. Samsung LCD monitor using the DVI connector. Running Windows XP with SP-2.

Was running an ATI card and had zero problems with this configuration. I've since gotten an MSI NX 6800GT TD-256 card, with an NV Silencer 5.

My pc is locking up while playing games. After a random amount of time, my screen gets covered in a weird pattern of small green pixels inside what look kinda like rectangular squares. The pc then locks up completely and sound will loop for 5-10 seconds or so. Hard to explain, but the artefacting is different between the two games and always looks the same. It'll happen the quickest in Warcraft III, say 10 minutes or so, but CS Source can be played for about 30-40 minutes at times. Also happens in Painkiller, DOD Source. The colour of the artefact pixels varies occasionally.

The pc needs to be rebooted when these lockups occur. There will be no artefacting present on the bios or in windows after these lockups (Bios and Windows usage gives me no problems)

No errors ever show in the Event Viewer in Windows XP.

Temps are fine, 47 degrees idle. I haven't checked the temps during gameplay, but if there's no artefacting if I reboot after a lockup, I presume theyre ok? I do though rarely get the weird stretched mess of polygon vertexes thing where you get weird long triangles of polygons stretched out all over the screen.

My psu is an Antec NeoPower 480watt PSU which has twin 12V rails. Should be adequate for running the card (it has a PCI-Express power connector)
I upgraded from a chieftec psu which was giving me the exact same problems in the exact same ways, so I don't believe it's the PSU. The card is on it's own independant power cable.

I've tried various drivers, Omegas, 3 official Nvidia drivers and about 4 or 5 beta Nvidia drivers. I've done driver cleans, and a windows XP repair and a clean install and the problem is still present. I've also tried various chipset drivers.

Tried fail safe settings in the motherboard, optimal settings, turning off practically everything in the BIOS I don't need.

I really haven't a clue what else could be causing this. If I've left out any important information, let me know.

christophicus

this is just a thought ,but your gpu might be overheating.Did you install the silencer your self?it might not be functioning properly,eveb if it was on the card when you baught it.

SirLemonhead

Since posting this I've had 2 windows lockups. Never had this before, just in games :mad:

I got the card second hand from someone on boards and he had put the cooler on himself I presume. It looks like everything is making contact though..

You'd recommend taking it off and redoing all the thermal paste?

christophicus

If you feel confidant that you wont break anything i would,just remember to remose any old paste first.by the way this is just a theory of mine i have no idea if it is a sound theory o0r not though.

RicardoSmith

Checked screen usually means video memory corruption. Which is usually.

A bad bios flash on the gfx
Overheating gfx (are you monitoring temps on the card?)
Faulty memory either through overclocking, or manufacturing fault.

If it was something else in the PC. like the rest of the system overheating, or a bad PSU, or 5v line, or power surges though the AC supply. I would expect it just to freeze or reboot. Not screen corruption like you describe.

On the off chance it is overheating. Remove the side of the case and direct a desk fan at it. See if you can run your game longer like that. Also try if someone else can run the card at all. If they have the same problem its a problem with the card. If not its a problem with your PC PSU. or AC power.

greglo23

just a thought but the nv silencers exhaust out through one of the pci card openings and you get a slotted version of the origional covers with the card. is this fitted ?

SirLemonhead

The idle temp for the card, according to the Nvidia control panel, is 47 degrees. Haven't got the card overclocked at all.

My Psu is reporting that my -5v line is running at 3.55v.... though on my old chieftec psu I still had the same video problems and the voltages on that were fine.

SirLemonhead

greglo23 wrote:

just a thought but the nv silencers exhaust out through one of the pci card openings and you get a slotted version of the origional covers with the card. is this fitted ?

The original metal bracket for the card is what's on it. Not the double one with slots for the exhaust. The exhaust is just "loose" and is blowing out of the empty case slot.

RicardoSmith

MasterOfSolos wrote:

The idle temp for the card, according to the Nvidia control panel, is 47 degrees. Haven't got the card overclocked at all.

My Psu is reporting that my -5v line is running at 3.55v.... though on my old chieftec psu I still had the same video problems and the voltages on that were fine.

For some reason I thought this card was 2nd hand, and meant if the previous user overclocked it. If its a new card, then I'd say its faulty. If you do any of the PC Mark tests I'd say you'd see the problem staightaway.

SirLemonhead

I did 2 runs through of 3D mark 2005 and nothing out of the ordinary happened. Ran through without artefacts and no lockup.

I did PM the guy I bought it from a few days back but haven't heard back from him, so I don't know. I presume he probably did overclock it if he changed the cooler..

RicardoSmith

MasterOfSolos wrote:

I did 2 runs through of 3D mark 2005 and nothing out of the ordinary happened. Ran through without artefacts and no lockup.
..

Umm well that is odd. Maybe your drivers are bunched? Seems unlikely.

SirLemonhead

Yeah it makes no sense Warcraft III locks up really quickly, but I can play a lot longer in Doom 3 without problem. I've tried lots of different drivers at this stage and none of them made a difference. On the 81.82 betas now.

RicardoSmith

Did you try the card in another PC?

SirLemonhead

Just tried it in a p4 1.3ghz pc with the chieftec PSU and had the same lockup problem on the menu of Warcraft III.

It's a problem with the cooler then?

christophicus

NO it is more then likely a problem with the ram.the problem could be caused by the cooler though.try lowering the clock speed ,and gpu ram frequency.

unkel

You haven't tried the most obvious: running your cpu at stock speed

MasterOfSolos wrote:

The idle temp for the card, according to the Nvidia control panel, is 47 degrees. Haven't got the card overclocked at all

That's what you think

Did you fix the PCI/AGP in BIOS to 33/66MHz? If not it is the most likely cause of your problem

SirLemonhead

Yep, tried it at stock settings. Everything at their defaults, looser ram timings than the ram is specced for etc etc.

The Nforce 2 board has a lock on the AGP and PCI bus set as on by default and i've never turned it off. My previous card never had a single problem.

And the problem has occured in the old P4 machine, which isn't overclocked (Gateway machine, so can't anyway..)

unkel

Last chance saloon:

Make sure your ram is ok by running memtest for a couple of hours. If your ram is ok, then flash the BIOS back to its original (if it has been flashed) if you're comfortable with that

Still problems? Your GFX is fooked...

SirLemonhead

Yeah, i'm running the ATI card at the moment so i'll try get the card back in and run that test. Hooking up a floppy drive again..argh! Not a single problem with the ATI card since I've put it back, i've finished a couple of Warcraft III levels and some on Painkiller without a single problem. Would indicate that the ram is probably ok or can problems become apparent with the other card?

The guy who sold me the card has seen my PM but hasn't replied

RicardoSmith

MasterOfSolos wrote:

Just tried it in a p4 1.3ghz pc with the chieftec PSU and had the same lockup problem on the menu of Warcraft III.

It's a problem with the cooler then?

I'd still reckon its the card. Temps are ok.

unkel

MasterOfSolos wrote:

Not a single problem with the ATI card since I've put it back, i've finished a couple of Warcraft III levels and some on Painkiller without a single problem. Would indicate that the ram is probably ok

More than likely. Give the guy you bought it from a bit of time to reply. Maybe pm him a link to this thread. It is possible he might sort out something on the warranty if he was the original owner

christophicus

when i was talking about the ram i ment the ram that is onboard the gpu.

SirLemonhead

christophicus wrote:

when i was talking about the ram i ment the ram that is onboard the gpu.

Yeah, I was replying to Unkel about not needing to do the memory tests.

There's no way I can test the memory on the GPU is there? I still havent got a chance to take a look at the cooler properly. Don't want to take it off until I can find my tube of thermal paste. Seems to have squishy grey pads for the memory too and I don't know what'll happen those if I pull the cooler off..

EL WOG

a mate of mine havd a problem with he's 6800 and it turned out that he was running it with he's agp port set at 4x, and the cards default is 8x, the only way he could solve it was to get a motherboard that had an 8x agp port.
its a long shot, but you could check the same

RicardoSmith

EL WOG wrote:

a mate of mine havd a problem with he's 6800 and it turned out that he was running it with he's agp port set at 4x, and the cards default is 8x, the only way he could solve it was to get a motherboard that had an 8x agp port.
its a long shot, but you could check the same

Did he have the same screen corruption problem?

MasterOfSolos wrote:

Yeah, I was replying to Unkel about not needing to do the memory tests.

There's no way I can test the memory on the GPU is there? I still havent got a chance to take a look at the cooler properly. Don't want to take it off until I can find my tube of thermal paste. Seems to have squishy grey pads for the memory too and I don't know what'll happen those if I pull the cooler off..

Usually running benchmark tests like 3D Mark will show up theres a problem with the memory. But apparently it runs them fine. Which is not what I would have expected.

SirLemonhead

EL WOG wrote:

a mate of mine havd a problem with he's 6800 and it turned out that he was running it with he's agp port set at 4x, and the cards default is 8x, the only way he could solve it was to get a motherboard that had an 8x agp port.
its a long shot, but you could check the same

Hi Warren

Nah, my board is an AGP 8X board.

Edit: just realised your name is in your profile, so before you start wondering..I know Paul from Youghal. heh.

christophicus

MasterOfSolos wrote:

Yeah, I was replying to Unkel about not needing to do the memory tests.

There's no way I can test the memory on the GPU is there? I still havent got a chance to take a look at the cooler properly. Don't want to take it off until I can find my tube of thermal paste. Seems to have squishy grey pads for the memory too and I don't know what'll happen those if I pull the cooler off..

Those squishy pads are supposed to do the same job as thermal paste but they are reall burish,paste is way better.

SirLemonhead

Don't think I could replace them with paste as the copper pads that sit onto the memory wouldn't reach down onto the tops of the chips without them. Think I have revision 2 of the cooler so the whole copper is flat..

I removed the cooler screws today with great difficulty, realised I wouldn't be able to take the cooler itself off without having to pry it a big, so I gave the whole thing a tiny move to make sure it was seated properly, redid the screws (though not as tight)

Ran 3 consecutive runs of 3D Mark 2003 without problem, and have been playing Painkiller the last hour or 2 and still nothing went wrong So it could be ok.. I'll keep trying different games I guess.

christophicus

Cool thats really good news ,and I CANT BELIEVE THAT I WAS RIGHT!!!!?????.lololol

SirLemonhead

It's been pretty random so far though. I bet i'll reload Painkiller and get a lockup after 5 minutes

Fingers crossed anyway...

Cheers for the help everyone

christophicus

ok good luck.

SirLemonhead

Happened in Warcraft 3 after 15-20 minutes.

Things do seem somewhat better though I think.. must just need new paste on the GPU.

Wouldn't the temps report wrong though if the gpu wasnt being cooled properly?

RicardoSmith

MasterOfSolos wrote:

...Wouldn't the temps report wrong though if the gpu wasnt being cooled properly?

Thats what you'd expect alright. Unless its not reading temps right...

SirLemonhead

Just got in over an hour of Warcraft III play without problem! (could have kept going I imagine but I stopped playing)
I got around to trying the card underclocked. First I tried just the memory, ran that at 900mhz and the GPU at stock. The lockup occured as usual. Then I dropped the GPU to 315 and kept the mem at 900mhz. Seemed to do the trick!

Also, according to Riva Tuner, the fan on the card was set to never go faster than 50% of it's maximum speed. Is this a normal setting or was something changed in the Bios by the previous user? I put the fan up to 100% before I ran the game, but didn't actually notice any increase in noise in my case as a result...

Next step I guess is to try the memory at stock speeds and keep the GPU underclock at 315.

Getting there, slowly. I hope this is just a cooling problem rather than slightly-borked core or whatever

christophicus

I dont think its normal to keep the fan at 50 percent.Push it up to 100% and run the card at stock settings.