GTX 260 probs continued - help/opinions plz

[njj1046]

I recently bought a Zotac GTX 260 Amp! edition (55nm, 216 shaders) from OcUK. It was factory overclocked to 650 core, 1400 shaders and 2100 memory.

I had problems with it straight away, and originally posted for help Here ...Only read of you have too much time on your hands

I decided to RMA it to Overclockers and a few working days later I received the same card back because they couldn't find any errors. Included with the card was a report saying which tests they'd done and screenshots etc. They tested it with crysis benchmark, world in conflict, furmark, 3dmark06 and grid. The conclusion of the report was:

"There was no fault found. Please check your other hardware or software installation" and "No fault was found GFX card passed all tests. When testing with grid the 1.2 patch is required as without graphical corruptions appear."

I put the card back in my machine and had the same problems immediately.

So then I took some pics of the texture corruptions with my camera phone and made a post about the problem in the 'OcUK Customer Service' forum including the pics. I also stated in that post that with the card underclocked to default GTX 260 speeds the card works flawlessly. Here are the pics..


Call of Duty: World at War after 15-30 mins of play..

Grid (with v1.2 patch installed) after a few minutes of play..

Half Life 2 Ep2 after a few minutes of play..

Call of Duty 4 after approx 45 minutes of play..

I got a reply there saying it could be an issue with my power supply or a strange compatibility issue with another part of my system Also raising my PCI-E frequency to 105 in the BIOS was suggested (which I tried and didn't make any difference).

My power supply is a brand new Be Quiet 750 Watt unit and my system is 100% stable and I never had any problems at all with my trusty 8800GTS 512.

I installed ATITools this afternoon as well and ran the artifact tester. I went off to make a cup of coffee and came back to a lot of beeping and this..

I just want to re-iterate here that these problems only occur when running the card at the factory overclocked speeds. When I underclock it to default GTX 260 speeds, there are no problems whatsoever.

Between sending the card back and being charged for testing and return postage I've spent an additional ~£28 on top of the price of the card.

Any ideas or suggestions will be most graciously welcomed.

 

[Jono8]

ignore.

yeh send it back saying it can't handle the factory overclock

 

[LoadsaMoney]

If it only works at stock clocks, then it can't handle the OC Zotac have put on it, faulty card needs RMA'ing, send a webnote and link to this thread so they can see from the pics.

 

[RJC]

Does look like the card can't handle the overclock, what are the temps like, is there a possibility to pu the fans on full speed and see what happens - Edit you might of tried this already

 

[njj1046]

If it only works at stock clocks, then it can't handle the OC Zotac have put on it, faulty card needs RMA'ing, send a webnote and link to this thread so they can see from the pics.

It's already been RMA'd m8.. because they couldn't find a fault when they tested it, they sent it back to me and charged me for testing it and sending it back to me.

I posted up these pics in the 'OcUK Customer Service' section of the forum. In there posts can only be viewed by the posters and OcUK staff/moderators. Same response there.. basically they reckon it's not a problem with the card, it's a problem with my machine

RJC: The card gets unrestricted airflow from the 120mm intake fan and wouldn't go over 80 degrees C with furmark in extreme burn mode.

 

[HeX]

I totally feel for you, I have had the EXACT same problem, and the exact same response from OCUK regarding the problem, only difference is I had an EVGA card:

HERE

Had the same report back saying it was fine, and they could find no problem, yet every test I do shows the card can't handle the stock overclock.

In the end I have RMA'd it back to EVGA and am awaiting their report.

As soon as they confirm the card is faulty I will be going back to OCUK for a refund on the testing and shipping charge.

 

[Biffa]

Before you go crazy see if you can get hold of another PSU to test with. I was going to get a BE Quiet range PSU when I got went from a 9800GT setup to a GTX260 setup and opted for the Corsair instead because of general opinion that the BEQuiet PSUs weren't up to the job.

Not saying difinitively that is the case, but its just something you may want to check out because the stock overclock will be making the card draw more power than the "default" GTX260 clocks.

 

[njj1046]

I totally feel for you, I have had the EXACT same problem, and the exact same response from OCUK regarding the problem, only difference is I had an EVGA card:

HERE

Had the same report back saying it was fine, and they could find no problem, yet every test I do shows the card can't handle the stock overclock.

In the end I have RMA'd it back to EVGA and am awaiting their report.

As soon as they confirm the card is faulty I will be going back to OCUK for a refund on the testing and shipping charge.

Good luck with that m8, I'll keep an eye out for any updates to your problem.

Before you go crazy see if you can get hold of another PSU to test with. I was going to get a BE Quiet range PSU when I got went from a 9800GT setup to a GTX260 setup and opted for the Corsair instead because of general opinion that the BEQuiet PSUs weren't up to the job.

Not saying difinitively that is the case, but its just something you may want to check out because the stock overclock will be making the card draw more power than the "default" GTX260 clocks.

I've still got my old Enermax Liberty 500W unit here. I'll give that a try just to be 100%

Strangely enough this new Be Quiet unit is giving me concerns atm. These units run the PSU fan, and any other fans you have connected to the PSU for 3 minutes after you power your PC down.

When I first had it, the noise it made whilst doing this was barely audible... now it's making a loud pulsating humming noise after powering off. Today I also heard a strange buzzing noise. I really hope that's the side of my case rattling

 

[Dutch Guy]

It looks like memory errors to me, possibly overheating as it only happens after playing for a while.

Can you hear the fan increase in speed (and noise) when gaming?

 

[njj1046]

I just ran Furmark in xtreme burn mode for 10 mins using GPU-Z to monitor the GPU temp and fan.

The fan at default runs at 40%. When the GPU temperature started to climb above 70c the fan speed increased also until it settled at 81c and 56% fan speed.

 

[Greebo]

That's a pretty hot card.

My EVGA running furmark hits 70C max at fan speed still at 30% with the clocks set at 729/1512. Never tried to test it at stock.

I do find these GTX260 do not like the heat and certainly make overclocks unstable way below their heat limit or fan profile.

For example I need to keep the card below 60C to keep it stable at 780/1572 which means ramping the fan up to 100%.

 

[Rroff]

Its both a blessing and a curse... if you can keep them really cool i.e. under 40C in extreme burn, they will clock pretty much as far as you can move the slider

 

[njj1046]

That's a pretty hot card.

My EVGA running furmark hits 70C max at fan speed still at 30% with the clocks set at 729/1512. Never tried to test it at stock.

How on earth do you manage to keep a GTX260 at 70C running Furmak with 30% fan with that overclock??

The test report I had back from the RMA stated that the average temp for my card when they tested it with Furmark was 80C.

 

[Greebo]

How on earth do you manage to keep a GTX260 at 70C running Furmak with 30% fan with that overclock??

The test report I had back from the RMA stated that the average temp for my card when they tested it with Furmark was 80C.

Lucky card I guess? Like I said, mine runs really cool and never steps up from 30% unless I go for silly overclocks.

I have a nice case though with a 120mm fan sucking nice cool air from the outside directly into the card. That might help

 

[njj1046]

I also have a 120mm intake fan blowing cool air directly over the gfx card.

 

[Dutch Guy]

My GTX280 hits 84C in Furmark and the fan runs at 100%, 3000rpm but it is stable.

The problem is that the memory temperature isn't measured and if that is causing the problem you have no idea about that.

 

[njj1046]

My GTX280 hits 84C in Furmark and the fan runs at 100%, 3000rpm but it is stable.

The problem is that the memory temperature isn't measured and if that is causing the problem you have no idea about that.

Some people may consider 84C to be too hot, but it's not that bad for a GTX200 chip running Furmark flat out. Also as you said, it doesn't cause your card any problems.

I've done some more testing with mine and ATITool... The artifact scanner starts beeping when my GPU temp reaches 70C then goes absolutely crazy when it hits 74C and the screen fills up with yellow artifacts... it just shouldn't be having such problems at that temperature. 74C is a reasonable temperature for one of these chips.

As you said though, monitoring the temp of the memory chips is out of the question. That might be playing a part in the problems too, I don't know.

 

[Greebo]

Some people may consider 84C to be too hot, but it's not that bad for a GTX200 chip running Furmark flat out. Also as you said, it doesn't cause your card any problems.

I've done some more testing with mine and ATITool... The artifact scanner starts beeping when my GPU temp reaches 70C then goes absolutely crazy when it hits 74C and the screen fills up with yellow artifacts... it just shouldn't be having such problems at that temperature. 74C is a reasonable temperature for one of these chips.

As you said though, monitoring the temp of the memory chips is out of the question. That might be playing a part in the problems too, I don't know.

Like you say the card is designed to take those temps and more so definitely something wrong. Send them the pics and RMA it directly to manufacturer. Then when you get a replacement with a report, ask ocuk for a refund of your postage costs to them.

 

[Dutch Guy]

I stopped using ATITool for stability testing, my old 8800GTS and now my GTX280 give corruption at stock speeds in ATITool but never in any game or other tester.

 

[HeX]

I stopped using ATITool for stability testing, my old 8800GTS and now my GTX280 give corruption at stock speeds in ATITool but never in any game or other tester.

It might just be that the artifacts they are giving are too subtle to notice.

I have found that if ATi Tool is giving artifacts, the card isn't fully stable, you will find at least one game/engine that will cause your card to do something it isn't meant to.

For example the card i've just RMA'd, gets arifacts in ATi Tool at stock clocks, will loop Crysis/Furmark/3dmark all day long and you won't notice anything.

Load a Source game and you'll get chequer boarding on model textures.

Obviously the other game engines don't use the GPU in the same way, or stress the same areas.

I had the same with an old 8800GTS 512, was solid in everything, but if I ran CoD4 i'd get model vertex errors with polys stretching over the screen etc, yet every other game would be fine.

In both cases the problems would go away if I downclocked the card.

So both cards were on the edge of stability and only showed errors in certain situations, and would look in 90% of stuff perfectly stable and fine.

ATi Tool knew otherwise however.

So I wouldn't be too quick in writing those test errors off.