Dying video card?

[mauser]

This started happening a couple days ago, and I had just got this card (9700 Pro) off of Ebay a couple of months ago to replace my 9500 Pro which was also giving me problems.  It started happening about the time I also replaced my monitor (Sony GDW-FW900 24 in).  I had previously been running 1024x768 in game and on desktop, but with the new monitor I started running it at the recommended 1920x1200 and 1600x1200 in game.  Would that have hastened this card's demise?  




Just yesterday I completely cleaned out any old drivers with ATI's driver uninstall utility, SMARTGART uninstall utility, and then Driver Cleaner.  Then reinstalled the old Cat 4.11 drivers.  Today I pulled out the card and didn't notice any dust build up, and the fan was running fine.  Didn't notice anything out of the ordinary upon a cursory inspection of the card.  

If this is the death knell for this card, that's two cards in less than a couple months - never had video cards die on me since I had a Voodoo1.  

mauser

[StarOfAfrica2]

I would imagine either a driver problem or a problem with the video settings before I'd say the card is bad.  First thing I'd check is that your video card supports those resolutions you are running.  Thats some really fine resolution there.  Its also going to eat up your resources running at that high a resolution, even if the hardware is fully capable of supporting it.  Also, to run that high, I'd say you are going to have to give up some quality settings somewhere else.  Try setting all your video card settings to performance and turn off AA and AF if you have them on.  Make sure you slide all your ingame settings to performance too.  See what happens.

[Schutt]

If the card has a fan on it check that, also clean it from dust.
It could be a heat problem, so after you checked that the card really supports that high resolution check for heat problem, that is open the case and blow in air somehow to cool it more.

Also its intresting if this happens after long playing time or right from switching on the comp.

[JB73]

looks like overheat problems

[Swoop]

Mauser, the second pic you've posted looks exactly like a problem I have with my card (9800).  It's not cooling, pull the card out and have a good long look at the connectors that run along the bottom of the card (the bit that slots into the....er.....slot).

My card has a slight nick on the connectors, I cure the problem by jamming a CD case in between the PC case and the card when it's in the slot, it pushes the card to the right and gets a better connection between the card and the PC.  No more problems.

And yes, I've had a CD case stuck out of my PC now for 6 months. It looks daft but hey, it works.

[38ruk]

i would download ati tool and check the temperatures on the card , even if there isnt any dust on or in the fan , you still might have a termal paste/ pad problem under the heatsink. i would also check the voltages for your agp slot , just to be sure that its running a 66mhz . or auto .  9700's werent good overclockers , so people would do a volt mod on them to get stable overclocks. since you bought it off ebay i would do alittle research on the mod and check if your card been modded. just a shot in the dark , but a possibilty .... gl   38

[humble]

Originally posted by 38ruk
i would download ati tool and check the temperatures on the card , even if there isnt any dust on or in the fan , you still might have a termal paste/ pad problem under the heatsink. i would also check the voltages for your agp slot , just to be sure that its running a 66mhz . or auto .  9700's werent good overclockers , so people would do a volt mod on them to get stable overclocks. since you bought it off ebay i would do alittle research on the mod and check if your card been modded. just a shot in the dark , but a possibilty .... gl   38

Assuming your other card went south I think ruk hit the nail on the head. Either your MB or power supply is probably going and spiking the voltage to the card. Most MB's have a utility that reports voltages. If your not getting the right voltage to the card your hosing it. One thing you can do is pull the power connectors and look for discoloration...if you see any of the pins have dark areas then either your PS is cooking the MB or the MB is going and toasting your PS....but I'd bet you've got a power issue at the root of the problem since it two different cards.

[mauser]

After I posted the screenshots, I uninstalled the drivers once again using the ATI cat uninstall utility, then booted to Safe Mode.  When the animated Windows 2000 logo screen came up, there was a grid of short black vertical lines on the screen.  They didn't go away after Safe Mode was completely loaded either - and Safe Mode was running with Standard VGA video drivers.  I have a screen shot of the desktop I'll post later when I get home in a couple hours.  Also of note is that I took the desktop and game resolutions down to 1280x1024 (and went from 512 to 256 textures) but it didn't cure the problem.  The card wouldn't even complete 3dmark03 - I get diagonal lines of square blocks on the screen as well as the "crumpled" textures.  

I'll take a look at the MB voltages in BIOS also.  One of the things I did try was take AGP voltage up to 1.7 I think (the first step off of the "Auto" setting), which didn't help yet.  

Then I'll take another look under the hood - I did pull the card out yesterday and took the heat sink off to see if it's the pad or compound.  Appeared to be compound.  Will check the AGP slot connectors also (maybe try to jiggle the card in the slot while the machine is running).  

Although I'm worried that the card is dying I'm suspicious about the MB also.  It is an MSI KT3Ultra2 which had been fine for years.  The power supply is an Antec 420W TruPower which was a recommended unit.  So many things to worry about...

mauser

edit: here is the safe mode screen

[SkyChimp]

I had the same problem with my 9800 pro (Heat Problem)


Get a small fan put it over video card anyways thats what i did.

[Callisto]

Originally posted by mauser
I started running it at the recommended 1920x1200 and 1600x1200 in game.  Would that have hastened this card's demise?  


mauser

Maybe you're pushing it too much with resolutions..

AH2 @ 1920X1200 on radeon 9700 pro.. thats crazy..

Even 1600X1200 is too much for that card, especially in AH2.

[humble]

Heat should not be an issue at start up, the card should get thru basic boot up....

Most major boards have a 4 or 6 layer laminate and over time they will crack do to heat. That can trigger problems directly or create a problem that leads to PS failure. Do you have any ability to swap the card to another system...you need to determine if the card is functioning...everytime you use your current system you might be further damaging the card...

[MANDO]

If you have these effects in safe mode just after switching on the computer, then it is not related to heat, neither refresh rate, nor drivers.

Try to switch on the computer with the monitor turned off (may be some dangerous voltage is comming from the monitor to the card by the vga connector). Keep the monitor switched off several minutes, turn it on and see if the effect was already present, if not, then the problem may be in your monitor.

If it is incorrect AGP voltage, you may change it (+- 0.1v) from the bios and see what happens. If it doesnt work, try to unplug and plug the card, be sure it is perfectly perpendicular to the mainboard to ensure the best possible contact with he AGP slot.

[mauser]

After a day of trying different things on my machine and also dealing with a bad HD on my father's computer, it looks like the AGP connection was the culprit.  I took the card out and inspected the contacts - they looked reasonable so I just got an eraser and polished them a bit.  Reinserted the card and noticed that I got it to sit slightly differently than before.  Instead of tightening the case-card screw I left it loose and ran 3dmark03 again.  No artifacts this time but jiggling the card lightly in the slot caused 3dmark to "lose focus."  Tightened the card down while making sure it was as straight as possible.  So far, I ran 3dmark03 a few more times without a hitch, and left my screensaver (3d nature with heavy textures and transparent water) on for most of the late afternoon.  The screensaver had been freezing a lot before, now it continued to run merrily on its own.  Tried ATI Tool also - unfortunately I guess the 9700 Pro doesn't support temperature monitoring.  I did run the artifact checker and got one artifact error after about a min, but nothing really significant.  

Haven't had a chance to play AH for any extended period, although offline everything looked back to normal for the 5 mins I ran it.  Will have to try tomorrow when I have more time on my hands.  

Thanks for your suggestions!  I had no idea that slot connections could get fidgety like that.  The connections seem to be pretty firm down there.  I'll have to cross my fingers and hope it doesn't come back.  

By the way, for some reason I had thought that you should turn the monitor on first, then the system.  Have I been doing it wrong all this time?

mauser

[mauser]

Well, crossing my fingers didn't work.  The problems came back again soon enough and didn't go away.  I reformatted and reinstalled Windows 2000 but the lines on the screen remained.  I then took both cards (9500 Pro and 9700 Pro) to my parents house and tried them in their machine.  The 9500 wouldn't boot at all and the system would beep multiple times, just as it did in my system (BIOS Beep Codes).  The 9700 still had the short vertical lines all over the screen.  So now I have two dead or dying video cards and am unable to play AH since the spare video card from my brother in law is a DX8.1 only NVIDIA Quadro NVS 280.  

I am now contemplating a new mobo, cpu, and video card.  It will most likely be an nforce4 board with an Athlon64 Venice core cpu and some PCI-x video card.  I figure I might as well go to PCI-x since both my AGP cards are no good.  However, I'm affraid of having this problem occur once again (and wasting hundreds of dollars in good hardware plus pissing off the wife).  Could running the video cards at such high resolutions (1900x1200 desktop, 1280x1024 or 1600x1200 in AH) really ruin a video card?  Or could the card have been fried by voltage spikes?  Here's how my rail voltages look from the BIOS screen:

Vcore = 1.728 V
+5.0V = 4.975-5.003 V
+12.0V = 11.741-11.802 V
-12.0V = -11.825 V
-5.0V = -4.926 to -4.976 V
Batt = 3.328 - 3.360 V
+5V SB = 4.921 - 4.945 V

mauser

[StarOfAfrica2]

You have to look at what resolutions the video card can put out.  If the video card doesnt support the kind of resolutions you want to run, then heck yeah pushing it can burn it out.  Usually its the monitor that is the limiting factor, but if you have a monitor that wants to run at resolutions like that, you'd better be prepared to buy a good video card.  Check the specs before you buy.



Most cards (including the 2 you listed in your post) support a max resolution of 2048x1536, but thats only at a 60hz refresh rate.  Matter of fact, once you get above 1024x768 I dont think there are many that go beyond 60hz.  If you pushed your refresh rate above the recommended limits, it could have caused the burnout, even if the monitor could support it.