New Video card, constant Game Crashing...

jtzako

Limp Gawd
Joined
May 7, 2005
Messages
299
I recently purchased a BFGTech 6800GT OC 256MB AGP card.. The first one I got I had to return as It instantly locked up if I tried to run a 3D Game.

I got another Identical one from the same store but I still think something is wrong..

Prior to this upgrade I had a 9800Pro in my system.. no problems at all with that one..

Now, I get a couple problems with the 6800GT

1) Games Crash.. Constantly.. Kotor2 crashes every 5 to 10 minutes. SWG crashes about Once an hour. Usually I see a very brief BSOD then computer reboots. I cant Identify what the error is and nothing shows up in event viewer. However, it doesnt always reboot, sometimes it just Hangs. The Latest 3DMark will not even play completely through the first test before lockup.

2) When I am in a Game such as SWG its very smooth except for slight "hiccups" every several seconds. Its almost like everything just Jumps briefly and continues on. This is very noticable in Space (if you've played SWG you know) because the enemy ships bounce around when this happens.


My System:
AMD64 3000+
1Gig Memory (2 sticks of 512, tried 2 different sets so far)
MSI K8N Neo-Platinum motherboard
SATA HDD
Sony DVD Burner (IDE)
Enermax 420 WATT PSU
Windows XP SP2

Everything I have has been updated to the latest Drivers and BIOS
Nothing is overclocked other than the "default" OC from BFGTech

12Volt rail reads as 12.10
Under load my CPU and GPU temps reach about 65C which I'm told by both MSI and BFG is well within normal range for them

Do you guys think this Card is also faulty somehow? Im almost hoping it is and I can just return it but I'm afraid Ill end up getting another Lemon now (since this is my 2nd one of these)

I also removed the 6800GT and put back my 9800 Pro and it ran perfectly for hours at a time and can complete the 3DMark tests also.. I have used Driver cleaner to get rid of extra drivers. I have defragmented the HDD using Diskeeper Pro and I have around 5+ Gig free on the C Drive.
 
uninstall driver in window SAFE MODE

then reinstall latest driver from Nvidia OR get the non-offical driver (check out nvidia video card thread)
 
holycpu said:
uninstall driver in window SAFE MODE

then reinstall latest driver from Nvidia OR get the non-offical driver (check out nvidia video card thread)

Did that stuff already..I always use safemode to uninstall (assuming it works as some apps dont) and I also used driver cleaner to make sure no remaining drivers.

I've also tried using Omega "unofficial" driver and had same issue.
 
What tempertures do you get? Is it possible the video card is overheating?
 
Your CPU should not be at 65C under full load, can you be more detailed on that? Also the 12V rail reading, where are you getting it from. Open the case, on the side of the PSU, write down what the 12v+ ratings are and the total wattage.
 
Did you uninstall all the ati stuff? Are fastwrites disabled in bios? What is the agp voltage set to?
 
FanATIc said:
Your CPU should not be at 65C under full load, can you be more detailed on that? Also the 12V rail reading, where are you getting it from. Open the case, on the side of the PSU, write down what the 12v+ ratings are and the total wattage.

I am definitely CPU at 65C at full load. Verified by the Mobo sensor and a 3rd party one. If 65C is bad I've never seen anything to indicate that.. If you think its wrong please be specific as to why. I am a computer tech and I can handle the real explanations.

My 12V rail reading is also from the sensors 12.10 all the time as far as I can tell.

I have an enermax 420watt PSU Whisper II EG465P-VE FMA
+12V1 is 15A
+12V2 is 14A
Total power is 420W
 
[F8] said:
Did you uninstall all the ati stuff? Are fastwrites disabled in bios? What is the agp voltage set to?

I have definitely uninstalled ATI stuff. I even use driver cleaner to make sure its gone. I also remove the Nvidia stuff to try different drivers.

I have fastwrites on "Auto" in the BIOS so it should be driver controlled. Should fastwrites be off? (If they should be off please explain why this would not hurt performance)

AGP Voltage says 1.50 V in the BIOS and is not changable
 
jtzako said:
I am definitely CPU at 65C at full load. Verified by the Mobo sensor and a 3rd party one. If 65C is bad I've never seen anything to indicate that.. If you think its wrong please be specific as to why. I am a computer tech and I can handle the real explanations.

My 12V rail reading is also from the sensors 12.10 all the time as far as I can tell.

I have an enermax 420watt PSU Whisper II EG465P-VE FMA
+12V1 is 15A
+12V2 is 14A
Total power is 420W


Why? 65C is above what a prescott works at :). You're cpu should never go above 50C when its running full load, even that is a little hot. You might want to remove the CPU heat sink check the CPU out, make sure thermal paste is applied, and check to make sure the heat sink is making good contact with the CPU. No A64 should be running that hott. You have a serious heat issue though, certainly will cause BSOD and crashing. For a 6800GT thats an ok temperature, but for an A64 thats rediculously hot. Way way above normal.
 
FanATIc said:
Why? 65C is above what a prescott works at :). You're cpu should never go above 50C when its running full load, even that is a little hot. You might want to remove the CPU heat sink check the CPU out, make sure thermal paste is applied, and check to make sure the heat sink is making good contact with the CPU. No A64 should be running that hott. You have a serious heat issue though, certainly will cause BSOD and crashing. For a 6800GT thats an ok temperature, but for an A64 thats rediculously hot. Way way above normal.

Ill look into that but im pretty sure the thermal paste is good. This CPU has been running like this since day 1 and I only recently got the 6800GT. I had a 9800Pro for a very long time and never had any issues.

AMD says the following about their AMD 64 Processors

Interenal Error Free Temp = 65C
Surface Error Free Temp = 55C

Im probably on the border of the temps they list as "error free" so Ill definitely see what can be done there.. but its not like its out of bounds persay..
 
jtzako said:
Ill look into that but im pretty sure the thermal paste is good. This CPU has been running like this since day 1 and I only recently got the 6800GT. I had a 9800Pro for a very long time and never had any issues.

AMD says the following about their AMD 64 Processors

Interenal Error Free Temp = 65C
Surface Error Free Temp = 55C

Im probably on the border of the temps they list as "error free" so Ill definitely see what can be done there.. but its not like its out of bounds persay..

If its been running like that from day 1 then thats an even better chance it was compeletly installed wrong. It wont destroy the CPU to work at that, but its not healthy either. It can cause the problems you are describing since it will throttle and fail on occasion. Make sure the heat sink/fan arent crap either. Some no names out there that do absolutly nothing. Check for dust as well.

You can post this in the AMD forum if you like further proof then my word, but im telling you thats crazy hot for that processor and would absolutly cause all the problems you're describing.

A 3000+ 130nm proc should be idling in the low 40s and hitting mid 50s tops during load.
A 3000+ 90nm proc would idle mid 30s and top at 45

edit- i'd completely ignore their little safe temp charts from all CPU and GPU manufacturers. I have found them to be very static and totally incorrect. ATI rates the R400 series core at a failure temp of 120C or some crazy number, but i'v noticed continous failures on multiple versions if the GPU goes 80+. Artifacting usually, then alot of lock ups.

CPUs are the same. Intel has some rediculous number but it was the same deal, 70-75C and i'd get BSOD and lock ups. 64AMDs run much cooler then Intels, talking a 10-15C difference, but yours is running hotter then a standard Revision 1 Prescott processor which had problems with warping silicone over time from the heat.

Even IF you find no problem you should absolutly go out and get a high quality heatsink/fan for that CPU. Rest assured its not your graphics card though.
 
FanATIc said:
If its been running like that from day 1 then thats an even better chance it was compeletly installed wrong. It wont destroy the CPU to work at that, but its not healthy either. It can cause the problems you are describing since it will throttle and fail on occasion. Make sure the heat sink/fan arent crap either. Some no names out there that do absolutly nothing. Check for dust as well.

You can post this in the AMD forum if you like further proof then my word, but im telling you thats crazy hot for that processor and would absolutly cause all the problems you're describing.

A 3000+ 130nm proc should be idling in the low 40s and hitting mid 50s tops during load.
A 3000+ 90nm proc would idle mid 30s and top at 45

One thing I should note.. Im starting to wonder if the heat sensors are accurate.. an AMD tech posted a suggestion for a rough "estimate" of the CPU temp.. He said to touch the base of the heatsync and see if its Warm, Hot, Burns to the touch etc.. I can barely tell mine is warm (midway through a 3DMark test which definitely puts it to full load)..

It seems to me that if it was really 65C then it couldnt possibly be that cool to the touch at the heatsync base.. Unfortunately both sensors I have available are made by the mobo company so I cant say whether they are accurate.
 
FanATIc said:
Rest assured its not your graphics card though.

I never crashed in any game or test (including a 24hr Prime95 run). I installed the BFG 6800GT OC and within 10 minutes of Gameplay or First test in 3DMark it would crash. This is why I suspect it has something to do with video. My system temps have not budged since intalling that card either.

I will be looking into the heatsync issue for sure though. I am using the stock heatsync/Fan combo with ArcticSilver paste currently.
 
jtzako said:
One thing I should note.. Im starting to wonder if the heat sensors are accurate.. an AMD tech posted a suggestion for a rough "estimate" of the CPU temp.. He said to touch the base of the heatsync and see if its Warm, Hot, Burns to the touch etc.. I can barely tell mine is warm (midway through a 3DMark test which definitely puts it to full load)..

It seems to me that if it was really 65C then it couldnt possibly be that cool to the touch at the heatsync base.. Unfortunately both sensors I have available are made by the mobo company so I cant say whether they are accurate.

possible they're inaccurate. you could update the bios as well. A heatsink will be warm to the touch though on any processor, so i really think its either making crappy contact or no contact at all with the processor.

Who made the PC, whos the motherboard manufacturer?

jtzako said:
I never crashed in any game or test (including a 24hr Prime95 run). I installed the BFG 6800GT OC and within 10 minutes of Gameplay or First test in 3DMark it would crash. This is why I suspect it has something to do with video. My system temps have not budged since intalling that card either.

I will be looking into the heatsync issue for sure though. I am using the stock heatsync/Fan combo with ArcticSilver paste currently.

But you said its your second time this happened though. Its extremely rare to get a double screwed RMA. I'v gone through 13 cards this year alone and never had a problem with any of them technically. Since it doesnt appear to be a driver problem, it has to be something else. And in what you have posted so far, the only thing totally odd is the CPU load readings. The rail seems to be a bit over volting for some reason as well. Though i dont know enough about PSU's to tell you anything more then that.
 
You should definitely try disabling fast writes and also slow you AGP slot down to 4X. That solves crash problems for lots of people. The both of them together barely have a measureable affect on performance (a few at percent at most and usually nothing). You can do it in the BIOS or usually with CoolBits or RivaTuner.
 
UncleMacro said:
You should definitely try disabling fast writes and also slow you AGP slot down to 4X. That solves crash problems for lots of people. The both of them together barely have a measureable affect on performance (a few at percent at most and usually nothing). You can do it in the BIOS or usually with CoolBits or RivaTuner.

Im currently testing the system again to find out what causes it to destabilize.

The 12.10 on the +12 is dead on according to the PSU company and several other places I've referenced (Including BFGTech support)

I did try disabling Fast Writes in the BIOS as well but I think it was already being disabled by the drivers so not sure if that will help.

I have gone through 2 3Dmark full runs without a crash tonight after all the other tweaking I've done. Will be trying my Game soon (the one that usually crashes most frequently)
 
FanATIc said:
possible they're inaccurate. you could update the bios as well. A heatsink will be warm to the touch though on any processor, so i really think its either making crappy contact or no contact at all with the processor.

Who made the PC, whos the motherboard manufacturer?



But you said its your second time this happened though. Its extremely rare to get a double screwed RMA. I'v gone through 13 cards this year alone and never had a problem with any of them technically. Since it doesnt appear to be a driver problem, it has to be something else. And in what you have posted so far, the only thing totally odd is the CPU load readings. The rail seems to be a bit over volting for some reason as well. Though i dont know enough about PSU's to tell you anything more then that.

Its a self built PC. MSI K8N Neo-Platinum motherboard. BIOS is updated to the latest available but I dont go by BIOS temp readings generally. One program I have that can read my sensors is called Everest Home Edition. Its been pretty accurate on my other PCs.

The only thing I havent done is try the card in other computers. Mainly because their PSU are max 11.86 on the 12V Rail and BFG Tech doesnt recommend that for this video card.
 
So after making a few changes to my system I played for several hours last night, including 3 full run 3Dmark05 tests and did not crash or spontaneously reboot.

I switched power from the normal rails to the "Extra" one per Enermax manual.

I swapped RAM out from my other system (I had done this the day before but hadnt really tested it more than 30minutes at that point)

Completely removed all Nvidia drivers from Safemode then reinstalled a fresh copy again

Disabled "Fast Writes" in BIOS

----------------------

So far I have not had the chance to check into my potential CPU heat issue but maybe Monday I can stop by a store and get a good CPU heatsync/Fan for that issue.

I plan to leave the system exactly as it is now and let it run games and tests all night again tonight to see if it has any more issues.
 
Back
Top