PDA

View Full Version : anyone have xfx 6600gt agp


elitebear4
01-19-05, 06:52 PM
im running on the xfx 6600gt agp with the new "improved" heatsink. it ****in sucks at cooling my gpu and doesnt even make full contact with it. anyways ive been wondering if anyone has ever applied as5 to this thing. im having so much trouble getting the heatsink off and it seems like the screws dont work with any of my screw drivers. has anyone else had this issue?

felony
01-19-05, 08:49 PM
I havent tried taking mine off, but I know it runs very hot.

Still trying to decide if I am going to RMA it.. but if i do I get another XfX anyway

dan

Despairl
01-19-05, 10:00 PM
my temps on mine run between 40-42c at idle and 53c at load oc'd at 590/1190......yes i put as5 on mine and also the bridge chip :)

elitebear4
01-19-05, 10:17 PM
despair, are you running on the newer heatsink (without the cartoon)? and if so, did you have any trouble getting the heatsink off?

Despairl
01-19-05, 10:32 PM
Yea that chromed out looking one? yea it was a pain in the ass to get off :mad: :temper: ....the one i got has push pins and u have to be really careful not to break anything off the bottom of the card....espeacially near the bridge chip. But once u have them off and have the as5 on there i'm sure u will notice a big drop.

jet_black
01-20-05, 02:06 AM
I have the cartoony dog heatsink.

I really want to put some new thermal compound on, but I am a little sketchy on how to do it.

Once I have the heatsink off, do I treat the CPU as a normal processor?

DirtyDogLives
01-20-05, 03:32 AM
I was hoping if anyone here can help me out with an issue I have with this card. It runs Command and Conquer Generals, Call of Duty:UO, Doom3 fine with no problems whatsoever......so far (after 6 hours of gameplay). My issue is whenever I run 3DMark03, from the second game test onwards, it artifacts like crazy(like glass breaking all over the screen). It runs through all the way to the end and gives me scores (7100-7200). I have tried drivers 66.81, 66.93, 67.03. I have updated all of my drivers on the system. I have plugged in the molex connector to my new enermax noisetaker 470w power supply, abit nf-7 v2.0, 1gb corsair xms pc3200.

What can I do??? I suspect it's a driver issue or some conflict somewhere as the games seem to work fine so far. It's just very annoying to see that much artifacting in benchmarking. GPU temp according to the nvidia software = 70'C idle and up to 84'C under load (taken directly after 3hours of gameplay on Doom3 and Call of Duty:UO)

Please help me someone!!!!

DirtyDogLives
01-20-05, 03:32 AM
And the videocard is running is stock mode, no overclocking.

ViewtifulMonkey
01-20-05, 03:39 AM
Which one is the "improved" heatsink? I have the chrome one on mine

DirtyDogLives
01-20-05, 03:53 AM
the chrome looking one. not the one with the plastic looking fan housing

DirtyDogLives
01-20-05, 04:38 AM
I just had doom3 lock up on me!!! Temp on GPU reads 78'C after playing.

Despairl
01-20-05, 08:01 AM
i'd say it's because of heat not being able to get to the heatsink. U need to take it off and put some as5 on there. Ok you'll notice in the pictures that i attached, i highlighted the push pins. Take a small tinkering flat bladed screw driver and find on the back of the card the little pin spots. Next you'll notice small little gap between the pin and the pin housing as shown in picture 3. Slightly push down on the pin only...not the housing and the pin. Please take care that the pin is spring loaded and not to lose any of them. Once you have both heatsinks off, apply the as5 to them. Remember not to much, on the GPU i would say about 1/4th the size of a BB would do and on the bridge chip even smaller. Once you've applied the as5 reassemble in reverse order, by pushing the pins in and making sure there is even contact made with the GPU and bridge chip at all times. I hope this helps some.

alinosa
01-20-05, 10:31 AM
Yea that chromed out looking one? yea it was a pain in the ass to get off :mad: :temper: ....the one i got has push pins and u have to be really careful not to break anything off the bottom of the card....espeacially near the bridge chip. But once u have them off and have the as5 on there i'm sure u will notice a big drop.

i wasn't careful with it and managed to shear some components off... be VERY careful.

Memphis69
01-20-05, 11:05 AM
damn that sucks

alinosa
01-20-05, 11:18 AM
i was soooo PO'd... but i've got a 6800U so i can live with myself now.

i just saw your oc on your vidcard... that rocks!

Despairl
01-20-05, 11:56 AM
yea thats one of the worst things about this card is the layout of the chipset.....i mean they put the little micro transistors so damn close to the push pins on the bottom for sure even alittle slip might cause some kind of damage.

alinosa
01-20-05, 12:09 PM
and it was a tiny little screwdriver!

Despairl
01-20-05, 12:13 PM
yea i just got lucky i guess i have very steady hands, it takes some time to do it right....i mean it took me 2 hours just to get the damn GPU heatsink off without damaging anything. But the help of the as5 was worth it.

alinosa
01-20-05, 12:21 PM
the first time i took it off it was fine... i lapped the copper plate on the heatsink. when i tried to take it off to see if a WB would fit...SCRATCH!

Waldo1967
01-20-05, 03:15 PM
I've been tweaking my XFX 6600GT for the last two days and I'm very upset. I ran a whole load of tests comparing my old 5900XT with this card and game after game, the 5900 equals the 6600GT. I spent $200 for nothing!@#$#%@!

I even checked the heatsink today, because my temps were going from 50C to 75C. Took it off no problem, and found it already has paste on it.

Now I'm beginning to think my system is bottlenecking, what do you think:

P4 2.4 @ 2.9
P4P800 Deluxe mobo
1024 MB PC3200 ram
WinXP Pro SP2

I even have my ram set to Turbo in the bios, and am overclocking the cpu.

alinosa
01-20-05, 03:31 PM
i don't know that you're bottlenecking your card with your cpu yet. what resolution are you testing at? the 6600gt has 8 pipes and i thought the 5900 only had four. not to mention the RAM should be a lot more efficient too. Try turning up the eye candy and test at those levels as well. are the tests you're running DX9? That might be a factor as well.