New Arctic Cooler NV5rev3 for 6800GT - disappointing results

I installed this last night and saw absolutely no difference in gpu temps, either at idle or load, which is very dissapointing. the only thing i can think that might have gone wrong is that i removed the thermal pad that came with it (which has some thickness, maybe 2 or 3 mm), and used AS5 on the gpu. Is it possible that the heatsink doesn't make it all the way to the gpu so that the AS5 isn't being spead over the gpu? I hand tightened the four screws on the rear clip as much as i could, and visually checked to see if the heatsink was in contact with the gpu, and it looked like it was. maybe i just need to wait until the AS5 cures?

Not that I know a lot about computers, but.. probably best to wait to see the results after it cures. If this happens to be a second computer (meanin you have another one around to use), I'd let it'd stay completely off and let it cure for 15 hours. If not, then don't do anything stressful on the card til about 24 hours pass.

AS5 cures after around 200 hours or so OF USE, but I've seen results on cpu's within a few hours. :shrug:

200 hours? holy christ, thought it was less

sorry :/

I'm not sure if thermal pads can be replaced with thermal grease, though I see no reason why not.


What's the comparison in noise? Are you sure you're not running it in some low RPM mode?

I would check to make sure all the contacts are good.

Also, how much paste did you use?

And no, the thermal pads cannot be replaced with grease. The pads are there becuase there is open space between the heatsink and the ram, which is what the thick pads are for.

EDIT: and while AS5 has a long curing time, this does not mean it does not work until it has cured.

you can replace thermal pads with paste as long as the thermal pads aren't filling a gap in that the paste can't fill. in other words, if the heatsink can make full contact with the surface of the chip without a thermal pad, then you can use paste.

i honestly don't see a situation where a chip manufacturer would design something so that there is an actual gap that REQUIRES a thermal pad. it doesn't make sense.

you can replace thermal pads with paste as long as the thermal pads aren't filling a gap in that the paste can't fill. in other words, if the heatsink can make full contact with the surface of the chip without a thermal pad, then you can use paste.

i honestly don't see a situation where a chip manufacturer would design something so that there is an actual gap that REQUIRES a thermal pad. it doesn't make sense.

my thoughts exactly. i've replaced thermal pads with top shelf paste literally dozens of times, so i can't see why this would be an issue this time. i visually inspected the contact between heatsink and gpu and it looked solid to me, so why the fuck are my temps still so high?

you can replace thermal pads with paste as long as the thermal pads aren't filling a gap in that the paste can't fill. in other words, if the heatsink can make full contact with the surface of the chip without a thermal pad, then you can use paste.

i honestly don't see a situation where a chip manufacturer would design something so that there is an actual gap that REQUIRES a thermal pad. it doesn't make sense.

The actual spec for the 6800 utilizes two sets of heatsinks ont he card itself, one set high that goes on the core and a lower set that goes on the memory. There is a noticeable gap between the base of the NV silencer 5 and the memory modules, thus why there are pads and not just grease.

note: the revision 1 did have extra plates of copper soldered to the base of the heatsink to make direct contact with the memory, but the soldering process of the plates created too many failed units since the plates could move while the solder was drying. Thus why they replaces the plates with thermal pads in rev 2 & 3.
If they wanted to do a good job making direct contact with the memory, they should have doen what leadtek did with their heatsinks, where the base of the heatsinks bulges out to contact the memory. By far a more expensive solution to execute, but will have a lot less failed units.

my thoughts exactly. i've replaced thermal pads with top shelf paste literally dozens of times, so i can't see why this would be an issue this time. i visually inspected the contact between heatsink and gpu and it looked solid to me, so why the fuck are my temps still so high?
Also the temp readings are from the core, not the memory. So that doesnt even matter....

What is the ambeint temp reading from both your mobo and video card say? and also the temps of the card's core for that matter...
Idle and load temps please


EDIT: as a note, your idle temps should hardly show any change at all, it is the load temps that you should be watching

What is the ambeint temp reading from both your mobo and video card say? and also the temps of the card's core for that matter...
Idle and load temps please

System idle: ~38^ C
System loaded: ~42^C
CPU idle: ~40^C
CPU loaded: ~45^C
GPU idle: ~55^C
GPU loaded: ~75^C

as you said, it's gpu load temps i should be looking at, and it hasn't dropped at all.

i did take the card out and tighten the fastener with a screwdriver, which seemed to make a minor difference (1 or 2 ^ C). also, i'm 100% sure the heatsink is in good contact with the gpu.