applying thermal paste to C2D

Dudeyourlame

[H]ard|Gawd
Joined
Jul 23, 2004
Messages
1,085
I recently noticed that when i applied my thermal paste when i built my computer last year that i did it wrong.. so i redid it to what Arctic Silvers site said.. My cores are now a lower temp but one is 32 C. and one is 37C is that normal?

So am i doing it right??

computer026xl9.jpg


computer028sy5.jpg
 
thats not bad...I would use about 1/2 of that though...the length of the line the same, just use 1/2 the ammount and stretch it..

I did it like im describing to you, and i get 22core0 23core1 idle and 41core0/1 at load with coretemp software. using a tuniq tower,

but its pretty common to have one core more than the other...i would still use less thermal paste though....although if its all installed, unless you want to overclock it more, and get the lowest possible temps, i wouldnt take it all apart just to change the paste...maybe if i was in there doing somthing else, i would do it aswell...up to you :)
 
thats not bad...I would use about 1/2 of that though...the length of the line the same, just use 1/2 the ammount and stretch it..

I did it like im describing to you, and i get 22core0 23core1 idle and 41core0/1 at load with coretemp software. using a tuniq tower,

but its pretty common to have one core more than the other...i would still use less thermal paste though....although if its all installed, unless you want to overclock it more, and get the lowest possible temps, i wouldnt take it all apart just to change the paste...maybe if i was in there doing somthing else, i would do it aswell...up to you :)


Thanks, least i know my line is the correct way, there was alot more paste on previously and my temps did go down so I'm just gonna leave it. Either i dont know what settings to do, or my config is a complete garbage OCer. Ive only managed (2.56ghz 320x8) @ 1.375 V
 
So a tiny drop like the old days isn't the "right" way to do it anymore?! Ugh wish i knew before i put mine together! :mad:
 
i dont believe this whole "too much thermal paste actually hinders performance" I think that theories full of poop. Seems to me that the big mac scandle where the macs had too much thermalpaste: the thermalpaste wasnt the issue... they just didnt apply enough pressure.

Anyways, back on track. It is certainly possible to get one side of the core much hotter than the other. If you were able to get a diode right on the silica, i think you would be suprised at the temperature difference.
 
Well. It kind of makes sense. As the paste is suppose to fill in the air/gaps. So if you put on too much then its filling in the air/gaps and then some. Kind of like wearing a glove when touching something hot versus not wearing one at all! ;)
 
So a tiny drop like the old days isn't the "right" way to do it anymore?! Ugh wish i knew before i put mine together! :mad:

i think that always was the right way, even according to artic silver.

the difference is that now you dont have a single core in the middle, you have 2 (more) offset to the edges of the heatspreader, which id imagine it why they're suggesting using a line now.

2 dabs directly over each core would increase the chances of getting an air bubble in the middle, so it makes sense.

edit: okay that was a bit off, the cores are actually alot closer together than what i was picturing. but it looks like the line will cover the cores (which are slightly offset from the center) as well as the cache:

http://www.xbitlabs.com/images/cpu/core2duo-e6300/c2d-illustration.jpg
 
it will also cover the convex/concave heatspreader on the chip

The dot is not the way to do it on dual core, because of the 2 chips as the other guy explained.

And there is definatly such thing has too much thermal paste...i have seated many heatsinks working in the techshop here comparing..and I always find "less is more"..even Artic Silver says that in the instructions.

just think about what it is doing, it us filling the extremly small gaps between the heatsink and the proccessor. microscopicly speaking, there are ALOT of gaps...if there is no themal paste it wont transfer to the heatsink execpt for the very few spots it would be touching, If you have too much, it will fill it up, but then the heat will act like a blanket and keep the heat on the chip...if you have a thin thin layer.. it flows right across, heat is energy

But if you put on the ammount you did, and get fine temps (I suspect you would), then you did a fine job, if you want it a little better, try a little less, thats all im saying [h]ard members take every degree into consideration...its part of the hobby
 
So, on the C2D's, do you just put that horizontal/vertical line of paste and put the heatsink on? Do you not have to distribute the paste all across the entire processor?
 
That is quite a bit of paste, I usually use about 1/4 of that but spread it out on both the chip and the heatsink using a saran wrapped finger.

There is little harm in using too much, as it will simply squeeze out the sides.
 
Well. It kind of makes sense. As the paste is suppose to fill in the air/gaps. So if you put on too much then its filling in the air/gaps and then some. Kind of like wearing a glove when touching something hot versus not wearing one at all! ;)

no... no that doesn't make sense, since the extra material would just extrude between the wafer/mobo/w.e.
 
Well. It kind of makes sense. As the paste is suppose to fill in the air/gaps. So if you put on too much then its filling in the air/gaps and then some. Kind of like wearing a glove when touching something hot versus not wearing one at all! ;)

But remember, when you wear a glove and touch something hot, you don't get burned because the heat cant get thru......which is exactly why too much paste is bad...you WANT the heat to get thru...otherwise your heatsink cant do its job, as all the heat is staying in the CPU....
 
I've tried putting a pea size down, making a line horizontal/vertical and my tried and true method, lightly covering the entire IHS. The most difference I have ever seen is 2c which is well in the margin of error.

People make this too damn scientific, when I have never seen any difference worth talking about.

So that puts MrWizard in the same *logical* camp as myself.
 
Uhmm, is that a Zalman HSF? If so i think you are supposed to rotate that bracket 90 degrees clockwise. At least that is what the instructions say.

I dont think it is a big deal, because from what i can tell, all it does is provide a little space for when you lift the CPU lever.

I just noticed because I am currently re-installing a HSF with the Zalman 9700 NT, and that brackets looks familiar ;-)
 
i dont believe this whole "too much thermal paste actually hinders performance" I think that theories full of poop. Seems to me that the big mac scandle where the macs had too much thermalpaste: the thermalpaste wasnt the issue... they just didnt apply enough pressure.

There's about 350 years of thermodynamics that disagree with you. The farther the distance between the hot and cold junction the higher the temperature difference required to move a given amount of energy between them. This isn't a theory, it's a proven scientific fact.

Assuming AS5 has a 0.0045°C-in2/Watt in a 0.001 inch layer, a CPU trying to dissipate 100W would have a 0.45C temperature rise due to a 0.001 in layer, but would have a rise of 1.8C with 0.004 inches of AS5 between them. The more thermal compound you goop on there above what's needed to displace the air, the higher your temperatures will be.
 
There's about 350 years of thermodynamics that disagree with you. The farther the distance between the hot and cold junction the higher the temperature difference required to move a given amount of energy between them. This isn't a theory, it's a proven scientific fact.

Assuming AS5 has a 0.0045°C-in2/Watt in a 0.001 inch layer, a CPU trying to dissipate 100W would have a 0.45C temperature rise due to a 0.001 in layer, but would have a rise of 1.8C with 0.004 inches of AS5 between them. The more thermal compound you goop on there above what's needed to displace the air, the higher your temperatures will be.

Ok that all makes perfect sense.. but doesnt the heatsink pushing down just spread it out evenly..spewing excess over the edge..if u have that much thermal paste in there?

Side note: funny how my asking am i doing it rite? turns into a debate! :O
 
Ok that all makes perfect sense.. but doesnt the heatsink pushing down just spread it out evenly..spewing excess over the edge..if u have that much thermal paste in there?

Side note: funny how my asking am i doing it rite? turns into a debate! :O

to an extent, but heatsink mounts dont apply arbitrarilly high amounts of pressure, the more thermal compound, the larger the gap between will be. variations from .001in to .004 in seems reasonable
 
People here overdo this kind of thing to death. Stop worrying about it. Some of these folks had concave/convex heatspreaders to start with and they are the same 'only use a grain of rice sized application' people.

Don't goop it up, and you'll be fine. A couple degrees difference, even using a lot is meaningless.

That dust would irritate me more than anything to do with the thermal paste.

Ask yourself "does this make a difference in the quality of my life?" - It's an easy question that alleviates a lot of minor decisions, and fretting behavior.
 
This is stupid. My processor is hotter than other people's for some reason. I just redid the thermal paste and reseated the heatsink today. First time I did what I normally do, a lot of thermal paste. Then I read this thread. I tried it with a dab and smeared it thinly with saran wrap around my finger. 5 celsius degrees higher now.
 
This is stupid. My processor is hotter than other people's for some reason. I just redid the thermal paste and reseated the heatsink today. First time I did what I normally do, a lot of thermal paste. Then I read this thread. I tried it with a dab and smeared it thinly with saran wrap around my finger. 5 celsius degrees higher now.
... if that is AS5, from reviews (one posting) of a CPU at newegg, it takes 10 days to cure before fully effective. (unless some other reason is at fault for the temperature)
 
Back
Top