Best way to apply thermal paste

  • Thread starter Deleted member 82943
  • Start date
D

Deleted member 82943

Guest
What's the best method in applying thermal paste to get the best coverage and the lowest temps provided I'm using Artic silver compound and the rosewill Z3 cooler?
 
I put a grain of rice sized dab in the middle, and let the HSF spread it.

Lots of people use a credit card/razor blade to get it into a thin coat, but this is against what AS advises...

I've done both, and neither produced a better temperature.
 
i heard a thin layer is the best, my friend tried both, and found a degree or 2 cooler with the layer, i have heard from many people not to put a blob in the middle, because you either dont put enough or put too much
 
It depends on the quality of the heatsinks surface.

If its maximally flat/smooth, the grain of rice method is perfect.

If the heatsink has a non smooth or grooved surface, a grain of rice may not be enough.
You can either use more paste than a grain of rice or ...
prefill the grooves with paste (use a credit card edge to remove any excess) and then use the grain of rice method.
 
Dethman said:
i heard a thin layer is the best, my friend tried both, and found a degree or 2 cooler with the layer, i have heard from many people not to put a blob in the middle, because you either dont put enough or put too much
Click to expand...

This method can trap air.
 
Chernobyl1 said:
This method can trap air.
Click to expand...

Umm...I don't really see how. If the blob is in the center, applying any pressure on the HS will only push the paste outwards along paths of least resistance (largest pockets of holes). Unless the person puts more than one blob, the grain of rice in the center should work fine.

Actually, applying with a razor is only suggestable for smooth surfaces on both HS and cpu IHS. For example, if your HS is crooked (slanted), I wouldn't suggest using the razor method since when you put it on, you still have the slant and zero contact on one side of the HS. Of course in these cases, you should really just sand the HS down.
 
Zoogle said:
Umm...I don't really see how. If the blob is in the center, applying any pressure on the HS will only push the paste outwards along paths of least resistance (largest pockets of holes). Unless the person puts more than one blob, the grain of rice in the center should work fine
Click to expand...

I was talking about this
Dethman said:
i heard a thin layer is the best, my friend tried both, and found a degree or 2 cooler with the layer, i have heard from many people not to put a blob in the middle, because you either dont put enough or put too much
Click to expand...

He doesnt use a blob so not sure why you thought I was talking about that?
 
I dunno, my bad...must be tired but I thought you were talking about the blob method. Haha...cheers....
 
The blob method is what a lot of big oems show in the internal pics for repair techs. Personaly that seems to work fine on chips without a ihs. For chips with a ihs I personaly like to apply a layer on it.
 
swatbat said:
The blob method is what a lot of big oems show in the internal pics for repair techs. Personaly that seems to work fine on chips without a ihs. For chips with a ihs I personaly like to apply a layer on it.
Click to expand...

That is exactly opposite to what you should do. *With* an IHS you should be using the grain of rice method. Without it you should be evenly spreading it across the chip.

The exact application somewhat depends upon what processor you are using. But the principle on letting the heatsink spread out the compound is the same. Pre spreading is bad mmmkay?
 
Chernobyl1 said:
This method can trap air.
Click to expand...


So what if it does? It won't make a difference.... you aren't going to be able to overclock any better because you took a degree or two off the temps. If that would make a difference then you are already WAY past the stability threshold and need to look to other means to cool/overclock your system. People put too much thought into this. Squirt a bit in the middle of the CPU and use the pressure of attaching the HSF to distribute the paste.
 
I dab on some AS and then cut a business card so that I have a piece about 1/2" wide to use to smooth is over the entire surface so there is just enough to cover the top of the cpu. I then add the HSF and as I'm attaching it there is usually enough horizontal movement, i.e. 1/4" each way, while doing this that I feel comfortable that the HSF is making 100% contact and mated well.

I see the point of the rice blob method, but I can't quite reconcile it that I get 100% coverage. I think that smothing it out over the entire cpu in a paper thin layer is more effective. I think the worst method would be to apply a circle or a spiral of compund and then just seat the HSF then because you could trap air.
 
i used AS ceranmique for my conroe e6600. what i did was put a little dab on the heatsink and spread it evenly around. then i wipe it off just enought so that there is still a little bit of the thermal paste on it. i do the same for the CPU. when you are doen you are not suppose to see or notice and evidence of thermal paste on either surface. then i put a little BB sized dropped in the middle of the HS and lay it gentally on the cpu.

this might seem like it;s too much but you'll get pretty good results.

what they say about letting the HS spread the paste is right. it might be too much or too little.
with a layer it has to be "PERFECTLY" flat or you'll get air bubbles and they are the devil.
 
Poncho said:
So what if it does? It won't make a difference.... you aren't going to be able to overclock any better because you took a degree or two off the temps. If that would make a difference then you are already WAY past the stability threshold and need to look to other means to cool/overclock your system. People put too much thought into this. Squirt a bit in the middle of the CPU and use the pressure of attaching the HSF to distribute the paste.
Click to expand...

Your argument made me look at the screen in disbelief.
I'm not sure you are in a position to dictate how stable anyones system is and certainly should not condone trapping air between the heatsink and CPU die!

You are very wrong stating: if a couple of degrees C can make a difference then we are already way past stability.
What kind of advice is that?

After making sure I have a GOOD contact with my CPU cooler, I overclock until its unstable and then back off a little for stability.
If I have air trapped between my CPU die and the heatsink, my overclock will be lower or less stable so I'll have to decline your wise words.
 
Chernobyl1 said:
Your argument made me look at the screen in disbelief.
I'm not sure you are in a position to dictate how stable anyones system is and certainly should not condone trapping air between the heatsink and CPU die!

You are very wrong stating: if a couple of degrees C can make a difference then we are already way past stability.
What kind of advice is that?

After making sure I have a GOOD contact with my CPU cooler, I overclock until its unstable and then back off a little for stability.
If I have air trapped between my CPU die and the heatsink, my overclock will be lower or less stable so I'll have to decline your wise words.
Click to expand...


Well let me ask you something... if a couple degrees matters, then how do you deal with changes in ambient temps? Do you keep you PC in a closed room with it's own cooling STRICTLY monitoring the temps? :rolleyes: Ambient temps can very in a house greatly and if you overclock needs those couple degrees that you got by being anal about your thermal paste... what do you do on a hot day when it may be 5-10degrees hotter in your house? Do you just not run that system? LOL... you guys crack me up.
 
Use a latex glove or a plastic card to spread the compound over the heat spreader, works great all the time, just dont use too much paste :D

oldmx
 
Bare Core CPU - spread a THIN layer evenly across the die
CPU with IHS - small "blob" of thermal compound in the middle of the IHS, letting the pressure of the heatsink spread it out (I never know how give an example of this.. the "grain of rice" is too small, at least the grains I eat. Maybe two grains.)
intel's "dual core" CPU with IHS - small/thin line of paste down the middle, parallel to the cores. Heatsink spreads it.

This minimizes the chances of air pockets forming. I've always found this to provide superior contact... you can tell the difference after taking the heatsink off (if done correctly).
If you spread it out over an IHS, there's a higher chance of air pockets forming, and it's easier to over-use.

For CPU's with an IHS, the most heat is transferred directly above the die, and not really on the edges of the IHS. So it's best to have the greatest contact in the middle of the IHS than to have less contact over the entire surface of the IHS.

more = less


edit: I should also add, that sometimes it helps to put a small amount on the heatsink (and/or IHS) and rub it into the heatsink using your finger inside a plastic bag. Then wipe it off. It helps fill the small gaps. (just gotta try not to let anything else get onto the heatsink while wiping off the extra grease)
 
Poncho said:
Well let me ask you something... if a couple degrees matters, then how do you deal with changes in ambient temps? Do you keep you PC in a closed room with it's own cooling STRICTLY monitoring the temps? :rolleyes: Ambient temps can very in a house greatly and if you overclock needs those couple degrees that you got by being anal about your thermal paste... what do you do on a hot day when it may be 5-10degrees hotter in your house? Do you just not run that system? LOL... you guys crack me up.
Click to expand...


No matter when you start calculating the difference a couple of degrees will make, it can easily reduce stability at some time in the day when clocked near the limit.

You gotta be joking making a statement like that.
My system is stable to cope with local weather variations but there are occasions when it does get too warm and I clock down a little. Thats in the nature of overclocking.
This is why we clean our fans of dust, or do you suggest we dont bother with that as well?

If I were to add 2C to my temps 24hrs a day, I would need to reduce my overclock to maintain overall stability.

2C is a low ballpark figure, allowing air to be trapped between the die and the heatsink is negligent and could make for a very bad contact with much higher temps!
 
Zoogle said:
Umm...I don't really see how. If the blob is in the center, applying any pressure on the HS will only push the paste outwards along paths of least resistance (largest pockets of holes). Unless the person puts more than one blob, the grain of rice in the center should work fine.

Actually, applying with a razor is only suggestable for smooth surfaces on both HS and cpu IHS. For example, if your HS is crooked (slanted), I wouldn't suggest using the razor method since when you put it on, you still have the slant and zero contact on one side of the HS. Of course in these cases, you should really just sand the HS down.
Click to expand...

so ya, you're saying always apply with a razor (if not then YOU MUST LAP!)

....I agree. Sorry for the Caps.
 
I think that brings up an important point. If you are prepping a system for high performance, there is a good deal of pre-prep that must be performed prior to applying TIM. Lapping the heatsink is a big step in making sure you have good heat transfer. Make sure to clean the surfaces completely after you finish the lapping. The technique used afterwards varies from person to person but several of the ones offered here all work well. Which one is best? Probably the one that you can successfully apply to your system. The weakness is not understanding fully what it is you are trying to get done. Applying and buffing in TIM prior to making the application is my favorite method, and i usually work the CPU first, then apply more TIM and use a razor blade to get it spread thinly and evenly, and then scrape the excess onto the heatsink and buff it into the surface. When I am finished I usually have a light grey sheen on the heatsink, and a thin translucent layer on the top of the chip. This method has certainly worked well for me.

Oh, and don't forget that when you attatch the heatsink, you want to get it set gently and evenly on the top of the chip, and if it is attatched with screws, take turns tightening them down. You want to sneak up on the finished tightness.
 
I think someone said this earlier..

But I use a latex glove to spread it thin across the entire CPU surface. Put 4 small drop in each corner of the CPU and start spreading well until the entire thing is covered. The HSF contact to it should be good. Spread it thin...

I tried this over the weekend and it dropped like 2c - 3c from the original method I did was to just a drop in the middle and had the headsink spread it upon pressure.

Use good alcohol to clean the surface of the headsink first before applying.
 
wow , getting scientific with it,

i put a drop of the compound on the cpu,
spread it with my finger
done
 
Just take a piece of saran wrap and put it on your finger. First time I tried it was easy as pie. It is not hard to get perfect looking thin coat on the cpu. Small bead in center and gently smooth around.
 
OR!

Use the manufacturer's specification for the application of the paste, apply a little dot in the middle and get no worse temps than all the rest of the methods :D
 
You must log in or register to reply here.
Back
Top