two silly questions about applying thermal compound

[toyota]

I will be installing the A70 this weekend and have never fooled with applying compound before. my previous coolers all had compound pre applied on them.

1: why does nearly every review show them applying compound to the cooler when instructions say to put a pea size drop on the cpu?

2: after applying thermal compound and placing the A70 on the cpu, is it okay to lift up the cooler and check to see if compound spread evenly?

 

[drizek]

1. Don't know. Pea sized in the center is the way to go, I think.

2. NO

 

[bononos]

I will be installing the A70 this weekend and have never fooled with applying compound before. my previous coolers all had compound pre applied on them.

1: why does nearly every review show them applying compound to the cooler when instructions say to put a pea size drop on the cpu?

2: after applying thermal compound and placing the A70 on the cpu, is it okay to lift up the cooler and check to see if compound spread evenly?

1. Peasized drops is the amount that people used in the old days when heatsinks were not finely milled unlike the very shiny mirror like finishes you get today. Slathering on a toothpaste sized amount is the wrong thing to do unless the manual says otherwise. A rice grain amount is more commonly used now plus 'hazing' both mating surfaces with a tiny amount of compound to fill in the microscopic valleys and carefully scraping over the tiny amount with a plastic credit card or razor blade.

2. I sometimes do that and smile afterwards when it looks right.

http://forums.anandtech.com/showthread.php?t=2203188&page=2

 

[lehtv]

2. NO

Yes.


(c) Idontcare

 

[dma0991]

1. Makes almost negligible difference as what matters is that the TIM is sandwiched between the two surface. Would putting ham on the top slice make my sandwich taste better compared to putting my ham on the bottom slice? :hmm:
2. Not too concerned about this. Just dab a little TIM and let viscosity does its magic.

 

[Shmee]

what compound are you using? different compounds can be applied differently.

 

[GLeeM]

2. Lifting up and then setting back down can cause air bubbles. If you want to lift it up to see what it looks like, then you should clean everything off and reapply. Also, it will look different after a couple days of heat on it.

 

[Revolution 11]

The application of thermal paste is not as important as the amount you apply. In general, the thinner the layer of paste, the better. Still, I think you should follow the thermal paste manufacturer's instructions on how to apply it. They would have the best advice on how to use their products.

No, you can't lift up the cooler after the paste has been applied. You might as well clean off the surfaces and reapply the paste. If you are concerned about how the paste is spreading, you can use a flat glass surface and test the cooler and paste on that to see the spread and coverage.

 

[lehtv]

Why are people saying you can't lift the cooler to check how the thermal paste you just applied was applied, then put it back? I already posted evidence of only 0.7 degree celsius difference at load with 2600K OC'd to 5Ghz.

 

[PlasmaBomb]

Why are people saying you can't lift the cooler to check how the thermal paste you just applied was applied, then put it back? I already posted evidence of only 0.7 degree celsius difference at load with 2600K OC'd to 5Ghz.

To be fair that was IDC on his lapped processor and heatsink. The detla would likely be large for an unlapped CPU and heatsink...

 

[toyota]

oh well its done. I have one core with a crazy low temp at idle thats 11C cooler than the rest. that core was always a few degrees cooler even with stock heatsink too though. under load the temps for all cores are within 3-4 degrees of each other.

even with the A70 fans at 1600 rpm instead of 2000 rpm, my highest temp using maximum setting in Intel burn test was 53 C. lol, with the stock cooler I would stop the test right before it hit 80 C. and that's with stock clocks.

also at 1600 rpm the fans are quieter than the stock Intel fan.

 

[Idontcare]

oh well its done. I have one core with a crazy low temp at idle thats 11C cooler than the rest. that core was always a few degrees cooler even with stock heatsink too though. under load the temps for all cores are within 3-4 degrees of each other.

The die is not symmetric so you would expect the cores to have differing tmeperatures based on their proximity to so-called "dark" silicon (areas of the die which are inactive, like the iGPU).

 

[lenjack]

I failed to see the silly parts of your questions.

 

[tranceport]

The purpose of the compound is to transfer heat between the two metals. Air is not a good conductor of heat so the compound is used. If the two pieces of metals are flat enough (lapped) then less compound is needed. You want the thinnest layer possible while still maintaining complete coverage without having the compound run out and all over your motherboard.

A decent site with some pictures...

http://www.techpowerup.com/printarticle.php?id=134

 

[john3850]

When ever I buy a cpu or a high end viedo card I always relace the tim.
How can you have that many posts from mostly cpu and gpu related subjets and never used TIM.
I will always check the flatness first then I lightly buff all related parts on a polisher with white rudge and then use the least amount tim and then shim the springs for more tension.
There is a difference in your core temps cause for the first time you have fair cooling.

 

[sm625]

2. Lifting up and then setting back down can cause air bubbles.

That's what she.... wait wrong thread. Do you have any evidence to support this statement?

 

[soccerballtux]

what a silly question.

 

[soccerballtux]

Yes.


(c) Idontcare

this is interesting.