Applying TIM to CPU

nheather

nheather

nheather

Associate
Joined
19 May 2004
Posts
947
Location
Horsham, West Sussex
Hi,

I have always applied a small amount of TIM to the CPU and spread it VERY thinly and then applied the heatsink, pushing down hard and twisting to bed it in.

With heatsinks which have copper pipes that touch the CPU I have also applied to the heatsink and then scraped it off leaving it only in the joints between the pipes and the heatsink body.

But I frequently see advice that I use a 'rice grain' sized blob in the centre of the CPU and let the pressure of the heatsink spread it out.

I understand the principle but unsure about the possible spread

1 - surely the spread is going to be circular whereas the CPU is square - does it really spread out to cover the entire face of the CPU?

2 - A modern AMD CPU is around 14mm square - is a rice grain really sufficient to cover the whole surface? With the method I have used in the past I put more than that on, mind you a fair bit is removed or left on the spreader so I'm not entirely sure how much I put on.

Cheers,

Nigel
 
Most modern cpus the die is a fraction of the size of the lid you see and its the die your cooling not the lid the reason not to spread in many peoples opinion is due to getting air bubbles in the tim and air as we know doesnt conduct heat too well. Go for whatever you feel more comfortable with, if the correct amount is used I doubt theres more than a degree or two difference between all methods
 
It depends on the TIM. Check the application instructions because they will differ between the materials and brands.

Generally a rice/pea size works fine. However, i tend to find it better to put a rice size in the middle and in each corner i put an additional tiny dot.

As for the direct heatpipe method you do, it is advised to do as you do and apply it to the gaps in addition to the CPU.
 
Generally a rice/pea size works fine.
Click to expand...

I've seen this quoted quite a lot to.

There is a huge difference in volume between a grain of rice and a per (even a petit pois) - reckon a pea could easy be 3x the volume of a grain of rice.

Looked at the video - actually have seen it before. All of those are using much more than a grain of rice.

Also the bubbles on the spreading example seems to fixed to support the posters argument. Pressure is applied and all is fine, no bubbles. It then looks like the acrylic plate is flexed at the corners to cause it to lift across the centre giving the pattern.

Also understand what you are saying about not needing to cover the whole area but the examples with Arctic Silver in the video look insufficient to me.

Cheers,

Nigel
 
Last edited:
OP, as you mention, the blob in the middle technique is for solid based heatinks, ones with heat pipes don't spread too well as you either need to pre-fill the voids on the heatsink or... what I do with heat pipe coolers is... put a thin line on each heat pipe, this tends to fill the voids as well as ensuring that each heat pipe is fully covered, which is the most important bit to my mind
 
Thermal paste on the whole of the spreader.

Thinking back to my heat transfer stuff at university many years ago ... my question is this.

The core is small, granted, but it is attached to the heat spreader. The prurpose of the spreader is to transfer the heat from the core to the outside world.

On it's own it isn't enough so you need to attach a heatsink.

Surely then if the core is transferring heat to the spreader you want the whole of the spreader to be in full contact with the heatsink for optimal heat transfer.

Cheers,

Nigel
 
nheather said:
Thermal paste on the whole of the spreader.

Thinking back to my heat transfer stuff at university many years ago ... my question is this.

The core is small, granted, but it is attached to the heat spreader. The prurpose of the spreader is to transfer the heat from the core to the outside world.

On it's own it isn't enough so you need to attach a heatsink.

Surely then if the core is transferring heat to the spreader you want the whole of the spreader to be in full contact with the heatsink for optimal heat transfer.

Cheers,

Nigel
Click to expand...

Exactly.
 
nheather said:
Thermal paste on the whole of the spreader.

Thinking back to my heat transfer stuff at university many years ago ... my question is this.

The core is small, granted, but it is attached to the heat spreader. The prurpose of the spreader is to transfer the heat from the core to the outside world.

On it's own it isn't enough so you need to attach a heatsink.

Surely then if the core is transferring heat to the spreader you want the whole of the spreader to be in full contact with the heatsink for optimal heat transfer.

Cheers,

Nigel
Click to expand...

Indeed. But spreading by hand creates air pockets.
 
nheather said:
Thermal paste on the whole of the spreader.

Thinking back to my heat transfer stuff at university many years ago ... my question is this.

The core is small, granted, but it is attached to the heat spreader. The prurpose of the spreader is to transfer the heat from the core to the outside world.

On it's own it isn't enough so you need to attach a heatsink.

Surely then if the core is transferring heat to the spreader you want the whole of the spreader to be in full contact with the heatsink for optimal heat transfer.

Cheers,
--

Spot on, more contact area the better
Nigel
Click to expand...
 
You must log in or register to reply here.
Back
Top Bottom