Thermal conductive adhesive (only 1.7 w/m*k)?

Just got a silicone thermal adhesive, ACC AS1803. It is only rated at 1.7 w/m*k and is classified as a thermal conductive paste.

With silver and copper being rated at 300-400+ w/m*k, is this thermal paste really good for conducting heat?

I used it to glue a few coins together to be used as a heatsink (saves me the trouble of filing the coins as the silicone is fairly thick and tends to 'even' out the surface when the coins are pressed together). But with its thermal rating, is it in fact impeding the conductivity as opposed to two filed coins connected directly?

Do I really have to get Arctic Silver Adhesive to obtain silver's conductive properties in an adhesive paste form? *edit : Ok found that they claim arctic silver is around 7.5 w/m*k. Still sounds low to me. Is it anything to do with its mass?

Also, water is only rated at 0.6 w/m*k. Why is it so popular as a heatsinking solution?

Why not solder the coins together?

A soldering iron is not very effective to do this - use a torch (flame type - not flashlight type ). I have done this a few times with very good effect. Make sure to do this on a heat/flame resistant surface. If these are used to mount an emitter, you should at least file/sand the emitter side of the heatsink smooth.

Paul

That is exactly why you want the layer of thermal goo to be as thin as possible.

You want the two surfaces to be as flat as possible, to minimize the thickness of the thermal paste layer.

People that try pushing their PCs as fast as possible (overclockers) have been known to lap both surfaces carefully of the CPU and heatsink, in order to remove the high and low spots, so that the layer is even thinner, which makes the thermal transfer even better. This lowers the CPU temperature.

Unfortunately, some folks don't understand that the common white paste has a rather low thermal conductivity (high thermal resistance), and will slap the stuff on like some folks do with cream cheese and bagels. Ack!

Once the surfaces are flat, which is what thermal paste is designed for, you apply it in a layer that is thin enough to see thru- just a little bit, and mate the two surfaces and apply pressure.

I remember one designer actually tried to use white thermal paste as a gap filler to fill a gap 1/4" wide. He had many field failures, and cursed the thermal paste to no end, once he found out that his power parts were not being kept cool.

A few minor modifications to the assembly resulted in achieving mating surfaces. He did not want to use thermal paste, since he considered it an insulator now. A few thermocouples put on his parts showed he still wasn't holding the parts as low as he thought he was, but it was much better than having 1/4" of thermal paste. Showing him that adding a little paste to his surfaces that now mated, then repeating the test brought a very big smile to his face. Guess what? His field failures ceased.

Hi Chimo, I did solder my first bunch of coins together (filed surfaces). I found filing a little tedious so I though I could even the 'bumps' with a layer of thermal paste and avoid filing.

Then I started looking up the specs and realised the conductivity was poor as hell. I guess my 2nd makeshift heatsink has to be dumped. And to think I was planning to paste a lux emitter on a heatsink with this gunk. haha.

Newbie, nice writeup. I guess I realize now that the thermal paste itself is a poor conductor when compared to the metal itself, but it's supposed to fill up microscopic uneven surfaces (thus a thin layer will do) to make contact better. The paste is still better than air in terms of conductivity and 'plugging' these microscopic holes will help with the heat transfer. Is that right?

LEDcandle said:

Newbie, nice writeup. I guess I realize now that the thermal paste itself is a poor conductor when compared to the metal itself, but it's supposed to fill up microscopic uneven surfaces (thus a thin layer will do) to make contact better. The paste is still better than air in terms of conductivity and 'plugging' these microscopic holes will help with the heat transfer. Is that right?

Click to expand...

You got it now.

LEDcandle said:

Hi Chimo, I did solder my first bunch of coins together (filed surfaces). I found filing a little tedious so I though I could even the 'bumps' with a layer of thermal paste and avoid filing.

Click to expand...

Hi LEDcandle, you should not have to file the coin surfaces that will be soldered. You can use a hammer to pound most of the ridges flat. If you tin the surfaces that will be mated, the solder will fill the voids. Ensure the coins are clean and use a little flux before tinning.

As I alluded to above, and as Newbie has pointed out, the surface that will accept the device you want to cool (likely an emitter in this case) should be smooth. Thermal paste or adhesive should be as thin as possible, that's why you want the emitter contact area of your heatsink as smooth as possible.

Good luck in your mod,

Paul

chimo said:

Hi LEDcandle, you should not have to file the coin surfaces that will be soldered. You can use a hammer to pound most of the ridges flat. If you tin the surfaces that will be mated, the solder will fill the voids. Ensure the coins are clean and use a little flux before tinning.

As I alluded to above, and as Newbie has pointed out, the surface that will accept the device you want to cool (likely an emitter in this case) should be smooth. Thermal paste or adhesive should be as thin as possible, that's why you want the emitter contact area of your heatsink as smooth as possible.

Good luck in your mod,

Paul

Click to expand...

Hi Paul,

Your mod looks nice and well-done. I have no torch type tho, I don't think an iron will work for this. I ordered some dremel bits and mebe I can make short work of the coin surface with a sanding stone of some sort (hopefully). Thanks!

Understand that sometimes the 'tedious' is the best way to get what you are after. From personal experience, I've never had the best results with half-assed fabrications. Good luck with the heat sink and post pics of build!

The thermal resistance of a layer is d/k, where

k = thermal conductivity of material
d = thickness of layer.

If d is very small, the resistance of the thermal grease layer could easily be smaller than several mm of copper. It will definitely be smaller than the thermal resistance of an air gap between two imperfectly mated pieces of metal.

Update:

Here is a list of thermal conductivities for different materials. Notice that air is ~100 times worse than thermal grease.

Don has told me to my face that I am "anal." I 1200 min.-grit lap the slugs of my LSs and their mating thermal surfaces when possible. When I Arctic Alumina Adhesive them down, I press as hard as I can! I have yet to fry an LS. (Leave convertor boards outta this. :green: )

Larry

tvodrd said:

Don has told me to my face that I am "anal." I 1200 min.-grit lap the slugs of my LSs and their mating thermal surfaces when possible. When I Arctic Alumina Adhesive them down, I press as hard as I can! I have yet to fry an LS. (Leave convertor boards outta this. :green: )

Larry

Click to expand...

I'm in the same boat. I'll lap my heatsinks until I don't have any fingerprints left and weigh down the objects being adhered until I'm damn sure they are closely secured as possible.