How Thermal Compound Spreads

hardtarget · 17 Sep 2009 at 20:20

I was bored.

While looking around deciding what parts I need for a new build I came across this very interesting video

Sorry if it's been posted already.

krooton · 17 Sep 2009 at 21:22

Win for the track in the MX-2 video

DavyBoy · 17 Sep 2009 at 22:39

Hmmm, so it would seem that the spreading method that most people suggest on here isn't the best method at all, as seen in the MX-2 video and the one above, it causes air bubbles which can't be good at all...

I used to always just do the line method, but after loads of people on here recommended spreading, I started doing it that way instead.

Gonna pop my TRUE off tommorow and go back to the line method again.

PermaBanned · 17 Sep 2009 at 22:55

Was a thread on this a while back. I just use the pea method, what this doesn't mention is that there's no performance difference between the spreading methods unless you put a stupid amount on. Also, it doesn't account for the heat that it's actually convecting, which will help with spreading also.

Merlin5 · 17 Sep 2009 at 22:57

It's funny you should post that. I was literally about to post this

http://www.youtube.com/watch?v=ffK7L0Qj13Q

JonJ678 · 17 Sep 2009 at 23:59

There may be an argument to be made for running it for a while, so heatsink and processor are hot, then swiftly reapplying. Lower viscosity when hot, spreads better.

Might even go so far as to suggest heating up the heatsink, applying paste to heatsink, then mounting. I haven't tried this though.

Edit: more intelligent approach would be heating up the paste itself thinking about it

DavyBoy · 18 Sep 2009 at 00:01

Jon which method do you use for applying TIM?

JonJ678 · 18 Sep 2009 at 00:02

The majority of the time I leave whatever was on before in place. Doubt I'm alone

Otherwise, clean with citrus degreaser then ipa, then dot in the middle. Mount, remove and look at how much has been covered. Then put more on and remount if need be. No doubt this leaves air bubbles, out of the less lazy methods I'd personally vote for a line over the cores or a cross.

Currently running a mix of mx2 and as5. Both silicone based, so no particular issues I can think of.

DavyBoy · 18 Sep 2009 at 00:07

I always clean my chip and HSF with Akasa TIM cleaner stuff before applying more.

I used to just use the line or dot in the center method, but several people on here convinced me that spreading it out was the better method, so I went back to doing it like that.

Though I am using MX3 and it has a really weird consitency, it's quite thick, so I am not sure how well it would spread by simply mounting a cooler directly onto the blob, I will test it out tomorrow to see how well it spreads.

JonJ678 · 18 Sep 2009 at 00:09

Try it hot too

I'm probably buying some mx3, seems too loved to pass on.

DavyBoy · 18 Sep 2009 at 00:13

Yeah, I guess if this MX3 was warmed up it would spread a lot easier.

I personally love it mate, I got a 5c drop in load temps going from AS5 to this.

Though, it probably wasn't just the change in TIM, it could have also had something to do with me reseating the HSF again that gave me a drop in temps to, 5c does seem like quite a difference from just changing my TIM..

But there have been others on here who experienced a drop in temps when swtiching to MX3.

game4it · 18 Sep 2009 at 00:21

DavyBoy said:
I always clean my chip and HSF with Akasa TIM cleaner stuff before applying more.

I used to just use the line or dot in the center method, but several people on here convinced me that spreading it out was the better method, so I went back to doing it like that.

Though I am using MX3 and it has a really weird consitency, it's quite thick, so I am not sure how well it would spread by simply mounting a cooler directly onto the blob, I will test it out tomorrow to see how well it spreads.

JonJ678 said:
Try it hot too

I'm probably buying some mx3, seems too loved to pass on.

I have tried spreading mx-3 after heating it up first.It made no difference as as the IHS was cold :mad:

I tried the cling film method too but once you use any pressure to spread it it just comes off.With gentle pressure it will spread after a while but i couldn't get it thin enough again without it rubbing off.
This is with my lapped cpu.

I will try heating up the heatsink first and try see if it speads on that tomorrow.

edit:
@ Davyboy My temps are the same as the noctua nh1 i was using and i have reseated lots of times already

Madskillz · 18 Sep 2009 at 14:14

Some good ideas there guys, nice one.

What's the consensus on how to get an even spread of TIM? I use the sticky label with a square cut-out and credit card method. Seems to work for me.

Madz

Gaidin109 · 18 Sep 2009 at 14:23

JonJ678 said:
The majority of the time I leave whatever was on before in place. Doubt I'm alone

Otherwise, clean with citrus degreaser then ipa, then dot in the middle. Mount, remove and look at how much has been covered. Then put more on and remount if need be. No doubt this leaves air bubbles, out of the less lazy methods I'd personally vote for a line over the cores or a cross.

Currently running a mix of mx2 and as5. Both silicone based, so no particular issues I can think of.

AS5 isn't silicone based, its Silver based. It also uses zinc oxide and aluminum oxide and boron nitride. This makes it capacitive so you have to be careful not to get it on cpu pins and the like. MX2 is not as its completely metal free. I would like to know what the ingredients are but they are not telling

I spread my MX2 with a credit card. seems to work.

tuppers360 · 18 Sep 2009 at 14:34

Cool I am going to be assembling my lot soon that certainly helped!!

JonJ678 · 18 Sep 2009 at 17:37

I disagree Gaidin. Both are solid particles suspended in silicone grease. MX3 is carbon, as5 has silver involved. I don't doubt there are other solids suspended as well, people occasionally try diamond. Since the solids are suspended in silicone, I consider them silicone based

Since AS5 is particles in silicone, and so is MX2, a mix of them is just a lot of different particles in silicone grease. It probably performs roughly as the weighted average of the two. Silicone grease by itself does a reasonable job.

The air bubbles from the even spread are interesting, but I'm not convinced they'd persist when repeatedly heated and cooled under pressure

Gaidin109 · 18 Sep 2009 at 18:09

JonJ678 said:
I disagree Gaidin. Both are solid particles suspended in silicone grease. MX3 is carbon, as5 has silver involved. I don't doubt there are other solids suspended as well, people occasionally try diamond. Since the solids are suspended in silicone, I consider them silicone based

Since AS5 is particles in silicone, and so is MX2, a mix of them is just a lot of different particles in silicone grease. It probably performs roughly as the weighted average of the two. Silicone grease by itself does a reasonable job.

The air bubbles from the even spread are interesting, but I'm not convinced they'd persist when repeatedly heated and cooled under pressure

AS5 uses a zinc/proprietary oil base in which three different sizes of silver particles are suspended. MX2 is ceramic/carbon particle based. You should not mix the two as this can impair the performance of both. Neither is silicone based.

http://www.arcticsilver.com/as5.htm

lbcliff · 18 Sep 2009 at 21:15

Has anyone tested temp difference for different methods?

madindehead · 18 Sep 2009 at 21:49

Hmm. What do you think will be the best way to apply thermal paste to my soon to come i7 and Megahalems then?

PermaBanned · 18 Sep 2009 at 21:50

I like the pea method, but it's your choice really.