New(ish) Thermal Compound?

[orclev]

Ok, so it's not really new, but it's newer than a lot of others. What does everyone think about Coollaboratory LIQUID PRO (http://www.performance-pcs.com/catalog/index.php?main_page=product_info&cPath=58&products_id=4373)? From looking at test results on various sites it looks like this stuff out performs AS5 by a pretty decent margin. The downside seems to be that it doesn't play well with aluminum (but who on here would be using aluminum heatsinks anyway?), and according to some that it will discolor your CPU as well, although once again, who cares.

So, my question is, has anyone on here actually used this stuff, and if so would you recommend it, or advise the rest of us to steer clear of it, and why?

[p2501]

This stuff is by no means new. I've used it for over a year on my old E6600 in combination with an Thermalright SI-128. It did well, but i didn't compare it to other compounds, as I had only good ol' AS3 lying around. Even if you cleaned all surfaces really well with alcohol it's a pain to spread, I've used a spare brush and once it got "grip" on the surface it was okay. But all in all I guess I was at least 15 minutes at it to spread it on the IHS and SI-128. Some will tell you that the stuff tends to glue components together, so that you'll be having hell getting the heatsink off again but that never happened to me. I think it's pretty good stuff IF you're willing to spend a bit of time applying it. AND if you're willing to invest even more time getting it off the components it was used on - forget alcohol or solvent - you'll have to _polish_ it off, maybe with stainless steel wool, but I used sandpaper.

[orclev]

Ok, well, I'm looking at using a D-TEK FuZion v2 CPU block, and I'm looking at the best thermal compound I can get. I don't really care how hard it is to spread as I'm already putting a lot of work into this system, and an extra 10 to 30 minutes is nothing. The only issue I could see would be having it stick the water block to the CPU, and having to grind it off, but since I was considering lapping the CPU anyway I'm not overly concerned about that. As long as this stuff out performs the AS5 I was planning on using, and doesn't have any other drawbacks, I'm inclined to use it. Thanks for the feedback.

[iboomalot]

**quietly chants**

OOOOOO CCCCCCCCCCCCCC ZZZZZZZZZZZZZZZZZZ

[Zaskar]

Why are people still worrying about how hard something is to spread? Thought it was mostly accepted nowadays that a pea sized dot in the center of the CPU or a line going across the cores (thats the official AS5 spreading procedure as per their site) was the best, letting the heatsink do the spreading out for you, unless you have a low tension mounting system this should be best.

[Martinm210]

Yes I've used it for a while now, although the TC Consultants Grease 0098 is very similar in performance and what I'm using now for testing (Needs easy application).

http://www.xtremesystems.org/forums/showthread.php?t=159752

[oerekum]

Not with this stuff. I remember that it had some special way to be applied.

[Jimmer411]

I hated this liquid metal crap. Spreading was a PITA for me when I tried it. After a good 6 or so months it was a PITA to remove my WB at the time.

[p2501]

Why are people still worrying about how hard something is to spread? Thought it was mostly accepted nowadays that a pea sized dot in the center of the CPU or a line going across the cores (thats the official AS5 spreading procedure as per their site) was the best, letting the heatsink do the spreading out for you, unless you have a low tension mounting system this should be best.

You obviously never used this compound my friend. If you had used it before you would know that this simply doesn't spread, it's like mercury. If you just put on the block or HSF you're chances are good that it just rolls off in the process which is what you absolutely don't want, because it's very conductive. So, other than most pastes you'll have to spread it.

Orclev, if you're searching for something that will outperform AS5 then this is what you want. But MX-2 or the stuff that Martin recommended are very good too and have little too no cure time, unlike AS5. It's your choice. :up:

[Serpentarius]

If you wan the best .. this is the best .... (and cheap - DIY)


Material.........Thermal conductivity W/(m·K)
---------------------------------------
Diamond........900 - 2320
Silver............429
Copper..........401
Gold.............318
Aluminium......237


DIY - BEST THERMAL COMPOUND (http://hackedgadgets.com/2007/12/05/diamond-thermal-grease/)


http://hackedgadgets.com/wp-content/2/_Diamond%20Thermal%20Grease.jpg

http://www.overclockers.com/articles1389/conductivity2.JPG

[Martinm210]

Yeah CLP is a pain to install and it's a little scary taking the block off when it's cured up (sort of feels fused on there), but it does work well. Before trying the Tim Consultants grease, I was more than willing to put up with the CLP installation and removal. I actually found that using a Q-Tip worked pretty well, but it's just not practical for testing.

Anyhow it was a worthwhile gain over AS5, but I never tried any of the other good ones people really like these days. The TC grease seems to perform nearly as well and it cures very fast so it's perfect for thermal testing where you need to mount multiple times.

[shachar2]

$15 for a spit of Compound.
How much does diamond dust costs? you'll have to mix it with glue for it to stick to the cpu/gpu

[p2501]

you'll have to mix it with glue for it to stick to the cpu/gpu

Not really, the DIY guide said something about a type of silicone paste. Don't have a link though. But I'd think for me the whole process would be just a bit too much hassle for 1 or 2 degrees better temps. :shrug:

[Nickel020]

It's very easy to install if you use a fine brush to spread it.

It's definitely better than any other compound around, but's it's kind of hard to remove once it's been on for a while. If you re-install your cooler every month or so, it's no problem, other wise it will get hard and then it's hard to remove. If you don't care much about the resale value of your CPU, get it, it's definitely worth it (2-3C temp difference to convetionla TIM, that's more than you'll get by moving from a Swiftech Triple to a Thermochill radiator!).

[orclev]

If you wan the best .. this is the best .... (and cheap - DIY)


Material.........Thermal conductivity W/(m·K)
---------------------------------------
Diamond........900 - 2320
Silver............429
Copper..........401
Gold.............318
Aluminium......237


DIY - BEST THERMAL COMPOUND (http://hackedgadgets.com/2007/12/05/diamond-thermal-grease/)


Embedded in the comments for that link was a link to http://www.heatsinkfactory.com/ic-diamond-7-carat-thermal-compound-15-gram-p-16605.html
which appears to be the compound shown in that performance graph. Interesting concept, although if this stuff is so awesome I wonder why it's relatively unheard of. Then again, the CLP also seems to be relatively unheard of as well. I do notice that the installation instructions given on that page are contradictory. The first thing it tells you to do is give time for the solvent used in it to evaporate before installing the heatsink, but the very next instruction is to put a dab on the CPU and then immediately install the heatsink.

Edit: It also occurs to me that because the diamond compound is inherently non-conductive it could be used on non-IHS chips unlike CLP such a the NB, and SB, which traditionally would have required the use of AS Ceramique.

Further Edit: Also available in larger quantities, but otherwise appears to be exactly the same stuff http://www.heatsinkfactory.com/ic-diamond-24-carat-thermal-compound.html

[Zaskar]

If you wan the best .. this is the best .... (and cheap - DIY)


Material.........Thermal conductivity W/(m·K)
---------------------------------------
Diamond........900 - 2320
Silver............429
Copper..........401
Gold.............318
Aluminium......237


Great performance, but after a Very short time all the liquid in the compound drys out and your left with a hard powder, great if you never want to move it, but dont expect to lets say change the tension of your bowed CPU block after the first day to try to change the temp load over the cores.

Anything that drys like that always makes me think that a slight bump could cause it to lose its sealed connection, and since its now a solid, could be an issue.

Its non conductive though which is great for non CPU applications like ram chips, but so is MX2.

[orclev]

FYI, little more digging on that diamond TIM turned up a forum thread on ocforums.com that apparently is frequented (or was) by one of the guys who invented the stuff: http://www.ocforums.com/showthread.php?t=517055&page=1
Check page 4 for some posts by the guy (JoeC). I'm still reading through that thread, but a number of people so far have reported favorable results using the stuff, at least compared to AS5, not sure how it compares to MX-2 though. So far the impression I'm getting is that if applied correctly, and allowed to evaporate properly it should give results as good as, or better than AS5, but that it's thicker and therefore more difficult to spread than AS5, although that seems to be par for the course if you want something better than AS5 (never used MX-2 so I don't know how that compares in terms of spreadability).

Edit: FYI, this stuff is carried by Petra's: http://www.shoppts.com/icdi7cathco.html and here's the manufacturers website: http://www.innovationcooling.com/

[Zaskar]

I think that was an older revision of the compound, the stuff they sell now doesn't recommend any evaporate time.

That compound is one of the types you do now want to spread around, they recommend the pea sized dot in the center. your just fooling yourself when you try to spread thick compounds like that. Someone did a test on 2 big sheets of glass pressed together with different spread methods, the dot in the center was the best, trying to spread it out still ended up with areas with differing amounts of compound after heated in an oven.

[orclev]

I think that was an older revision of the compound, the stuff they sell now doesn't recommend any evaporate time.

That compound is one of the types you do now want to spread around, they recommend the pea sized dot in the center. your just fooling yourself when you try to spread thick compounds like that. Someone did a test on 2 big sheets of glass pressed together with different spread methods, the dot in the center was the best, trying to spread it out still ended up with areas with differing amounts of compound after heated in an oven.
Ah, they must have decided to remove the solvent. JoeC mentioned that they added the solvent to make it more spreadable, and that once the solvent evaporates it performs better but is closer to putty in consistency. They must have decided that spreadability didn't matter, and to just use the apply and squish method for an even spread. I'm sort of leaning more towards this stuff than the CLP at this point, although I wonder how it compares to MX-2. Man, we really need someone with a lab setup to get a sample of a bunch of the top TIMs out there and do a real shootout between them all.

[orclev]

Just found a thread on here with results of reviewing IC D7: http://www.xtremesystems.org/forums/showthread.php?t=182778&page=10

Heh, who knew. I'm so glad I found this forum.:D

[Zaskar]

Man, we really need someone with a lab setup to get a sample of a bunch of the top TIMs out there and do a real shootout between them all.

Its just so hard to do. the top Tims are so close that any difference could Easily simply be the mount.

Would need like 10-15 mounts and avg the temps to get any sort of useful data, and that's Allot of work.

[orclev]

Its just so hard to do. the top Tims are so close that any difference could Easily simply be the mount.

Would need like 10-15 mounts and avg the temps to get any sort of useful data, and that's Allot of work.
That thread on here that I posted a link to seems to have a pretty good testing setup. Later on in the thread they started using some sort of pressure film to measure the contact areas and pressure applied between heatsink and CPU. It looks like with the D7 it's critical to have a large contact area with high pressure, even more so than it normally is. So, seems like if you're going to be using this stuff you need to have a lapped CPU and heatsink, and a good quality backplate, or else you're better off with another TIM. FYI, seems the IC D7... uh, guy... rep... inventor... has some sort of connection to the company anyway, goes by the name of Tastymannatees on this forum, and if you agree to do testing might be able to hook you up with a sample of D7 to test. Not sure on the requirements to be a tester for them though, or if they're willing to take on some more testers.

[Zaskar]

I thought that sticky was outdated though? If I remember right, it tested MX-1, things have changed a little bit.

[orclev]

I thought that sticky was outdated though? If I remember right, it tested MX-1, things have changed a little bit.
I didn't notice if it was a sticky, I found it linked from somewhere else (and I can't remember where at this point, probably another forum). The thread is doing testing on D7, not in comparison to any other specific TIM, but more of a general performance benchmark. Part of the testing being documented includes what compound was used previously, what it looked like after the heatsink removal, how long the TIM had been in place, etc. etc. and the later posts were from a couple weeks ago, so even though the first few posts were a bit old (even then not much, few months), the latest posts are quite new.

Edit: Hmm, just went back and looked at the thread and the last posts were from dec 2007, not sure what I was looking at with newer posts. To many tabs open, maybe I was looking at a different thread.

Edit 2: Going to post a message to that thread asking what the final results where and if any comparisons to MX-2 or CLP where done.

[orclev]

It was compared to MX-2 and both of the tests it stomped MX-2 by a few degrees (which when you're talking high high end TIM like this is pretty significant). I think I'm going to use this stuff on my entire system when I switch over to my new waterblocks, so CPU/GPU and all major mobo components. Along with the new radiators and dual WC loops I'm really looking forward to what kind of temps and OC I can achieve with this.