Best thermal paste right know?

sauron18

n00b
Joined
Nov 2, 2012
Messages
10
I'm building a pc again... So i need to know which is the best thermal paste or the top 5 at least. Help me out
 
Here you go:

1) Tim Tronics Grey Ice 4200
2) MX-4
3) Arctic Silver 5
4) Noctua NT-NH 1
5) Prolimatech PK-1
 
1. Liquid Pro.

It's a pain to clean off but it is the best that I have used. Works great on bare die low pressure mounts. Cannot be mixed with aluminum. Only on high quality copper blocks and surfaces.
 
been using AS ceramique 2 and haven't had no problems. imo its doing on par or better than AS5 was when i was using it
 
Guys, OP is asking for the best thermal paste. Unless you've done testing yourself with a dozen pastes and have recorded the results, it would be nice for you to link reviews or other data so that the OP can get an objective answer to his question.

http://skinneelabs.com/2011-thermal-paste-review-comparison/

Indigo Xtreme is the best paste but it costs about $10 per application and it's very difficult to install correctly. When you first install it, you have to run your heatsink without a fan so that it heats up a lot, that way the metal alloy turns into a liquid and flows. I've read reviews where people have tried to install it and it didn't flow like the instructions say, so they wasted their money.

Instead I recommend Prolimatech PK-1 or Phobya Hegrease. They perform within half a degree of Indigo Xtreme and they are both very affordable (in comparison to Indigo). Also, Prolimatech has released two new versions called PK-2 and PK-3. The technical specifications say that these two pastes both have better performance than PK-1. I've sent Prolimatech an email asking them which paste is best, and they told me that PK-3 is the best, followed by PK-2 and PK-1.

So to sum it up you can get either Phobya Hegrease or Prolimatech PK-1 or PK-2 or PK-3.
 
Guys, OP is asking for the best thermal paste. Unless you've done testing yourself with a dozen pastes and have recorded the results, it would be nice for you to link reviews or other data so that the OP can get an objective answer to his question.

We are giving our opinions based on our experiences and what we have read just as you are. I have used at least 15 types over 12-14 years.

EDIT: For the performance, affordability, ease of use and availability of MX-4, it wins in my opinion.
 
Last edited:
What I'm trying to say is that anecdotal evidence is not very useful if someone is asking for the best item. Scientific data is much more useful.
 
While I tend to agree, if OP wanted nothing but scientific results he/she could have just googled/searched for reviews on thermal paste.
 
While I tend to agree, if OP wanted nothing but scientific results he/she could have just googled/searched for reviews on thermal paste.
Amen. Google exists for a reason. We are assuming the best of the OP in that he wants the opinions of [H]'ers and not that he's too lazy to google.

Indigo Xtreme is the best, and it's the easiest to apply of the "liquid metal" TIM's. However, that doesn't mean it's easy and it's still very expensive at $20/application.

I currently use MX4. I have NT-H1 as well, which I suppose I could use just as well. PK-1 is also highly recommended but I have no experience with it.
 
I'll give another +1 on the IX, it's pretty binary as to whether it's applied or not and if it's applied, it's a good application, lapping doesn't help it significantly and no messing with trying to get it perfect.
 
+1 for MX-4. No cure time, nonconductive, noncapacitive, inexpensive, and available at most computer shops or popular e-tailers.

Also, +1 for Google if you want scientific results, but the common trait amongst many enthusiasts who have used the products is that the application is far more important than the product. Many threads available where people used to use the 'spread' method and saw major improvements by switching to 'drop' method.
 
Been using AS5 for years. One tube will last you nearly a lifetime.

I've tried some of the "what's hot" thermal compounds a few times, only to find out that I was a guinea pig.

I'm with evilsofa on that most thermal paste are more or less the same shit. I feel the same way with the application procedure as well. I still spread it across with a credit card regardless if the "correct" way may be the bb or other method.
 
IMO, since these are all opinions, use nothing but Indigo Xtreme (on my i7 930@4GHz/ H100 right now) if you want the absolute best and like spending money on the best.

If that's not the case, use mayonnaise (any of the popular TIM's).

Skinneelabs_TIMOverallTemp.png


http://www.indigo-xtreme.com/
 
CoolLaboratory Liquid Pro on my delidded 3570k.

Ambient Temp 17.6C
[email protected] 1.14V
http://valid.canardpc.com/show_oc.php?id=2578618

quickndirtymaxdelidmod.jpg



Minumum temps
16C 19C 16C 11C

Maximum temps
43C 47C 46C 40C


I tried PK-1 and IC Diamond on this same chip and PK-1 was about 1C warmer than IC Diamond but the Liquid Pro with a nice low pressure mount on my die is over 15C cooler than IC Diamond.
 
SonDa5 makes a really good point in that if you have low pressure mount than nothing can match liquid metal pastes like Liquid Ultra or Indigo Xtreme. If you're using a de-lidded CPU or a push-pin mounting system then it might be worth it to invest in one of those pastes.

OTOH if you have a regular heatsink with good mounting pressure you have very little to gain. So it all depends on your situation, do whatever works for you.
 
I wish people would call stop calling Indigo Xtreme "paste".

Arctic Cooling's MX line is good all-around paste, but people's testing of it seems very inconsistent. Skinnee wrote that he believes that this is due to inconsistency across the different batches they put out. I'd tend to agree since I've tried two different tubes of MX-3 on my current processor; the older tube kept in-line with original testing when it was first released, but the new tube could never match the performance (~0.8C off on average) over multiple reapplications... I'm not new to testing so yes, all variables were kept tight.
 
IMO, since these are all opinions, use nothing but Indigo Xtreme (on my i7 930@4GHz/ H100 right now) if you want the absolute best and like spending money on the best.

If that's not the case, use mayonnaise (any of the popular TIM's).

Skinneelabs_TIMOverallTemp.png


http://www.indigo-xtreme.com/

People should use Indigo Xtreme between the cpu & integrated heat spreader on the Intel Ivy bridge chips that Intel didn't solder the heat spreader on. I likey my temps with Phobya He Grease & a Noctua NH-D14.
 
I wish people would call stop calling Indigo Xtreme "paste".

I don't think most people care whether their TIM is a paste, phase change metal, or pad so long as it performs well, "paste" is just a generalized term for TIM.
 
People should use Indigo Xtreme between the cpu & integrated heat spreader on the Intel Ivy bridge chips that Intel didn't solder the heat spreader on. I likey my temps with Phobya He Grease & a Noctua NH-D14.

If you do a good job cleaning up the epoxy and lap the base of the IHS a bit like GTX480 owners were doing you'll end up with very good temps on Ivy with just some MX-2.

I would rather spend some time getting a good mount than using something like indigo xtreme.

MX-2 is dirt cheap and works great. With a good mount it'll work as well as any other paste.
 
SonDa5 makes a really good point in that if you have low pressure mount than nothing can match liquid metal pastes like Liquid Ultra or Indigo Xtreme. If you're using a de-lidded CPU or a push-pin mounting system then it might be worth it to invest in one of those pastes.

OTOH if you have a regular heatsink with good mounting pressure you have very little to gain. So it all depends on your situation, do whatever works for you.

This isn't always true..I am also running a de-lidded cpu, except I am running an i7 3700K, with IC diamond..I applied a small rice grain sized amount to the die, then heated it for roughly 5 minutes with a 1850W hair dryer to make it soft like my MX-2..I then spread a THIN layer across the entire die, as IC recommended to me via email..

I told SonDa5 that he used WAY to much IC Diamond in the picture he posted, and he used the traditional method of letting the block squish the TIM around..This method, coupled with his "low pressure" mount as he stated, I feel did not do the IC Diamond justice what so ever..

Finally, using IBT with just a standard mode (1GB of installed ram) is far from providing the amount of heat that a 24 hour run of Prime95 with AVX, using a "custom" run of max memory, with a ten minute interval between FFTs..You see what I mean when you see that my third core spiked @ 89C, despite having an average range of 75-78C..

I like SonDa5 much as a fellow member, and have no personal axe to grind, but tests done by very well respected members going from IC Diamond to the IX and CL Ultra metal TIM did not create anywhere near a 19C drop, maybe 4-5C in the extreme cases..

4.8Ghz, 1.288Vcore 24 hour prime stable..
Max load temps were: 69,82,89,77C..Please note that during this run, my ambient temp was 23C, which is nearly 3C higher then normal..

I also had to run one of my GT-15s @ 7V, instead of the normal 12V, due to some horrible high pitched noise coming from it..Replacement fan from NCIX just arrived yesterday, so I will have it in tomorrow.

P95_12VFan_PLL165V_V1288.png


While I feel that the metal based TIM(s) are the "best", when you compare their prices to IC Diamond, I just don't see the benefit from a cost/performance ratio..
.
 
I like SonDa5 much as a fellow member, and have no personal axe to grind, but tests done by very well respected members going from IC Diamond to the IX and CL Ultra metal TIM did not create anywhere near a 19C drop, maybe 4-5C in the extreme cases..

I've seen a few people report that with their delidded ivy bridge cpus too. Its mostly that the CL ultra still performs well with a bad mount. I saw that much of a difference when I did a better job cleaning up the epoxy and lapping the base of the IHS a bit. So that was just fron correcting the mount.

Also IC diamond sucks. Its thick, its a pain to clean up, and its super sensitive to bad mounts. Most of the time mx-2 will actually perform better, at least in my experience.
 
What are you intending to cool? under what conditions?
For most builds, MX-4, Noctua NT-NH 1, and Prolimatech PK1, perform about equally well,
I usually go by price/availability and ease of application.

If you are going to extreme temperatures, overclocking on inadequate coolers or benchmarking for record and are fighting for every degree that is different

Arctic Silver 5, is this still conductive? I do not see an advantage to a "long curing" conductive paste.

Unfamiliar with Tim Tronics Grey Ice 4200.

My feeling for a new build is spending a little more on Cooler means more than using the absolute highest performing paste. At $10 or $20 an application a Silver Arrow or H100 seem better options.

Prolimatech PK1 can be a nuisance to find/order ship. Or was.
 
I've seen a few people report that with their delidded ivy bridge cpus too. Its mostly that the CL ultra still performs well with a bad mount. I saw that much of a difference when I did a better job cleaning up the epoxy and lapping the base of the IHS a bit. So that was just fron correcting the mount.

Also IC diamond sucks. Its thick, its a pain to clean up, and its super sensitive to bad mounts. Most of the time mx-2 will actually perform better, at least in my experience.

Funny, you mention that you prefer MX-2..When I started this build I had a ton of MX-2, but it was getting nearly a year old, and while always kept away from light and extreme heat/cold, I was a bit leery..The reason MX-2 works so well is because it has a very high carbon content, just like Diamonds..(see what I did there?) ;)

I was turned onto the IC Diamond because during a massive months long testing project some of us did @ Anandtech, ICD was the only compound not suffering from "thermal pump out" that many users were reporting (keep in mind we are talking about 1 specific application, in this case, de-lidded Ivy with the TIM replaced and the IHS reinstalled)...Also keep in mind that I installed the IC D. a specific way, as recommended by their support team when advised of my intentions of a direct die mount, and after heating it, it spread like smooth butter..

After these reports, with some of the users' having the issues with MX-4, I decided to get the IC Diamond since I planned on running de-lidded..I thought about the Metal TIMs, but since I was not 100% sure everything would be perfect the first attempt (it was:eek: ), I did not want to run into the issue of possibly having to lap the actual die of the cpu..

I feel SonDa5 is trustworthy, I just wanted more then one opinion, and one that offered cold hard facts like he does when he posts with data backing it up...As soon as I get a new 320mm rad, I am ditching the current 320mm slim + 120mm and this will be a Folding@Home box 24/7, except for my gaming every now and then..It has been too long and I am itching to get back to Folding..

*Ninja Edit:

Bababooey, I wanted to clarify a few things..In my first post you quoted, I meant to refer to IX only, not CL Ultra or Pro..I have heard and seen hard numbers with the CL, and it is good stuff..Again though, they recommend lapping both the HS/Waterblock and the IHS after you remove it each time, and I don't have one!!...Again, while the die surface is pretty tough, and I am quite experienced with lapping, I would honestly say that I would be moderately scared to attempt it on my 3770K..

Now that being said, that does not mean that I DO NOT want to DO IT, lol..I have thought about it several times, and its just that the cpu is is so light and fragile without the IHS installed..I am afraid I might crack the PCB in half..Based on my testing, and being Prime95 stable @ a pretty low voltage for IVY @ 4.8Ghz, I and others have said that they feel 5Ghz daily is within reach...If you guys really think I can knock another 10C+ off the 5+C I should definitely get from going to a much better 320mm rad and also reducing a massive amount of restriction in my loop @ the same time, then I think I am 99% in!!:p
 
Last edited:
.

I feel SonDa5 is trustworthy, I just wanted more then one opinion, and one that offered cold hard facts like he does when he posts with data backing it up...

I forgot to mention that with my latest mount of CL Liquid Pro I tweaked the mount with a .5mm layer of Fujipoly extreme around the die. I think it may be helping cool of PCB as well. My temps are really good with the last mount with the fujipoly thermal pad.


attachment.php


Had this idea over a month ago.


http://www.xtremesystems.org/forums...Idea-help....)&p=5151814&posted=1#post5151814
 
That's a cool idea. I'm playing devil's advocate here: how do you know there is heat coming off of the carrier package that the CPU is mounted on? Because if there is no heat coming off of the carrier package nothing will happen (or you will send heat from the IHS into the carrier package instead of the other way around if the IHS is hotter).

EDIT but if you are getting better temps than before then that make's it more believable too...
 
SonDa5, is your cooling setup allowing you to overclock the CPU higher than a regular heatsink?
 
SonDa5, is your cooling setup allowing you to overclock the CPU higher than a regular heatsink?

Stock heat sink was way hotter than my water loop. If I remember right with Intel Burn Test and stock heat sink at 4.5GHZ my 3570k was reaching TJ MAX temps of 105C on 1 core and others were about 100C.




I haven't been trying to find max over clock since doing my delid cooling mods but I have tested at 5GHZ and it is stable and cool now while before it was not stable and hot.

So yes my cooling mods have worked out great.

Before it took more voltage and I was not stable. My Intel Burn Test temps were exceeding TJ MAX at 5GHZ.

Huge improvement now.


Ambient Temps 18.6C
IBT quick and dirty standard run

[email protected]

http://valid.canardpc.com/show_oc.php?id=2578924

Victory5GHZIBTSTANDARD.jpg


Idle Temps

21C 24C 21C 18C


Maximum Temps

62C 70C 67C 62C
 
Back
Top