Best current TIM?

S

silent-circuit

[H]F Junkie
Joined
Sep 18, 2005
Messages
16,136
I don't even know what's on my chip right now (used what was laying around and forgot), and I might be getting a new chip soon.... Is Shin-Etsu X23 still the best? (as of 2, 3, 4 years ago? Whenever I last looked?) Doing my own research given I'm not seeing any super-recent threads on the topic, so I'll update when I know more, but would love to see other opinions.
 
Thermal Grizzly Kryonaut is the best rated "regular" (i.e. non-crazy, special metal super hard to apply you may blow it and need to order it again to re-apply) type paste on the market as of now.

But its pricey and hard to find anywhere locally. Both Amazon and NewEgg carry it though... (~$12-14 a small tube)
 
Phanteks, Coolermaster Master 5, NTH1, Indigo Extreme, all are very good to excellent never heard about that Kryonaut stuff, no review site I ever frequent ever listed it that I had noticed, generally speaking, very crap to ok thermal paste, seems ~10c difference at most average, ok to great thermal paste difference is generally quite small ~5c most cases, usually a good application is around the same at most, so I would imagine considering cost, why pay some 12-14$ per tube when example one can get some of that Master 5 stuff that was just released, get within 3c of same and save a fair amount while still not being that off the best top performing stuff out there (Col Pro paste) :)
 
When I researched the topic recently I found that Gelid's GC-Extreme was consistently rated at the top of the bunch of 'normal' thermal compounds. 'Normal' meaning excluding fancy liquid metal or diamond compounds...
 
Liquid metal is something I'd never recommend unless you plan on reapplying it every 6 months - the Liquid Metal Ultra stuff dries out after ~8 months. Happened on both my old 3930K and 3920XM. Dry as a bone, and temps were up.
 
So many TIMs are within 1-2C of each other, it actually matters more how you apply the TIM than which TIM you use. I still have some AS5 from over 10 years ago (have yet to run out) and it still does a good enough job for me.
 
Its nice to see some updated thermal paste reviews. It does seem that Thermal Grizzly Kryonaut is pretty much the best "conventional" paste around, but Gelid GC Extreme and Prolimatech PK-3 are close to the top as well. Liquid metal pastes/pads in theory have the very best cooling of all, but are conductive, have tricky applications, and I've heard reports they "wear out" and need reapplication with relative frequency, so I've never found them worth the aggravation. In most comparisons, the 5 or so best performing conventional pastes are typically within 1 or 2 degrees of each other, so don't stress too much about not being able to find one brand or the other. Oh and as a side note - I find it rather refreshing that Thermal Grizzly also seems to make some thermal pads that (hopefully) don't suck, which is a relatively underserved market; the only pads I've heard of frequently used are OEM garbage with crappy performance. If these work well, then using the pads for say...GPU memory or other difficult areas to cover with paste, seems like a nice idea.

While we're talking about TIM, lets talk about modern application regarding (non-delidded) modern processors. I noticed that the Thermal Grizzly Kryonaut instructions mentions spreading a thin layer of paste over the surface of the processor heatspreader. Kryonaut even comes with a special application flat tip which screws into the syringe, to make it easier to spread - a nice touch. Years ago, the "place a dot in the center the size of a pea, lower your heatsink onto it" method was king. When quad cores became popular, there was also the "thin line across the middle of the heat spreader, horizontally" option as well. Do people still use the "dot or line, + pressure" method, or do they spread the paste manually? Any differences between using a heatsink/air cooler, versus as water block for a CPU? What do you find works with today's Haswell (-E) and Skylake procs?
 
Xaeos said:
Its nice to see some updated thermal paste reviews. It does seem that Thermal Grizzly Kryonaut is pretty much the best "conventional" paste around, but Gelid GC Extreme and Prolimatech PK-3 are close to the top as well. Liquid metal pastes/pads in theory have the very best cooling of all, but are conductive, have tricky applications, and I've heard reports they "wear out" and need reapplication with relative frequency, so I've never found them worth the aggravation. In most comparisons, the 5 or so best performing conventional pastes are typically within 1 or 2 degrees of each other, so don't stress too much about not being able to find one brand or the other. Oh and as a side note - I find it rather refreshing that Thermal Grizzly also seems to make some thermal pads that (hopefully) don't suck, which is a relatively underserved market; the only pads I've heard of frequently used are OEM garbage with crappy performance. If these work well, then using the pads for say...GPU memory or other difficult areas to cover with paste, seems like a nice idea.

While we're talking about TIM, lets talk about modern application regarding (non-delidded) modern processors. I noticed that the Thermal Grizzly Kryonaut instructions mentions spreading a thin layer of paste over the surface of the processor heatspreader. Kryonaut even comes with a special application flat tip which screws into the syringe, to make it easier to spread - a nice touch. Years ago, the "place a dot in the center the size of a pea, lower your heatsink onto it" method was king. When quad cores became popular, there was also the "thin line across the middle of the heat spreader, horizontally" option as well. Do people still use the "dot or line, + pressure" method, or do they spread the paste manually? Any differences between using a heatsink/air cooler, versus as water block for a CPU? What do you find works with today's Haswell (-E) and Skylake procs?
Click to expand...

I just tried the Kryonaut TIM this week. The instructions on the back of the retail packaging has two application methods. One is spreading it across the CPU yourself. The Second method is a dot in the middle with a little more TIM in an X pattern across the heat spreader.

I ended up trying to do the spread method but this stuff feels like wet sanding a car when you spread it. Most of the TIM stuck to the spreading tool so I ended up with just a slight "haze" of TIM across the entire heatspreader. You can still read every letter on it. Temps seem about 2 degrees lower than with MX-4.

Also, have you ever tried the Fujipoly TIM pads? They were very effective when I gave them a shot for VRM cooling.
 
Last edited:
silent-circuit said:
Considering I'm looking at buying a de-lidded chip, liquid metal is out of the question.
Click to expand...

Still more than possible, but getting it on anything surrounding the die would always be a risk when applying. They are some of the best however, I am talking by 10C, that is on a prepped part though. With a normal IHS and HS base, I found the "metal" TIMs to only be as good or sometimes worse, depending on the finish quality of the base, I found them to be just slightly better than normal TIMs on waterblocks, but they are often much nicer finish than HS bases. The 10C drop I have seen was repeated a few times on lapped CPU/HS/WB combos, high heat conductivity and its natural thin nature helped a great deal. Dies are very flat and a lapped HS/WB with the metal TIM would probably work amazing, though I am not sure worth the risk unless this is just a benching machine only. I remember seeing a few sites a while back with all the IHS/TIM hate with Intel going on and some sites delidded and put quality TIM under them and they all saw large drops, when they used coollaboratory, it dropped another 6-8C (depending on site) over the next best TIM used under the IHS.

Down sides, it does dry out, however I never noticed any rise in temp with this, one rig ran non stopped for 4 years, no issues. It almost glues everything together though, and needs to be heated to remove easy. It will also without a doubt "etch" into the HS and IHS, you WILL have to sand both to remove all residue, in my cases it did not matter as I was lapping anyway.
 
many people here talk without any truly knowledge, OP you are already using the best TIM in the market, second only to Liquid Metal Stuff....

What model of Shin Etsu X23 are you using? G751 vary from the X7783D and X7762 depending on usage..

X7783D and X7762 are better suited for direct-die cooling and between CPU core and IHS, also for GPUs, G751 work better for typical usage. the only options to replace shin etsu I would use ever are Prolimatech PK3, Gelid GC extreme and Arctic Ceramique 2 as only options.
G765 for
 
http://www.overclockersclub.com/reviews/noctua_nt_h1/5.htm

http://www.play3r.net/reviews/cooling/best-thermal-paste-2015-update/

Just two I picked out of checking out some slightly older and newer reviews.

IMHO I'd avoid the liquid metal stuff because it does tend to be rather permanent, relatively speaking, on top of being conductive, needing to be reapplied often and not safe for some metals.

I'd go with stuff that measured use time in years, not months.

Unfortunately, what used to be among the best, Ceramique 2, barely passes for better than generic these days.

Of all the stuff mentioned in this thread, across several recent reviews...

Gelic GC Extreme, Thermal Grizzly Kyronaut, IC Diamond, Phanteks TH-NDC all seem to rate extremely well with Noctua NT-H1 rounding off the top.
 
I still use MX-2. Cooler I just ordered comes with MX-4.

Honesty, any of the better brands is only going to be a degree or two difference between them. Pick one that works for you and use it.
 
I use MX-4 but i've heard good things about thermal grizzy kryonaut .
 
Finished up my tube of MX-4 on a friend's build recently, going to have to try some of this Thermal Grizzly next.
 
Using the Noctua NTH1. Was having issues with my MX4. Temps are about 3C cooler between MX4 and NTH1.
 
I'm a really big fan of Zalman STG-1 because of how easy it is to apply.
 
First time i've heard of this Grizzly TIM. How easily does it spread? I'm used to MX4 and I love how easy it is to use.
 
I used the noctua nt h1 for a while before switching to the gelid gc. I will go back to the noctua when this tube runs out. Application is a little more difficult as it is a looser TIM. And temps were within a few degrees which is within the range of other variables being at play.
 
All of the recent TIM's are going to come within a degree or two of each other. And that's not accounting for user application error or environment variables. Pick one, use it, move on with life. ;)
 
MX2 seems to last a long time as well. Mine was applied for about 2 years until last week when I installed a new AIO. It was still soft and easy to clean up. I went with MX4 this time. Very easy to apply.
 
I've always been a fan of the Gelid GC-Extreme. Top-end performance, easy to use, and not too expensive. One $10 tube has lasted me probably a dozen applications, and I've probably got a few dozen more apps left in it.
 
Some of you who say that you found that one TIM is 2 or 3 C cooler than the one you used before may have been subject to the margin of user error.

Apply the same TIM to the same CPU and cooler the same way 5 or 10 times in a row, do temp testing each time, and you'll find a variance. Some installs are good, some are bad. You'll need to do a lot of testing to get a good average on each TIM, more than most of us have time for.
 
evilsofa said:
Some of you who say that you found that one TIM is 2 or 3 C cooler than the one you used before may have been subject to the margin of user error.

Apply the same TIM to the same CPU and cooler the same way 5 or 10 times in a row, do temp testing each time, and you'll find a variance. Some installs are good, some are bad. You'll need to do a lot of testing to get a good average on each TIM, more than most of us have time for.
Click to expand...

Yep, sometimes just replacing old TIM with new will lower temps. I think my next buy will be something other than Arctic Silver 5 though, just for shits.
 
AceGoober said:
I personally prefer Arctic Cooling Ceramique 2. It applies easily and doesn't wear out like others. I normally change thermal compound once a year. With ACC2 I change it every 2 years.
Click to expand...

I'm going to second this one. If you don't plan to re-apply every 6-12 months I would go with Céramique 2 . The 2 is SO MUCH easier to work with then the first release that was basically spreading a pencil eraser . But it will last forever.

On the short term. I've played with IC-Diamond, MX7, AS5 and honestly so long as it's a high quality TIM and it's correctly applied to a C L E A N surface (I can't stress that enough, drives me insane) you're going to get about the same withing a few % of a degree.

Also, the tube of AS5 you picked up with your build 6 years ago. Toss it out and get new.
 
The Ceramique is my constant standby for the very reasons you mentioned. It's the one tube of TIM I keep in my repair kit for house calls. I've tried several TIMs over the years and some were better than others, but the Ceramique is the only TIM I ever use on my GPUs and it's been like that since I first discovered it. It's easy to clean, it's performance is good even if you get into really low temps, and it's just so easy to spread.
 
I use both Tuniq TX-4 and Noctua NTH1 both non conductive and easier to work with than all the ones I've tried. The temps are really good too, I've used them on gpu waterblock vrm's and on cpu.
 
I'm using Gelid Extreme on both rigs below (comes with the EK Supremacy EVO cpu cooler) and impressed with results. I bought a separate tube of it from Newegg.
 
Riccochet said:
All of the recent TIM's are going to come within a degree or two of each other. And that's not accounting for user application error or environment variables. Pick one, use it, move on with life. ;)
Click to expand...

Hoping for a good amount of time before the EOL there is a question which TIM is most resistant against aging:happy:? There are reports claiming TIM performance degrading over time with repasting being a cure. But if one wants to avoid the repasting as long as possible which TIM to choose?
 
wirk said:
Hoping for a good amount of time before the EOL there is a question which TIM is most resistant against aging:happy:? There are reports claiming TIM performance degrading over time with repasting being a cure. But if one wants to avoid the repasting as long as possible which TIM to choose?
Click to expand...

I had MX-2 on my rig with an H80i for a couple years until I recently switched to a different AIO cooler. It was still as moist and pliable as the day it was put on. I switched to MX-4 since that's what came with the new cooler. Both perform very well and I can't see playing the "switching TIM's over and over" game to find the one that nets me an extra 1-2c cooler temps. My temps are fine at the highest overclock I can get with my chip, and that's all that matters really.
 
Not bad, but as others have posted and pointed out, the performance range (as in additional performance gains) going from any particular reputable thermal paste brand to another is negligible. (Perhaps gaining at best an extra 2-3 degrees in some cases).

For those that don't care about the 2 or 3 degrees of difference this can make, just pick one and run with it. AS-5 is a great choice. So is MX-4, MX-2, Prolimatech, Cooler Master, etc. They will all work just fine.

For those that want the very best, and don't want to bother with exotic liquid metals with their application headaches and other drawbacks, then you'll want to choose from something like the Thermal Grizzly Kryonaut or Gelid GC-Extreme.

The Thermal Grizzly Kryonaut has a thermal conductivity 12,5 W/mk which is amazingly high for a thermal paste. It's also non-conductive and has a temperature range of -200 °C / +350 °C, so it can also be used in extreme cooling applications such as LN2 cooling as well, which is what it was specifically designed for.

I consider the Thermal Grizzly Kryonaut the best thermal pastes on the market right now... but as to overall performance, it's only going to edge out other top tier offerings by a max of a degree or so.
 
I used to use the Arctic Silver Ceramique 2. Before that I used the original Arctic Silver Ceramique. As others have said, most quality thermal pastes are within ~5 C of each other. Longevity is where things matter if you're not doing extreme overclocking. I used to really like the Ceramique pastes as they're not electrically conductive or capacitive. Right now I'm using Antec Formula 7. It's more difficult to clean off than Ceramique pastes, but seems to be easier to apply. Some people have noted that other "diamond" pastes like IC 7 sand etching off of surfaces. I haven't personally experienced this, but you should be aware. I think that I'm going to try the Noctua NT-H1 next due to this.

I like to have one do all go to paste. Everything is conductive if you have enough juice running through it. But I try to stay with pastes that specifically state that they're electrically non-conductive and electrically non-capacitive. You probably know this, but in reference to electricity when something is conductive it means that electricity passes through it more or less easily. When something is electrically capacitive, what's being referenced is the materials ability to store a charge. Someone could probably school me on this, but both are bad. Highly electrically capacitive things are probably more dangerous to electronics than electrically conductive things. But I believe conductivity is an issue you're going to run into more often with thermal paste due to improper application. Nobody is perfect.

Speaking of perfection. Technique wise I used to use the renowned pea method. I use the X method right now in an attempt to cover more surface area. The more 'contact' for heat transfer the better in this instance. I've seen a lot of people in articles and videos highlight the importance of centering the paste over the area of the heat spreader where the die lies. However, though the entirety of the heat spreader may not be in direct contact with the heat source, it can and does absorb heat in a radiating fashion which is dissipated into the heatsink. Therefor, more contact in this instance will net you better cooling up to a point. But that's as far as we'll go down that rabbit hole.

A big issue not many people talk about is the lifespan and more importantly, thermal expansion and it's relation to the lifespan of the product. It's one of the reasons I switched over from Arctic Silver Ceramique products. Supposedly the Antec Formula 7 holds up great to thermal expansion or expands at a similar rate as the die, heat spreader and heatsink which leads to a fairly stable product over time per a lab test I reviewed. I can't find the article anymore though. To explain why this is an issue, when a paste cures and doesn't cope well with thermal expansion it can tear for lack of a better word and form gaps over time. The die or heatspreader and heatsink can expand and contract at different rates which can separate the TIM (thermal interface material) - known as thermal paste / pads to the lot of us. I haven't done any tests myself, but I have had the original Arctic Silver Ceramique thicken up on me over time and turn to an oil pastel like consistency - I'm talking a year or two though. But that's why I dumped it. So far so good with the Antec Formula 7.

Long story short, buy something of renown quality and longevity, but don't go overboard in price or difficulty of installation. Stay away from gimmicks.
 
You must log in or register to reply here.
Back
Top