Best thermal paste for lowest temps

G

Guest

Guest
Just curious what's the best thermal paste that yields the lowest temps out there?Heard some good reviews about AS5 but not sure if newer thermal pastes are out on the market.

Is the Gelid extreme good?

Thanks in advance.

 

Nedal0

Distinguished
Jan 12, 2012
427
0
18,860
I have heard about Gelid but have never used it. I have Used AS5 and a bunch of others in the past but was never really happy with the Temps it produced.

I then decided to go through Intel Thermal Solutions and Guide documents.

Intel recommends Shin-Etsu X23-7762 so I purchased a tube and have never looked back since. Shin-Etsu X23-7762 is more like Thermal Pad in a tube, and the composition is very different from all the others.

Currently maintaining 35°C on Idle and Max 45°C on Load. When I say Load, trust me I do really Load the PC as I run VM's, Photoshop, 3 Web Browsers and Other programs simultaneously.

Mind you I'm using an AIO PC and those are known to get a bit hot when it comes to Temps but Shin-Etsu X23-7762 does the job perfectly.

At the end of the day, its your decision, but do your own research and testing as you will have to find what works best for you.
 

aramisathei

Honorable
Aug 25, 2012
303
0
10,810
http://skinneelabs.com/2011-thermal-paste-review-comparison/2/

Should cover your question.

Personally, I use Prolimatech and I'm pretty happy with it.
Used AS5 and it was ok; Shin-Etsu and Tuniq-4 have also worked pretty well for me.

The biggest things to look for is know what kind of heatsink your cooler uses (copper vs aluminum, for example) as some don't mesh well depending on the type of material used.

An example is Coollaboratory Liquid Ultra, which hates aluminum and will significantly reduce the effectiveness of the heatsink.
 
G

Guest

Guest



Thanks guys! i did get a gelid extreme yesterday as i suspected my thermal paste was wearing out. Previously i was using a stock thermal paste from coolermaster that came bundled with my hyper 212 evo.

How is it on the heatsink of the my current evo?
 

PS-Matt

Honorable
Jan 11, 2013
21
0
10,540
Here are a few thermal paste comparisons:

http://www.hardwaresecrets.com/article/Thermal-Compound-Roundup-January-2012/1468/5
http://www.maximumpc.com/article/features/geek_tested_17_thermal_pastes_face?page=0,1
http://www.pugetsystems.com/labs/articles/Thermal-Paste-Roundup-172

The Arctic MX-2, Prolimatech PK-1, Tuniq TX-2 and Thermaltake TG-1 are the ones that are consistently the top performers. I personally like the Arctic MX-2, but any one of those will work great.

And I second on not trying the liquid ultra, that stuff is scary if it comes into contact with aluminum: http://www.youtube.com/watch?v=JHHI2Lk79cY
 

Nikky23

Honorable
Feb 1, 2013
2
0
10,510
Hi
I have a question as well.
What thermal compound/paste is best to use for a PS3 system 3 years old. I want to replace the paste that is already in there before the system over heats and crashes.
 

Nedal0

Distinguished
Jan 12, 2012
427
0
18,860
I believe there is an untouched subject when we talk about Thermal Paste. No one really takes the kind of PC into consideration and everyone automatically assumes its a Desktop PC.

Thermal Paste for Desktop PC, Laptops and Netbooks and All in One PC's really vary dues to the amount of Heat and type of Heat sinks each type of PC is restricted to.

It would be good if other form factors can be taken into consideration because Thermal Paste thats works well on a Desktop PC might just not cut it on a Laptop or AIO PC and vice versa.

 

Nedal0

Distinguished
Jan 12, 2012
427
0
18,860
So you think everyone who replaces their Thermal Paste is overclocking ?

We are not talking safe here. There are lots of people on forums worldwide with PC's, Laptops and All In One PC's asking questions.

I'm not sure what your point is really you should not even be answering to this as what you are saying is pointless.

Put your hands in the air, step away from the keyboard and back away slowly ........
 

Nedal0

Distinguished
Jan 12, 2012
427
0
18,860
The OP asked "Best thermal paste for lowest temps".

It would be nice if the question was answered accurately and kept short.

This whole rant does not even make any sense or even answers the OP's question

Simple "Thermal Narcissism"

Oh well... we are all entitled to an opinion but...........

I wouldn't recommend this rant any day. It was written whiles discrediting others. What makes it any good ? Nothing !

Just my Opinion again !






As usual, some really bad advice in this thread.

Gelid GC extreme is a high end, modern paste, but it's far from best. It's decent, you won't see significant temps using any other ceramique (well, significant relatively, it'd be significant by TIM standards).

Intel recommends Shin-Etsu X23-7762 so I purchased a tube and have never looked back since. Shin-Etsu X23-7762 is more like Thermal Pad in a tube, and the composition is very different from all the others.

Shin etsu is okay. It's a bit dated, used to be the hot stuff a few years ago. As is common with heatsinks and thermal pastes, what is good is not what gets recommended, but what has been around long enough to get outdated and used by enough people to get in a vicious cycle of being recommended over and over again.

http://skinneelabs.com/2011-therma [...] parison/2/

Should cover your question.

Personally, I use Prolimatech and I'm pretty happy with it.
Used AS5 and it was ok; Shin-Etsu and Tuniq-4 have also worked pretty well for me.

The biggest things to look for is know what kind of heatsink your cooler uses (copper vs aluminum, for example) as some don't mesh well depending on the type of material used.

An example is Coollaboratory Liquid Ultra, which hates aluminum and will significantly reduce the effectiveness of the heatsink.

Skinnee was awesome but he's not around anymore, unfortunately (inactive, not dead). Prolimatech is great by PK-1 is not. It's terribly outdated. To give you some perspective, PK-1 is not a good paste (i mean it was what, a decade ago, it's still 'okay'), and every paste that pK-1 beats on that benchmark (skinnee's tim review is probably one of the few large scale benches you can trust), is terrible nowadays.

coollaboratory liquid ultra ! :)

CLU is an extreme solution. It's fine, but prohibitively expensive (i use it, btw). 1.5g for $18+. You can find 1.5g of pk-3 or masscool for under $4, if not for free from some places.


Here are a few thermal paste comparisons:

http://www.hardwaresecrets.com/art [...] 012/1468/5
http://www.maximumpc.com/article/f [...] e?page=0,1
http://www.pugetsystems.com/labs/a [...] oundup-172

The Arctic MX-2, Prolimatech PK-1, Tuniq TX-2 and Thermaltake TG-1 are the ones that are consistently the top performers. I personally like the Arctic MX-2, but any one of those will work great.

And I second on not trying the liquid ultra, that stuff is scary if it comes into contact with aluminum: http://www.youtube.com/watch?v=JHHI2Lk79cY

God bless hardwaresecrets articles, but their TIM and heatsink benches are absolute crap. Phobya Hegrease Extreme is the best ceramique (and it's price reflects that). As5, pk-1, etc, are NOT better than it. Also, testing a 3.3ghz overclock on a lga1156 CPU is hardly conclusive evidence, you might as well test on stock cooling on an amd chip.

TIMs and heatsinks don't reveal their true performance profiles until you really push them. It would be more useful if we saw 4.5ghz@1.4vcore on Ivy/Sandy. Not a mild overclock on a old chip run at low voltage. They also ran large ffts in p95, which is extremely variable, and is going to be a good 10-20*C lower than small fft (in particular, a custom 8 fft min/max length). They also used way too much thermal paste, as evident in the pictures.

Finally, they take an arithmetic reading of all cores (not all the cores even get calibrated, most people have one or two cores reading as ambient or below ambient when that simply isn't possible), instead of using a max core temp (which isn't very reliable either, but it's more reliable than what they are doing and I'll accept a max core instead of a thermal diode placed in a milled CPU just for a general picture).

Oh, they report delta temps. Which is a terrible, as there is an exponential impact of ambient temps on load temps, ie 1*C change in ambient temp will lead to something like a 4*C variance in load temp. Gotta retest temps in a controlled setting, ie all the same session. There isn't enough spread on these pastes to really show the difference in them. And no wonder, using basically a low overclock system, it isn't stressing it.

Like a Hyper 212+ will outperform custom water loops on stock/low overclock. Not because it's better, but because Heatpipe Direct Contact works really well at low temps/powers, and is super cheap (ie its good for budget and lower overclocks), but custom water will destroy the hyper 212 in any raised voltage.

Puget review is good but really dated. Those pastes are all outdated. Like no one is benching an 4850 anymore to see how it compares to a 4870, no one cares what outdated equipment does. They also use too much paste, you might think that's okay if they use too much on each paste, but some pastes relatively, do better, than others when you got too much of it. That's enough to influence temps at least 3-5*C.

Maximumpc doesnt reveal how they test. I wouldn't trust their article.

MX-2, PK-1, TX-2... you realize there is an MX-4, a PK-3, and TX-4 right? You realize that in the last decade, CPUs and GPUs have advanced quite a bit right? Well, thermal pastes have advanced just as much.

You have people spreading misinformation because they have never used the aforementioned pastes and don't know how to read a review. It doesn't take much to realize that maybe you shouldn't recommend 10 year old pastes, that maybe you should recommend PK-3, or MX-4, or TX-4 instead.

There is very little difference in modern, high end pastes, yes. But there is a HUGE difference, between a high end, modern paste, and a paste that was crap 5 years ago.

Just for reference, PK-1 beats the popularly recommended, and total crap, AS5, by about 5-10*C just on an AMD system where the temp range is 50-60. That's as large a difference you get from going from a hyper 212 with 2 powerful fans and decent tim, and an Nh-d14!

Arctic MX-4 is the best stuff out right now. Hands down. Ehh, tuniq TX-4 might be alittle better, but either are topline products.

Please do not spread total misinformation. Someone read your post and think you know what you are talking about.


Hi
I have a question as well.
What thermal compound/paste is best to use for a PS3 system 3 years old. I want to replace the paste that is already in there before the system over heats and crashes.

Anything is fine, get what you can get for the cheapest, really. if you can't use some from a leftover tube or from a friend, get a 1.5g tube of some PK-3 or masscool for under $4. High end, modern, pastes, with enough paste for at least 10 applications (unfortunately they do not sell .1g tubes).

Right now, good pastes are:
Noctua NH-T1 (eh, more like decent, then great, but it's better than any outdated paste for sure)
Masscool
PK-3
MX-4 is okay too
Generally you can find these for free to under $4 for 1.5g, $6-8 for 3.5g, and $6-9 for 5g. There's a few others out there too, but PK-3 is the best 'standard' paste and you can find 1.5g tube for $3 on ebay.

For extreme pastes, you got:
- Phobya hegrease (it's a ceramique but the price is much higher than any other ceramique, better than any ceramique though) about $13 for what is usually $8-9 for most 5g ceramique tubes.

Then the Coollaboratories. Ultra > Pro, in terms of ease of application, ease of removal, how corrosive it is, and it's performance. These are the best 'TIMs' money can buy. They stain heatsinks and IHS though, and you can't use them on aluminum (they don't 'melt' aluminum as some uninformed people say, they just corrode it after long term use, slightly, but enough you dont want to use a hyper 212 with this stuff, for example). $18+ though, so they are bit more expensive, and not as ubiquotous (ie maybe you just want to have a single 1.5g last the rest of your life and be able to be used on your chipsets, GPU blocks, aftermarket cooling, low end heatsinks, et cetera).

(aluminum does not equal silver, aluminum is like the hyper 212 heatsink base besides the copper, most heatsnks, particularly higher end ones, are not copper but nickel-plated copper, nickel is slightly lower thermal conductivitiy, just barely, but WAY more durable, that's why you never see 'copper' on heatsinks anymore, ie the nh-d14, phanteks, it's all nickel plated copper - all the benefits of copper without the negatives)

And then Indigo Extreme is the best material, it's 2 applications for $20 and it's like a pad that you have to burn in at 90*c and it's quite difficult to work with, damages ihs/heatsink.

basically, pk-3 is your best bet. I haven't used indigo extreme because I am always unmounting/remounting my cooler, but I've used lots of pastes, I've even posted articles and reviews of pastes, so I at least know what I'm talking about here.

Unlike what most people say, using a good paste can make a huge difference in temps. There's not much difference in modern pastes, but the problem is that everyone recommends terrible pastes like as5. As5 is over 13 years old. It's as bad to modern pastes as a pentium 2 is to an ivy bridge. It's ridiciulous. And stock pastes are usually crap, tbh.

edit: its easy to use too much paste, but about impossible to use too little. a rice grain borders on too much paste, you really dont want to use enough paste to cover the IHS, you just need to cover the die. Like a fat rice grain would drop you ~5*C in temps for being too much. Also don't spread it (except clu/clp have to be spread due to their mercury like nature), always use dot/ricegrain. air bubbles arise from spread method.
 


I second your opinion! Ryan
 

klepp0906

Distinguished
Apr 29, 2013
150
0
18,710


that's not load. id be surprised if half your cores are at half utilization half the time =p run intel burn test or linpack or prime95 small FTT for an hour or so and get back to us on your temps =p
 
G

Guest

Guest
hey all, question I have is what if any are the differences between arctic silver 2, arctic silver 3, arctic silver 5 ? I have tubes of these and just wondered the difference in them. I get excellent results from all of them. tried to email artic silver and was not able to get through. thanks in advance. poorguy
 
G

Guest

Guest
hey all, i used to think that the type of thermal paste used really mattered but in my opinion it really doesn't. i have been in the electronics repair for years and yes all of the thermal pastes are cool with all of there frills but have never ever noticed any more difference then 1 to 3 degrees difference when compared with the old white transistor heat sink compound paste that i have been using for years. for along time i have searched and have not found any real proof that the silver or ceramic base compounds are really that much better. i have never seen any of them dry out or harden and i have used most of the arctic silvers and they are good but 1 to 3 degrees just ain't that impressive to me.
all my computers are desktops and all are running stock heat sink fans. i have also used the killer super heat sink fans also. for best results i have found big cases and large fans that move lots of cfm are the best solution for a cool running tower.
all in ones are nothing more than a laptop with a stand on it and don't care for them at all. a tablet is way better.
just my thoughts and experience and also only my opinion. thanks the poorguy