The best thermal paste to get?

[PaulW]

Hi All,

Going to make a cheap AM4 + a AM5 monster build in the new year and was wondering which thermal paste people recommend these days? many moons ago i used Arctic Silver. what its the go to that everyone swears by now?

Other question is isopropyl alcohol, I may use the opportunity to add new paste to my old ageing CPU and cooler. Can you get isopropyl alcohol wipes for cleaning CPUs? I really don’t think I need loads of the stuff.

Thanks

Paul

 

[JoelLehtinen]

you can use Cooler Master MasterGel Pro v2

 

[Armageus]

Arctic MX-4 is a decent "go-to" paste, that's got great longevity.

There are slightly better performing pastes (e.g. MX-6, the various Thermal Grizzly pastes etc), but generally they are only a couple of degrees better, and some of them come with bigger disadvantages - e.g. needing to be reapplied within a year, thermal pumpout with newer hotter CPUs etc.


My basket at OcUK:

• 1 x Arctic MX-4 Thermal Compound (4g) (SKU: TH-01B-AR) = £4.99

Total: £8.99 (includes delivery: £3.99)​

​

 

[Glanza]

I recently tried Arctic MX-6 and been happy with it, granted needs more effort to spread than other pastes.

 

[Troezar]

Have a look at thermal pads, they seem pretty good now, Thermal Grizzly for instance.

 

[MartinPrince]

Arctic MX-4 is a decent "go-to" paste, that's got great longevity.

There are slightly better performing pastes (e.g. MX-6, the various Thermal Grizzly pastes etc), but generally they are only a couple of degrees better, and some of them come with bigger disadvantages - e.g. needing to be reapplied within a year, thermal pumpout with newer hotter CPUs etc.


My basket at OcUK:

• 1 x Arctic MX-4 Thermal Compound (4g) (SKU: TH-01B-AR) = £4.99

Total: £8.99 (includes delivery: £3.99)​

​

​

​

This.

Just get the cheapest per gram decent thermal paste going as there is little difference between them. So MX-4 will do you just fine. Just don't buy Thermal Grizzly Kryonaut as that is an over-hyped and over priced rip off in my personal testing and opinion.

I personally use Noctua NT-H1/H2 as when I tested some of the top one's this came out as the best, both from temps, ease of use and longevity.

 

[Skully178]

^^ Most of the (decent) ones are the same. I've bought some MX-4 in the past and it's fine. Recently used NT-H2 and am happy with it.

 

[SpudMaster]

I'm currently using Thermal Grizzly Kryonaut and TBH i doubt I will buy it again. Theres nothing wrong with the paste, relatively easy to appy and good temps but it really isn't worth the cost. Barely noticed any difference verse MX-4, maybe a degree or 2. I'll probably buy MX-6 when I need to repaste

 

[PaulW]

Thinking of going with MX-6, seems pretty good and affordable.

 

[Littlenose]

that's a really interesting chart - thanks for posting it.
Confirms for me that I'll sticking with the standard pastes rather than the exotic/expensive ones.

 

[edGfaCTor]

Price wise it looks like older MX-4 is still by far best value and very easy to apply

Liquid metal is the best thermal wise.

 

[random_matt]

NT-H2 or PK-3.

 

[rare]

I always use MX-4, it’s cheap and easy just how I like em

 

[Gray2233]

Another mx4 advocate here, you don't gain any appreciable advantage with more expensive pastes.

 

[srekal34]

I can vouch for Honeywell ptm7950 as an alternative. Super easy to use and only beaten by liquid metal.

 

[Pooh]

I’ve used liquid metal in the past. Wonderful performance but very fiddly to use and easy to create a world of pain.

Concur with consensus above so if you’re not pushing the envelope that any of the usual suspects will be just fine

 

[Rroff]

I recently tried Arctic MX-6 and been happy with it, granted needs more effort to spread than other pastes.

Though there is a variant with a spatula you generally want to avoid hand spreading unless necessary on a large IHS - even LGA1700 probably on balance is best not hand spread.

MX-6 uses a silicone gel carrier which achieves optimal distribution under mounting pressure in most cases.

MX-4 is generally fine for AM 4/5, but I'd lean towards MX-6 for Intel LGA1700 CPUs.

And ignore most reviews of AS-5 as most use sub-optimal application methods, often out of a misguided sense of "fairness" in testing, and don't bother with the curing time it needs to get optimal results.

 

[Rroff]

The How-FixIT video you got that from, though they did briefly mention specific instructions for MX-6, appears to have spread all the pastes by hand for testing and not take into account curing time.

EDIT: In my own testing I've got AS-5 as close as makes no difference to Kryonaut and I know it will stand the test of time, but for my 14700K build I went for MX-6 for whatever that is worth.

 

[MartinPrince]

The How-FixIT video you got that from, though they did briefly mention specific instructions for MX-6, appears to have spread all the pastes by hand for testing and not take into account curing time.

Yes, none of the other Thermal paste tests I have seen over the years incorporate curing time as it is really not feasible when testing 10+ paste. How long do you leave it for?

Plus when testing multiples of pastes ensuring there are no air gaps or areas not fully covered is a prerequisite; spreading the paste by hand is the best way to do this. So I commend then for doing that as it is much more time consuming.

 

[Rroff]

Yes, none of the other Thermal paste tests I have seen over the years incorporate curing time as it is really not feasible when testing 10+ paste. How long do you leave it for?

Plus when testing multiples of pastes ensuring there are no air gaps or areas not fully covered is a prerequisite; spreading the paste by hand is the best way to do this. So I commend then for doing that as it is much more time consuming.

AS-5 can be as much as 200 hours, but most happens in the first few hours ~1C, with maybe 0.25-0.5C reduction after that.

For AS-5 potentially you are making it worse for air gaps, etc. hand spreading, for other pastes a more complex story, though a lot depends on the surfaces of the IHS and heatsink as to how much there is too that, it can also affect longevity not just performance. Most pastes have specific application instructions for optimal results which can vary a lot from Thermal Grizzly stuff which is often designed for hand spreading and can suffer from other methods, through to pastes which get very poor results from hand spreading so it isn't really good testing methodology except maybe as a guide for people who persist with hand spreading irrespective.