What the best thermal paste in 2019?

[Vayra86]

Y'all are way overcomplicating the thermal paste race.

If you can get no air bubble application done, you're done. Its that simple. Paste is the least of your worries. Dust, case airflow, heatsink dissipation capacity versus the CPU underneath, those are the real things to look at. And that is why a middle of the road paste is JUST FINE and if you want to spend big for peace of mind, by all means, but its nothing more or other than just that.

/thread. And that also goes for the 2020 version of this topic...

 

[Bill_Bright]

It does have actual silver in it right? And silver is metal right? I'll still pass on AS5

So what if it has metal in it? Metal typically is a great conductor of heat, and silver is one of the best. That's a good thing. And that's why the metal TIMs top that list. So what if a 20 year old formula? Again, is 3°C really that important? NO!!!!!!

Why is everyone so butt hurt about AS5?

Because of nonsense like yours above, and below.

Electronically speaking everything is conductive

Just because you may apply a big enough potential to cause voltage to arc across a gap, that does not mean electricity was conducted. Tunneling is not conduction.

I guess if you are so careless you squeeze out the entire tube of AS5, causing it ooze all over the motherboard, then you have a problem. But in that scenario, you have no business being inside the computer in the first place.

Geez. I had no idea the situation was so bad.

It's not.

 

[Raven Rampkin]

If you can get no air bubble application done, you're done. Its that simple. Paste is the least of your worries.

Unless it's KPT-8 from the Soviet times or straight up mayo

 

[TheoneandonlyMrK]

But, damn man, now I want to try liquid metal on my 1070s...

Nowt beats it IMHO.

 

[jayjr1105]

Again, is 3°C really that important? NO!!!!!!

Because of nonsense like yours above, and below.

I'm sorry, was the title of this thread "What's the most mediocre thermal paste in 2019 with a few drawbacks" or "What's the best thermal paste in 2019"???

If you want to get super technical and pick apart my posts with a fine tooth comb, then AS5 isn't even in the ballpark of discussion and shouldn't even have been mentioned in this thread. The thread starter clearly asked about the best thermal paste in 2019. Simple as that. And AS5 isn't in this discussion anymore because it simply doesn't perform as well as a dozen more modern pastes, not to mention it's difficult to clean up compared to carbon based TIM's, has a 200 hour break in period, and gives you no incentive to buy it from a price perspective.

If someone starts a thread about what's the best TV in 2019, nobody is suggesting a Plasma or CRT just because technically it still "works".

 

[lexluthermiester]

Why is everyone so butt hurt about AS5? It's a 20 year old formula.

It's a formula that works. It's age is irrelevant, something I suspect you have yet to understand.

 

[Bones]

Ceramique is junk.

Have to disagree, I use it all the time and it works fine...... Plus it's cheap too.
With what I do it's gotta be cheap.

Having not even tried it once, I already feel allergic to the Kryonaut and the buzz surrounding it. Sorry about that ¯\_(ツ)_/¯ Probably a game changer at an edible price but some sellers suuuuuure have decided to cash in on the hype.
Otherwise, MX-2 if you're a retrofreak/pervert, MX-4 for general use, GC-Extreme for extreme use, and GD900 if you're a pervert once again Minor experience + major Google-fu is the source in case you wonder There are the Noctuas, the Mastergel and some other options as well if you love experiments.
Also, wasn't the Arctic Silver's manufacturer the one behind that major customer service phVckuppe a few years ago? (I mean, that one with grandmas, where a rep went nuts in the respective forum thread.) Had something to do with paste abrasiveness iirc. EDIT they weren't

I've tried Kryonaut and it's..... OK.

Mainly meant and formulated for extreme use (Subzero), for everyday use it's not worth the expense at all.
I've been using the newer MX-4 Carbon stuff https://www.newegg.com/arctic-cooling-ac-mx4/p/2MB-000S-00017 and that works good without being too expensive for daily or any other use. Works with Ln2 as well, did that a couple weeks ago with it - No problems at all, even at full pot.

I'm sorry, was the title of this thread "What's the most mediocre thermal paste in 2019 with a few drawbacks" or "What's the best thermal paste in 2019"???

If you want to get super technical and pick apart my posts with a fine tooth comb, then AS5 isn't even in the ballpark of discussion and shouldn't even have been mentioned in this thread. The thread starter clearly asked about the best thermal paste in 2019. Simple as that. And AS5 isn't in this discussion anymore because it simply doesn't perform as well as a dozen more modern pastes, not to mention it's difficult to clean up compared to carbon based TIM's, has a 200 hour break in period, and gives you no incentive to buy it from a price perspective.

If someone starts a thread about what's the best TV in 2019, nobody is suggesting a Plasma or CRT just because technically it still "works".

View attachment 133248

Technically speaking, since it's still being made, sold and used in 2019 it does matter.

 

[killster1]

Have been using Thermal Grizzly Kryonaut since 2016.

yea same here but i recently found out that its only good for 4 years or less? really i didnt even read much about it, its pretty cheap on ebay 8$ for maybe 40 applications.

 

[jayjr1105]

It's a formula that works. It's age is irrelevant, something I suspect you have yet to understand.

Is the tile of this thread, "Thermal paste that works"? or "Best Thermal Paste 2019"?

 

[lexluthermiester]

Is the tile of this thread, "Thermal paste that works"? or "Best Thermal Paste 2019"?

It's still being manufactured and will continue to be, and is still in the top 3 in performance/cost ratio. You were saying?

 

[jayjr1105]

It's still being manufactured and will continue to be, and is still in the top 3 in performance/cost ratio. You were saying?

Lots of old crap is still manufactured because of it's mind share. Top 3 in the perfomance to cost ratio? You realize we're talking about pennies here right? #17 in that chart with a 9900K. 16 TIM's performed better (that they tested), there's probably more.

 

[FordGT90Concept]

If you want to get super technical and pick apart my posts with a fine tooth comb, then AS5 isn't even in the ballpark of discussion and shouldn't even have been mentioned in this thread.

AS5 has to sinter which takes hundreds of hours and thermal cycling. It's not an easy thing to benchmark.

I installed this HSF with a decade old AS5 (purchased around 2005) back in 2015. 4 years later, the temps under 100% load are either the same or less as they were in 2015 (~50C across all cores). Of course I don't know what the ambient temperature was then versus now. BOINC may also be working on something computationally different then versus now. But yeah... computer goes off almost every night and when it's running, it's at 100% like it is now. The AS5 is definitely fully sintered by now.

Been using AS5 exclusively for over a decade, no complaints here.

 

[jayjr1105]

AS5 has to sinter which takes hundreds of hours and thermal cycling. It's not an easy thing to benchmark.

Good point, another reason to not buy it.

Been using AS5 exclusively for over a decade, no complaints here.

Not saying it's terrible or that it will blow up your system. There are just better pastse nowadays, that's it.

 

[FordGT90Concept]

Good point, another reason to not buy it.

If you're changing HSFs every day/week/month, definitely. AS5 is only for the long haul.

 

[lexluthermiester]

AS5 has to sinter which takes hundreds of hours and thermal cycling. It's not an easy thing to benchmark.

Oh please stop that myth. Yes it has cure time, however it works within 96% of maximum until it does. Mountains out of molehills, please stop making them.

Good point, another reason to not buy it.

Not saying it's terrible or that it will blow up your system. There are just better pastse nowadays, that's it.

Ok, we understand your opinion. We don't have to agree with you though..

 

[FordGT90Concept]

Oh please stop that myth. Yes it has cure time, however it works within 96% of maximum until it does. Mountains out of molehills, please stop making them.

Considering the worst score and best score is a split of only 8%, you're random number stipulating a 4% change is huge.

It's not a myth, a Department of Energy study looked in to it. Of all the compounds they tried, sintering silver (Arctic Silver 5) is the only one that improved from the initial test. The rest (except graphite thermoplastic which never changes) stayed roughly the same until they started degrading.
Final summary: https://www.nrel.gov/docs/fy16osti/66754.pdf

 

[FreedomEclipse]

 

[FordGT90Concept]

His method of spreading paste is terrible. GD-900 may have performed the best for him simply because it's so viscous that it worked all the bubbles out itself.

 

[lexluthermiester]

It's not a myth, a Department of Energy study looked in to it.

I remember that thread and nothing has changed. The testing method was flawed and the use case scenario is not applicable to cooling PC components. Perhaps you would be well served by re-reading that thread yourself.

His method of spreading paste is terrible.

Saw that a day after it was posted, excellent testing methods as well as conclusions.

 

[FordGT90Concept]

The testing method was flawed and the use case scenario is not applicable to cooling PC components.

By "flawed" you mean scientific? They used a thermal couple as a heat source which provides constant wattage, they apply exactly 100 micrometers of gap with exactly 60 (I think it was) bars of pressure, and they cycle the thermal couple in a predefined pattern. Yeah, I'm gonna go with NERL results over random YouTube guy, spank you very much.

Saw that a day after it was posted, excellent testing methods as well as conclusions.

[facepalm.jpg]
I marked the obvious places that will form an air pocket in just one shot:

I have no confidence in his testing. One air bubble over the sensor and the reading is wrong. His application methodology makes it inevitable.

Air pockets are second only to complete vacuums in terms of thermal resistance.

 

[freeagent]

How do you get air bubbles when you plant your cooler, and clamp it down to proper specifications, ie all the way..? You get air pockets when you lift your cooler off to check your tim print, but at that point I'm sure you already know to wipe it off and start fresh.

I find a lot of the stuff on youtube to be entertaining. I don't take most of it seriously, and neither should you..

 

[FordGT90Concept]

Valleys become air pockets. If you're not hearing pops as it is being tightened, you either have no air pockets or the air is trapped inside. In general, there isn't enough force (because PCBs are weak) in computing applications to force the air pockets out. That's why proper application of the TIM is critical.

 

[freeagent]

I haven't heard any pops like you describe. That's hardcore. I guess if people mount their coolers with the plastic still on, anything is possible.

Edit:

How much paste do you have to use to get that to happen? Lol.

 

[jayjr1105]

Exactly.

You also get bubbles if you do a piss poor job of applying the TIM, not just if you lift it off and press right back down.

 

[FordGT90Concept]

I haven't heard any pops like you describe. That's hardcore. I guess if people mount their coolers with the plastic still on, anything is possible.

As I said, you'll hear the PCB crack before you hear the pops of air bubbles escaping (it's a function of air pressure, applied pressure, and viscosity). That's why you have to apply thermal paste/grease in a way that prohibits the formation of valleys that will become air bubbles.


You know what? I do have two pieces of dense, clear plexiglass here, I could demonstrate how air bubbles form but...meh...can't be arsed to do it when it should be common sense. Case in point, ever get an air pocket when applying a screen protector? Same thing.