Arctic MX-6 shelf life is just a couple months?

[natr0n]

Any emulsion can and will separate over time or if not mixed well.

Like mayonnaise.

You can always shoot some out and mix it with a stick/spatula and apply.

 

[BoggledBeagle]

The user replaced pads may be of the same dimension but the stiffness of the pads create a gap to the Heatsink area and the GPU die

The extremely thick layer of paste is really suspitious, the paste is meant to work as a thinnest layer between two bodies that touch each other. It is not meant as a gap filler, for that there are thermal PADS or even the graphene sheet.

If used as a gap filler, the paste is misused and you really cannot draw any conclusions about a product used in an unintended way.

 

[GerKNG]

The extremely thick layer of paste is really suspitious, the paste is meant to work as a thinnest layer between two bodies that touch each other. It is not meant as a gap filler, for that there are thermal PADS or even the graphene sheet.

If used as a gap filler, the paste is misused and you really cannot draw any conclusions about a product used in an unintended way.

Regarding the TF7 photo: why are you applying so extremely thick layer? You applied 20-50X more paste than needed.

What was the reason you are repasting a card under warranty? Are you using an original cooler? Are you sure you are mounting it correctly?

"extremely thick layer"
first of all you can dunk a whole tube of paste on a GPU and nothing happens. and this was a hairthin layer of TF7 that is like clay and had to be spreaded for a minute before it was even all over the die.

and if you want to know how the MX6 looked like:
pretty much like on one of my 6800XTs.

 

[BoggledBeagle]

"extremely thick layer"
first of all you dunk a whole tube of paste on a anything and nothing happens. and this was a hairthin layer of TF7 that is like clay and had to be spreaded for a minute before it was even all over the die.

and if you want to know how the MX6 looked like:
pretty much like this one of my 6800XTs.

View attachment 346663View attachment 346664

That does not look good...

I do not want to promote one product, but have you tried the graphene sheet by DerBauer?

 

[GerKNG]

i plan to buy a sheet of PTM7950 if the TF7 does not work.

 

[VT1185MF_Rosaria]

The extremely thick layer of paste is really suspitious, the paste is meant to work as a thinnest layer between two bodies that touch each other. It is not meant as a gap filler, for that there are thermal PADS or even the graphene sheet.

If used as a gap filler, the paste is misused and you really cannot draw any conclusions about a product used in an unintended way.

I think you got my statement wrong.
The scenario in which I explained pump-out acceleration DOES NOT imply if paste is used as gap filler.
The scenario I described refers to the unexpected creation of micro-gaps due to the difference in pliability of thermal pad material differences between user applied pads and factory applied pads, therefore accelerating the pump-out effect due to having too much of a gap between the die and heat sink contact.

And if we are being needlessly pendatic, thermal paste IS used as gap filler as there are microscopic imperfections between the die and the heat sink contact area that will prevent effective heat transfer due to not being able to contact the surface area of each other.

GPU dies are not as susceptible to paste overfill issues leading to gaps compared to other direct die TIM applications (such as laptop heat sink mounts) as GPU heat sink assemblies have significantly stronger mounting pressure that can and will push out excess paste (Unless there is a significant curving on the GPU heat sink contact, but that is another issue).

I overfilled Thermalright TFX application my Colorful RTX 3070, and it had no adverse effects on the specified heat sink gap tolerances other than consuming too much thermal paste from the syringe. I can provide temperatures as proof, and this is at an area where 35 degrees Celsius is the standard average daytime ambient temperature.


I see no issue with the amount of paste that the original user posted, not unless the mounting pressure applied by the heat sink mounting screws is highly unequal in nature (which is very, very unlikely).

And to the original poster, Honeywell PTM 7950 (and similar products under other brands) will work well on that application. I used it on my Sapphire Nitro RX 5700 XT last year and it still does not show any signs of significant temperature deviations between core to junction temperatures. I can run a stress test later to show proof, if needed.

 

[lexluthermiester]

Well... 4th tube of MX6, massive pump out within a couple days.
View attachment 346644

Paste turned into soup again.
Repasted with TF7 which i found in the box of an old thermalright cooler from ~2 years ago.
First test with TF7: Temps are 3-4°C Lower than MX6 and the Hotspot delta is so far just ~8°C.
View attachment 346648

I call moose muffins. Show us evidence.

Where are the photos of this happening? If you post this sort of information, you should provide some evidence.

Right?

Regarding the TF7 photo: why are you applying so extremely thick layer? You applied 20-50X more paste than needed.

That photo speaks volumes. I see an application of TIM with no regard for care & attention.

first of all you can dunk a whole tube of paste on a GPU and nothing happens.

Nonsense. Many TIMs are harmless, but not all.

 

[GerKNG]

after more than a week of testing, even the old tube of TF7 that was free in the box of a 20 dollar cooler is still perfect after ~ 100 hours of full load. (had to repaste twice with MX6 in the same timeframe)
no more 25°C Hotspot deltas and hitting 90s at stock clocks after a couple days.

overclocked, 300 RPM lower fan speed than usual.
Cyberpunk with Pathtracing.

 

[BoggledBeagle]

I removed a CPU cooler after a few months and I noticed wet (oily) patches at the edges:





I cannot find any date or batch number on the tube or the package, it is a 4g tube.

 

[FreedomEclipse]

I cannot find any date or batch number on the tube or the package, it is a 4g tube.

It will be on the side of the box next to any printed barcodes. There is something called 'UFI' followed by some numbers like a windows activation key. Im assuming this is the products unique identifier which Arctic might ask for if you report something wrong with their product as it will give them the relevant information they need to troubleshoot.

There is also a P/N but im guessing that there is nothing unique about that but i could be wrong.

In any case. its all about the PTM/Thermalright Heilos now.

 

[lexluthermiester]

I removed a CPU cooler after a few months and I noticed wet (oily) patches at the edges:

View attachment 355476
View attachment 355477
View attachment 355478

I cannot find any date or batch number on the tube or the package, it is a 4g tube.

Looks good to me. Not seeing anything out of sorts.

 

[Brotato]

Man, really? I just ordered Mx-6 2 days ago. Should I cancel it?

I bought some recently from frozencpu. Squeezed out just fine. No separation or anything.performs as expected.

there is one paste that i have been very impressed with however, arctic silver ceramique. Version 1 not 2. Not a top performer, probably bottom barrel these das but when i ran out of paste last year, i found some ancient tube of the stuff in my parts box. No separation and spread well. Temps were fine.

 

[ShrimpBrime]

I bought some recently from frozencpu. Squeezed out just fine. No separation or anything.performs as expected.

there is one paste that i have been very impressed with however, arctic silver ceramique. Version 1 not 2. Not a top performer, probably bottom barrel these das but when i ran out of paste last year, i found some ancient tube of the stuff in my parts box. No separation and spread well. Temps were fine.

Ceramic on what cpu? What is "temps just fine"?
Only curios, because many years ago was utilizing this paste on 125w processors and it performed considerably worse than As5.

 

[Brotato]

Cpu. 7800x3d. I don’t remeber the exact temps but i remember noticing that it was about 3c higher than the thermalrightt paste that comes with a thermalright frozen prism which is supposedly pretty good. Only remembered the 3c as i was like “the fuck?”.
i was remounting the aio in an a4-h20 after noticing that one of the tubes was pushing up against my ram and had to remount the block to move the tube.

edit: it was a bitch and a half to remove later though when i eventually took apart the case build and moved the aio to a relative’s pc and stuck the build on a bench table. It was like cement.
Sort of wish i didnt toss the rest out. I’d be curious to test it out on my loop.

 

[Hankieroseman]

There might or might not be enough for one more new processor. It doesn't really get old.

I got the new stuff-6

 

[freeagent]

There is lots in there, the plunger goes all the way down.

I am assuming that you have not seen AS5 separate?

It happens..

 

[lexluthermiester]

I am assuming that you have not seen AS5 separate?

It happens..

Not here. I'm sure it's happened, but it's got to be on the rare side.

 

[Bones]

I've honestly seen it before myself concerning AS5.

I can't speak to MX-6 since I've never tried it before.

 

[Hankieroseman]

I've honestly seen it before myself concerning AS5.

I can't speak to MX-6 since I've never tried it before.

I just did and they recommend the pea sized drop in the center and let the heatsink do the rest. easy peasy really

 

[Space Lynx]

I'm just going to do a Thermal Grizzly Kryosheet, its just as good as any paste, no mess, and will last a life time. I honestly see no reason why not at this point.

 

[Scooby24]

Joined so I could post this - there is DEFINITELY something wrong with MX-6 paste. I had repasted a Gigabyte Gaming OC 4090 November of 2023. Temps were great 60c GPU, 58C Memory junction, 69c Hotspot. After 1 year at roughly the same fan speed - temps were 79c, 58c, 101c. So pads were keeping memory the same temp still but GPU and Hotspot were way higher. Repasted and new pads...repasted with Noctua NT-H2...now GPU was at 62c, 60c, 70c. So basically about the same as the new MX-6...but 17c lower than old MX-6. Yikes.

attaching pictures of results:
Factory paste/pads left - Fresh MX-6 and Oddtone 12.8 W/mK pads right applied on 11/02/2023: (max values from benchmark highlighted)
60/58/69

Here's after about 1 year - note the memory junction temps (pads) were completely unchanged, versus the GPU/Hotspot temps taken 12/08/2024
79/58/101

Here's after removal and reapplication of paste and new pads - this time NT-H2 and using Arctic's TP-3 Pads. 12/09/2024
62/60/70

The Tube of NT-H2 I have is at least 2 years older than the MX-6 but it's shown no signs of age and on my other applications, no degradation. I did open a ticket up with Arctic and offered to send them my sample in but there doesn't appear to be any consistency issues with it coming out of the tube. By all accounts it LOOKS and applies fine. Its performance just doesn't last.

 

[Scooby24]

Chiming back in as I think my issue with MX-6 with my GPU was the less a problem with the paste degrading and more of an issue with pump out. I've been monitoring the NT-H2 that I applied temps are starting to creep up a little bit (3-4c higher than first tested). That got me thinking there's something more at play...I didn't realize the direct die nature of the GPUs causes the pump out effect until I read more from Igor's Lab.

I've decided to redo it again but use Thermal Grizzly's Phasesheet (PTM 7950).

MX-5 related news though - I just upgraded my daughter's 3600 to a 5700x3d and took off the cooler that's been using the MX-5 since I installed the cooler in 2021....and the paste was no longer liquid. Temps were fine on the CPU but it was definitely dry. I'm just going to avoid arctic pastes altogether.

Based on Igor's testing I'm going to use Kooling Monster KOLD-01 on my CPUs...that seems to have tested very promising.

 

[freeagent]

I like a nice, thick TIM.

Not thin TIM put on thick.

 

[Krit]

I'm always using MX-2 or MX-4 the best value for the money. Also used MX-5 but not MX-6.

MX-2 and MX-4 shelf life will last for many years.

 

[_roman_]

I have just checked my emails

I ordered in june 2023 that Arctic MX-6 thermal paste. I was very upset with it. Luckily I could sell the rest of it for a very cheap price.
Igorslab Igor loves it, as far as I know. I consider htat thermal paste as unsuitable for application.

13€ put in the waste basket. The worst paste I ever used in past 20 years so far.
ARCTIC MX-6 (8 g) - Ultimate Performance Wärmeleitpaste für CPU, Konsolen, Grafikkarten, Laptops, sehr hohe Wärmeleitfähigkeit, Lange Haltbarkeit,
Verkauft von: ARCTIC GmbH (translates to ... sold by)



I replaced several times for many people thermal paste on laptops, apple computers and desktop computers. I doubt there was a handling issue. Maybe I got a bad product from amazon.de or a fake one.


edit: I removed that arctic thermal paste around two weeks later - what i remember. Visually nothing wrong. Just bad - bad - bad thermal paste. You know - you check how is the visual on the processor - you check how is the visual on the cooler - you check too much applied - you check not enough applied - and such. arctic has great 5€ p12 and p14 fans. The arctic freezer 33 is a bad product - same as that thermal paste. arctic mx4 was decent product many many years ago.