[EOL] Arctic MX-5 is here!!Tests incoming! Completed. Now its MX-6 testing time!

[lexluthermiester]

Any great advantage applying it to both surfaces?

Yes. I've found over the years that applying a very thin layer to both surfaces renders the best results over any other method of application.

It's atrocious on laptops though. Too thin. Same problem as "new" KPx (old KPx was awesome on laptops).

Not true. I've already applied MX-5 to several laptops, including 2 of my own, and the results were equally impressive.

Cryorig instructions recommend tinting the cold plate, so I'm inclined to say "probably".

"tinting" What does that mean in the context of applying a TIM to a surface?

 

[Shrek]

For me, the problem with all this testing is that I am tempted to draw the conclusion 'it doesn't really matter what one uses, the difference is minor'

I know that isn't true, but that is the impression I get.

 

[Borc]

Yes. I've found over the years that applying a very thin layer to both surfaces renders the best results over any other method of application.


Not true. I've already applied MX-5 to several laptops, including 2 of my own, and the results were equally as impressive.


"tinting" What does that mean in the context of applying a TIM to a surface?

Impressive to what? What paste did you compare? Performance depends on the contact anyways, MX5 could work on a laptop with a good contact, however the best laptop pastes should beat MX5 convincingly. edit, based on your posting in #520 you didn't test any of the top laptop pastes. The only meaningful comparison is Noctua NT-H1 which is outdated, Noctua NT-H2 would be more interesting.

 

[Space Lynx]

For me, the problem with all this testing is that I am tempted to draw the conclusion 'it doesn't really matter what one uses, the difference is minor'

I know that isn't true, but that is the impression I get.

I'm pretty sure it is true actually. lol which is why I decided to stop caring and just stick with my mx-5. if you can get official mx-4 direct from arctic store though, i think mx-4 is worth it, same price and 2 celsius better. at least for me. but since I have mx-5 meh.

 

[Shrek]

OK, if MX-4 is 'better' than MX-5... what has MX-5 got to offer?

Top Thermal Paste for All CPU Coolers | ARCTIC
"The MX-5 impresses with ideal thermal properties, guaranteeing reliable heat dissipation over long periods of time."

perhaps it is longevity

 

[Space Lynx]

OK, if MX-4 is 'better' than MX-5... what has MX-5 got to offer?

Top Thermal Paste for All CPU Coolers | ARCTIC
"The MX-5 impresses with ideal thermal properties, guaranteeing reliable heat dissipation over long periods of time."

perhaps it is longevity

mx-4 also advertises 8 years though on the packaging. that's the problem, i think both are the same mostly, mx-5 maybe was just some cost saving measures? i dunno maybe mx-4 was costing them to much make so they switched to a more sustainable paste for the long term. prob not, just an idea

 

[toilet pepper]

I just used the Thermalright TFX on my RTX 3080 and I can say it improved the core temp by 3-4C compared to MX4. The application was a pain though. It had the consistency of sticky rice all squeezed together in a tube. The spatula had more paste than the die.

 

[80-watt Hamster]

"tinting" What does that mean in the context of applying a TIM to a surface?

Exactly what you did: a very thin, bordering on translucent layer. Although it's possible I'm mis-remembering the exact term used.

 

[lexluthermiester]

Impressive to what? What paste did you compare? Performance depends on the contact anyways, MX5 could work on a laptop with a good contact, however the best laptop pastes should beat MX5 convincingly. edit, based on your posting in #520 you didn't test any of the top laptop pastes. The only meaningful comparison is Noctua NT-H1 which is outdated, Noctua NT-H2 would be more interesting.

If you don't like what I posted, buy your own samples of TIMs, do your own tests and post your own results.

 

[Shrek]

Exactly what you did: a very thin, bordering on translucent layer. Although it's possible I'm mis-remembering the exact term used.

Makes one wonder if a surfactant is involved.

NOW I am overthinking things...

 

[lexluthermiester]

Makes one wonder if a surfactant is involved.

That's possible.

NOW I am overthinking things...

Maybe just a bit...

I was perfectly happy with the TIMs I had previously. Now that I've seen MX-5 and NT-H1 in action, the comparison to how far the science of TIM formulation has come is impressive. MX-4 was good and still is. But I have run tests like this before and against the DRG102 as a baseline. MX-4 did not perform as well as the NT-H1 and MX-5 when comparing relative margins.

I think MX-5 will now be my go-to TIM going forward.

 

[Shrek]

If you don't like what I posted, buy your own samples of TIMs, do your own tests and post your own results.

Yeah... I'm sore that you didn't include diaper cream... sorry, couldn't resist the pun

In reality let me thank you once again for taking the time to test all those pastes (I'm not allowed to call them greases)

 

[Space Lynx]

I just used the Thermalright TFX on my RTX 3080 and I can say it improved the core temp by 3-4C compared to MX4. The application was a pain though. It had the consistency of sticky rice all squeezed together in a tube. The spatula had more paste than the die.


View attachment 202594

that tfx stuff sounds pretty great. if mx-5 made me lose 2 celsius... over mx-4, and that made you gain 3-4 celsius over mx-4 i could be looking at a net gain of 5-6 celsius over mx-5... that's no joke.

 

[lexluthermiester]

In reality let me thank you once again for taking the time to test

You're welcome. I actually do this every few years just to keep up on the best TIMs in the industry. Wasn't going to do another round of testing until next year, but the debate in this thread intrigued me. This is the first time I've publicly posted my efforts and results.

(I'm not allowed to call them greases)

That's why you see me calling them " TIM ". That's what they are, Thermal Interface Materials.

 

[Shrek]

Now that I have 120g of GD900 I don't think I'll be changing anytime soon.

 

[lexluthermiester]

For those who might have complaints, please see edit at the bottom of post #520.

 

[toilet pepper]

that tfx stuff sounds pretty great. if mx-5 made me lose 2 celsius... over mx-4, and that made you gain 3-4 celsius over mx-4 i could be looking at a net gain of 5-6 celsius over mx-5... that's no joke.

I dont suggest it though. I just repasted mine as it is very easy to do since my card is already deshrouded and the disassembly is quite easy. It felt like MX4 wasnt up to task with high consistent temps like mining.

 

[Solid State Soul ( SSS )]

nothing in a PC is gunna run at 120C

GDDR6X: allow me to introduce myself

 

[lexluthermiester]

I dont suggest it though. I just repasted mine as it is very easy to do since my card is already deshrouded and the disassembly is quite easy. It felt like MX4 wasnt up to task with high consistent temps like mining.

I wonder how MX-5 would do in your situation...

GDDR6X: allow me to introduce myself

No VRAM runs that hot and stays stable.

 

[Falkentyne]

Yes. I've found over the years that applying a very thin layer to both surfaces renders the best results over any other method of application.


Not true. I've already applied MX-5 to several laptops, including 2 of my own, and the results were equally as impressive.


"tinting" What does that mean in the context of applying a TIM to a surface?

LGA or BGA laptop?
Because BGA Was bad according to rooter (Not talking about video card application).

TechnologyGuide

Thank you for visiting the TechnologyGuide network. Unfortunately, these forums are no longer active. We extend a heartfelt thank you to the entire community for their steadfast support—it is really you, our readers, that drove

forum.notebookreview.com

Pretty sure the heatsink fit is everything here. Some of these dellienware tripod heatsinks are worse than atrocious.

I do agree with applying to both surfaces. I got the best results with TFX on my 3090 FE (full spread and applying a second layer directly on the heatsink itself)+ Gelid Extreme 1.5mm pads.
I believe people talked about 'tinning' the heatsink many years ago but um...I think that was applying a micro residue layer (like as little as possible)? but that definitely doesn't work with TFX. Gotta do it thick. For MX-5 and Kryonaut Extreme I just do a normal thin layer on the heatsink and a thicker layer on the GPU.

 

[Shrek]

No VRAM runs that hot and stays stable.

I have always wondered about such heat, is it the silicone that fails or the solder attaching the chip?

 

[Falkentyne]

I just used the Thermalright TFX on my RTX 3080 and I can say it improved the core temp by 3-4C compared to MX4. The application was a pain though. It had the consistency of sticky rice all squeezed together in a tube. The spatula had more paste than the die.


View attachment 202594

Spreading TFX Was the most annoying thing in the world.

I couldn't even do it without losing the half (not all) of the freaking tube to the spatula (I was using ZF-EX 2g tubes though, which is the same paste, the TFX itself I did the X pattern + 4 dots in quadrants), until I found this post, which explained how to spread TFX easier. You basically have to yeet it right away and then apply an even stroke to it.

TechnologyGuide

Thank you for visiting the TechnologyGuide network. Unfortunately, these forums are no longer active. We extend a heartfelt thank you to the entire community for their steadfast support—it is really you, our readers, that drove

forum.notebookreview.com

When I used that exact method (except on a much larger 3090 die rather than smaller laptop chips), I got a nice complete thick spread.
End result: 3 weeks later, core to core hotspot temp deltas are almost the same as day 1, sometimes varying by 1C depending on when I test it or how the chip cools earlier (Delta ranges from 10.0C to 11.6C depending on power limit and what I'm running. Had 9.9C at 530W last night, 10.1C this morning. Then 10.6C two days ago, whatever...).

That being said, TFX wet vs dry consistency can vary wildly.

I had several 2g Thermagic ZF-EX tubes (still three unopened ones left, the current opened one is perfect). One of them was the consistency of rubber. Almost like there was no liquid left lol. Only way to apply that was X + dots pattern. The very next one was perfect consistency (still super thick with most sticking to the spatula). The one I spread fully manually on my 3090 FE was the perfect one. Same thing with the TFX. I've had four total 6.2g tubes. Two: absolutely perfect. One tube: a bit "too" perfect (actually seemed a slightly bit wetter than normal, even wetter than the perfect ZF-EX). Fourth tube: dry as living hell (still came out of the syringe fine though). No way you're spreading the "super dry" ones. The method I linked screenshots above works fine with the 'wet' or 'normal' consistency tubes. The super dry ones--no chance in hell.

I wasted way too much expensive TFX and ZF-EX messing around with my 3090 FE and different thermal pads until I found the Gelid Extreme 1.5mm soft pads.

I have always wondered about such heat, is it the silicone that fails or the solder attaching the chip?

RTX "Space Invaders" (Turing cards) happened because the solder got loose on the BGA chip closest to the PCIE slot due to crappy soldering and very high heat (an expert solderer from Germany found this out by testing).

 

[Shrek]

Fourth tube: dry as living hell

TIMs that are drying out in storage scare me, unless they are designed to dry out

I have some that is perhaps 30 years old and it seems the exact same as it was when new.

Silicone oil should be good to 500°C; Zinc Oxide much higher.

 

[Falkentyne]

TIMs that are drying out in storage scare me, unless they are designed to dry out

I have some that is perhaps 30 years old and it seems the exact same as it was when new.

Probably a batch problem. Package was sealed tight and completely locked down.
My "super wet" TFX sample is just sitting in a ziploc bag (not airtight) with the lid closed and hasn't dried one bit. Wonder if the storage temp or location has something to do with the iffy tubes...
Because I purchased a bunch of ZF-EX from the same supplier in China, all at the same time. One tube was really dry, it would barely even stick to the die when pushing the plunger! The next tube was absolutely perfect (for TFX of course, still going to stick to the spatula unless you do the spread method I posted above). No idea how it could vary that much.

BTW Andy you should try SYY-157. Probably the most pleasing paste I've ever used.
Thicker than Kryonaut Extreme, MUCH easier to spread than TFX, performs the same as TFX and doesn't cost your liver, kidney or firstborn son to purchase.

 

[freeagent]

I love the performance of TFX. I dislike how long you have to futz with it when you are applying it. And I also dislike the price. But it really does work well.. best I’ve used so far.. I’m going to try SYY-157 if I can find it.