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

[AlwaysHope]

Interesting results with MX-5, meh... personally I'll stick with thermal grizzly range of TIMs.

 

[ir_cow]

I started to use MX-5 recently. I as well ran out of Thermal Grizzly. Its comparable to thermal grizzly kryonaut in temps. Not that I did any serious testing but it would be hard to say its "worse". Its a bit more gummy and is hard to remove, but not by much. Also 20g of Thermal Grizzly is half the amount of MX-5 if not more. I didn't realize how much Thermal Grizzly was ripping me off.

 

[lexluthermiester]

Ok, I'll admit that my take was overly cynical. All the arguing about AL must have me a bit edgy.

No worries. And I feel you! This latest Intel CPU release has brought out the fanboying something fierce... What I don't get is the lack of objectivity. Why is it that people forget the idea's of merit and scientific thinking where it comes to product releases? That can easily apply to this thread too. Are people really that emotionally invested in brand loyalty that they forget how to think without bias?

Its a bit more gummy and is hard to remove, but not by much.

This is true. MX-5 is very gooey and sticky, which makes cleaning it off surfaces more of a task, but IMHO, the results are worth it, especially for the price.

 

[ViperXTR]

Ill look at it then, but the high viscosity might be annoying

 

[lexluthermiester]

Ill look at it then, but the high viscosity might be annoying

It's not that annoying. Clean up is just more involved. Not a real problem unless you're going to be swapping CPU's frequently..

 

[Shrek]

I suspect that viscosity may be critical to longevity and so not something to be avoided.

Are people really that emotionally invested in brand loyalty that they forget how to think without bias?

This may be sewn into our genetics and part of our tribal mentality; a reflection of group loyalty.

 

[tabascosauz]

Tried to read some and yeah, mixed bag. Also considering MX-4, NT-H2 and Gelid GC Extreme, whichever is cheaper i guess

Some of this thread is for bare die applications (GPU, mobile CPU), so performance is going to look quite different for a lot of these pastes. MX-5 should be better than MX-4 there, but then again, I've never had a problem repasting my GPUs with MX-4.

On CPU however, I'd rule out MX-4 completely for modern or high end CPUs. The temp difference is kinda obvious, MX-4 sucks - tested it mainly against NT-H1 and NT-H2 on 4790K, 3700X, 5700G, 5900X. NT-H1 and NT-H2 seem to perform the same (maybe slight edge to NT-H2), I've tried SYY157 a few times and it's lost to the Noctuas by about 0.5-1C every time (also much harder to spread), it was hella cheap though. Havent tried MX-5 yet, have a big tube of NT-H1 and SYY157

 

[Space Lynx]

Some of this thread is for bare die applications (GPU, mobile CPU), so performance is going to look quite different for a lot of these pastes. MX-5 should be better than MX-4 there, but then again, I've never had a problem repasting my GPUs with MX-4.

On CPU however, I'd rule out MX-4 completely for modern or high end CPUs. The temp difference is kinda obvious, MX-4 sucks - tested it mainly against NT-H1 and NT-H2 on 4790K, 3700X, 5700G, 5900X. NT-H1 and NT-H2 seem to perform the same (maybe slight edge to NT-H2), I've tried SYY157 a few times and it's lost to the Noctuas by about 0.5-1C every time (also much harder to spread), it was hella cheap though. Havent tried MX-5 yet, have a big tube of NT-H1 and SYY157

any particular reason you dislike mx-4 so much? I found that it cooled 1-2 Celsius better than mx-5, mx-5 probably lasts longer though, i think both are rated for 8 years, but for some reason my mx-5 tube when i look at mx-5, i don't know, i just feel like it will last 20 years for some reason, but its no real reason for me to think that

 

[Shrek]

A conjecture (and only that): perhaps the high MX-5 viscosity means the layer is thicker and that explains the higher temperatures.

For me 8 years goes by so very fast and your feeling that MX-5 will last longer may be well founded.

 

[Falkentyne]

MX-5 works fine but if you want the best temps on ADL, or on a hot power guggling RTX 3090, you definitely want to consider Thermalright TFX for its higher viscosity and overall better temps (since Ampere has the 'temp throttle' thing where you lose 15 mhz at certain temp steps (the card manually adjusts its V/F curve down).

On both 10900k and RTX 1070 (laptop, modded), MX-5 was about 3-4C worse than Thermalright TFX. But TFX is freaking expensive. Only thing worse in price is Kryonaut Extreme and Coolermaster Cryofuze, both are which are complete highway robbery and bend over and get violated (and even worse, Cryofuze was tested on a laptop to perform WORSE than TFX).

 

[lexluthermiester]

MX-5 works fine but if you want the best temps on ADL, or on a hot power guggling RTX 3090, you definitely want to consider Thermalright TFX for its higher viscosity

This tells me you've never used MX-5 as I have used both of those and the MX-5 is about the same viscosity if not a bit more.

On both 10900k and RTX 1070 (laptop, modded), MX-5 was about 3-4C worse than Thermalright TFX.

I'm not saying it doesn't perform well, because it does. However, I'm not willing to accept 3-4c difference without testing to show such because that is not what I experienced when I tested it.

But TFX is freaking expensive.

No doubt. $20 for 2grams? Yeah stupid expensive. And when I tested it last, it performed well, but not $20 for 2g well.

 

[ViperXTR]

Noctua is more expensive here now so got the MX-5, will test it once it arrives.

 

[ViperXTR]

View attachment 224293
Noctua is more expensive here now so got the MX-5, will test it once it arrives.

Item has arrived, tested it, and its around 3-4C worse than Kryonaut so far, maybe thats to be expected? Also, does this have burn in/curing time?

 

[tabascosauz]

Item has arrived, tested it, and its around 3-4C worse than Kryonaut so far, maybe thats to be expected? Also, does this have burn in/curing time?

Big gap but doesn't seem surprising - iirc Tom's tested Kryonaut on par with NT-H1, I think I had a 2-3C gap between NT-H1 and MX-4 (the latter being worse). Seems from earlier in the thread that MX-5 doesn't improve significantly on MX-4 for IHS CPUs.

The only paste that has curing time is Arctic Silver I think. Might be a tiny change as the paste spreads in the few minutes after application but not much more

 

[lexluthermiester]

Item has arrived, tested it, and its around 3-4C worse than Kryonaut so far, maybe thats to be expected?

Weird.

Also, does this have burn in/curing time?

Nope, no curing or settling time.

Tom's tested Kryonaut on par with NT-H1

If this is true, I'm wondering what the differences are between my testing and ViperXTR's operating environment. My tests showed that MX-5 beat out NT-H1 by 3C idle and load and if Kryonaut is on par, then there must be a variance somewhere.

Still 3-4C difference is not the end of the world. Heck it's barely a blip on the temp radar..

 

[ViperXTR]

Im on a tropical environment, so my temps are warmer than most of you folks. With semi freshly applied Kryonaut, highest i got on a warm day running 1hr stress test was 82C (gets around 85C after few months, and Kryonaut becomes solid material like thing). For MX-5 i get around 83-85 on fresh application.
It only reaches these temps on extreme workloads like stress test, handbrake it does a bit but slightly lower. RPCS3 (PS3 emulator) is pretty much like a CPU stress test as well (on some heavy hitter games), it loads all your physical cores and extensive use of AVX2 (and it will use AVX512 if your CPU is capable)

 

[lexluthermiester]

Im on a tropical environment, so my temps are warmer than most of you folks. With semi freshly applied Kryonaut, highest i got on a warm day running 1hr stress test was 82C (gets around 85C after few months, and Kryonaut becomes solid material like thing). For MX-5 i get around 83-85 on fresh application.
It only reaches these temps on extreme workloads like stress test, handbrake it does a bit but slightly lower. RPCS3 (PS3 emulator) is pretty much like a CPU stress test as well (on some heavy hitter games), it loads all your physical cores and extensive use of AVX2 (and it will use AVX512 if your CPU is capable)

Fair enough. I don't think MX-5 is ever going to become a solid material over time like many TIM's do.

 

[Falkentyne]

This tells me you've never used MX-5 as I have used both of those and the MX-5 is about the same viscosity if not a bit more.

I'm not saying it doesn't perform well, because it does. However, I'm not willing to accept 3-4c difference without testing to show such because that is not what I experienced when I tested it.

No doubt. $20 for 2grams? Yeah stupid expensive. And when I tested it last, it performed well, but not $20 for 2g well.

I've never used MX-5? Oh really? Do you have any idea how much thermal paste I have?
You really shouldn't assume things about people that you know absolutely *nothing* about.

I have *40 grams* of MX-5. As well as 150 grams of homemade Galinstan, among many others.
In my own specific tests, MX-5 is about 3-4C worse than TFX.

 

[tabascosauz]

I have *40 grams* of MX-5. As well as 150 grams of homemade Galinstan, among many others.
In my own specific tests, MX-5 is about 3-4C worse than TFX.

What technique do you use for spreading TFX and SYY157? I'm thinking of revisiting my SYY results - might be leaving some performance on the table by not manually spreading

 

[lexluthermiester]

You really shouldn't assume things about people that you know absolutely *nothing* about.

It wasn't an assumption. You said the following...

MX-5 works fine but if you want the best temps on ADL, or on a hot power guggling RTX 3090, you definitely want to consider Thermalright TFX for its higher viscosity

...which is not the experience I had, my experience is that MX-5 is the more viscous TIM.

I'm not saying that TFX does not perform well because it does, but $14 for 2G? That price/performance ratio is pathetic. Thermalright TFX is not worth that. But Arctic's MX-5 at $12 for 8G? Hell yeah.

Someone said this;

Tom's tested Kryonaut on par with NT-H1

And I found a recent article at TH which compared TF4 to NT-H1 along with many others including MX-5. Tom's findings almost identically match mine in comparative temperature differential, NT-H1 VS MX-5. Courtesy of Toms Hardware, the images below show their results. One set of tests with a Noctua NT-D15 with low tension, another with the D15 and high tension and a third set with an Elite360 AIO;

As you can see, with both sets of tests using the D15, the results are the same, MX-5 beating NT-H1 by about 3C and matching NT-H2. With the Elite360, all of the premium TIMs were on par with each other indicating that the AIO was more than able to handle the thermal load and heat saturation seemed to not be a factor in that test. The conclusion one can easily see is that MX-5 a top performer and stands nose to nose with far more expensive brands.

What I did find for TFX is the following;

Unboxing and Review of Thermalright TFX Thermal Compound | UnbxTech

www.unbxtech.com

It ranks third and forth in the two other tests, trailing behind Noctua NT-H1 and NT-H2 (its review here).

Their NT-H2 review shows a similar result to Toms, thus giving their result credibility and merit. Their testing shows TFX falling behind both NT-H1 and NT-H2.

In both Toms tests and mine, MX-5 performs better than NT-H1. The conclusion is clear.

So if you're saying that TFX is doing better than MX-5, then there is something very hinky going on with your testing setup because the 3-4C better performance you claim TFX is getting does not match UnbxTech & Toms testing, nor mine. Add to this the fact that you claim TFX is more viscous than MX-5 and I have to raise an eyebrow, thus my above statement.

PS, Just in case anyone attempts to make the accusation, I did my testing in June. Toms did their testing last month. So no, I did not copy them. The fact that our comparative results match well only proves two things, 1. My testing methodology is spot on and valid, followed by 2. That MX-5 is one of the best performing TIMs on the market, handily beating out TIMs 3 or more times it's cost.

 

[Falkentyne]

What technique do you use for spreading TFX and SYY157? I'm thinking of revisiting my SYY results - might be leaving some performance on the table by not manually spreading

Thermalright TFX I simply yeet it Boil some water to "almost" boiling in a pot, remove the pot from stove, throw the TFX syringe in there for about 10 minutes, nozzle covered, drain the hot water, let it cool a few minutes, then spread the TFX with a spatula in even "long" strokes across the entire die. Works perfectly. If you have a hair dryer, that's a lot quicker and easier.
Hot air stations I know nothing about. I would assume they don't go as low as 80C on the lowest setting, so I'm not getting into that.

That being said, the TFX tube I have I didn't even need to heat the syringe when I applied a layer to my RTX 3090 FE after I re-did the thermal pads on that card. It was harder to spread than SYY without pre-heating but it worked fine. Some TFX tubes tend to be massively difficult. The 6.2g tubes seem to be a more consistent from apparent Amazon reviews than the 2g tubes are.

SYY-157 is much easier to spread as long as you don't get a dried out syringe. Just apply it and off you go. Use even, full strokes.
That Zezzio 14.8 w/mk thermal paste is identical 100% to SYY-157...it's the exact same paste.

FuzeIce Plus and Alseye T9+ Platinum both perform about the same as SYY-157, however they both seem to be a lot "stickier" than SYY-157. SYY-157 is sort of between FuzeIce Plus and TFX.
TFX performs the best (maybe about 1-2C better than SYY-157).

 

[Space Lynx]

Thermalright TFX I simply yeet it Boil some water to "almost" boiling in a pot, remove the pot from stove, throw the TFX syringe in there for about 10 minutes, nozzle covered, drain the hot water, let it cool a few minutes, then spread the TFX with a spatula in even "long" strokes across the entire die. Works perfectly. If you have a hair dryer, that's a lot quicker and easier.
Hot air stations I know nothing about. I would assume they don't go as low as 80C on the lowest setting, so I'm not getting into that.

That being said, the TFX tube I have I didn't even need to heat the syringe when I applied a layer to my RTX 3090 FE after I re-did the thermal pads on that card. It was harder to spread than SYY without pre-heating but it worked fine. Some TFX tubes tend to be massively difficult. The 6.2g tubes seem to be a more consistent from apparent Amazon reviews than the 2g tubes are.

SYY-157 is much easier to spread as long as you don't get a dried out syringe. Just apply it and off you go. Use even, full strokes.
That Zezzio 14.8 w/mk thermal paste is identical 100% to SYY-157...it's the exact same paste.

FuzeIce Plus and Alseye T9+ Platinum both perform about the same as SYY-157, however they both seem to be a lot "stickier" than SYY-157. SYY-157 is sort of between FuzeIce Plus and TFX.
TFX performs the best (maybe about 1-2C better than SYY-157).

for my next build I think I will give this thermalright tfx a try and stop being a cheap ass. lol

 

[Falkentyne]

It wasn't an assumption. You said the following...

...which is not the experience I had, my experience is that MX-5 is the more viscous TIM.

I'm not saying that TFX does not perform well because it does, but $14 for 2G? That price/performance ratio is pathetic. Thermalright TFX is not worth that. But Arctic's MX-5 at $12 for 8G? Hell yeah.

Someone said this;

And I found a recent article at TH which compared TF4 to NT-H1 along with many others including MX-5. Tom's findings almost identically match mine in comparative temperature differential, NT-H1 VS MX-5. Courtesy of Toms Hardware, the images below show their results. One set of tests with a Noctua NT-D15 with low tension, another with the D15 and high tension and a third set with an Elite360 AIO;
View attachment 224608View attachment 224609View attachment 224610
As you can see, with both sets of tests using the D15, the results are the same, MX-5 beating NT-H1 by about 3C and matching NT-H2. With the Elite360, all of the premium TIMs were on par with each other indicating that the AIO was more than able to handle the thermal load and heat saturation seemed to not be a factor in that test. The conclusion one can easily see is that MX-5 a top performer and stands nose to nose with far more expensive brands.

What I did find for TFX is the following;

Unboxing and Review of Thermalright TFX Thermal Compound | UnbxTech

www.unbxtech.com

Their NT-H2 review shows a similar result to Toms, thus giving their result credibility and merit. Their testing shows TFX falling behind both NT-H1 and NT-H2.

In both Toms tests and mine, MX-5 performs better than NT-H1. The conclusion is clear.

So if you're saying that TFX is doing better than MX-5, then there is something very hinky going on with your testing setup because the 3-4C better performance you claim TFX is getting does not match UnbxTech & Toms testing, nor mine. Add to this the fact that you claim TFX is more viscous than MX-5 and I have to raise an eyebrow, thus my above statement.

PS, Just in case anyone attempts to make the accusation, I did my testing in June. Toms did their testing last month. So no, I did not copy them. The fact that our comparative results match well only proves two things, 1. My testing methodology is spot on and valid, followed by 2. That MX-5 is one of the best performing TIMs on the market, handily beating out TIMs 3 or more times it's cost.

Tom's hardware is full of crap.

The Brazilian site (which I got banned from for just trying to REGISTER there) had TFX in the lead ahead of all the other pastes.
Laptop tests have also shown Thermalright TFX to be in the lead, with SYY-157, FuzeIce Plus and Alseye T9+ Platinum very close behind it.
MX-5 did terribly in these tests, but better than MX-4.

I'm not posting the notebookreview laptop thread again. I posted it before. I don't need to repeat myself. If you dont want to read it that's on you.

On desktops, there is far less of a difference between any of these pastes.

Even Mr. Fox on notebookreview has tested these pastes and on desktops with decent PSI pressure, most of these pastes are within 2C of each other. Then what matters is long term durability.

You are 100% free to bash my "imperfect" tests (I have OCD--i make 100% sure I test perfectly and consistently), but I tested MX-5 on a 1) 9900k, 2) 10900k, 3) GTX 1070 Laptop (TDP modded) and in each case, MX-5 was behind SYY-157 by a few C. (laptop was 74C for MX-5, 72C for SYY-157 with TFX at 69C on the GPU).

On the 9900k, TFX, SYY-157 and FuzeIce Plus seemed to be within 1C of each other. (I've seen TFX on my 10900k improve by 2C After a week however, which is equal to what some Japanese users found).

This is a free country. You're free to disagree with my tests. Other people have tested TFX and have been very happy with it. Including people on OCN who were having Kryonaut and Kryonaut Extreme both "dry out" on RTX 3090 video cards, and switched to TFX and had it a lot more consistent.

If your results have MX-5 ahead of TFX, then enjoy your MX-5. I'll enjoy my TFX.
Take care.

 

[lexluthermiester]

Tom's hardware is full of crap.

And you must think UnbxTech & I am too... Whatever.

Take care.

Yup..

 

[Shrek]

Anyone tried Cooler Master Ice Fusion V2?

Ice Fusion V2 High Thermal Conductivity Compound | Cooler Master