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

[Space Lynx]

Are you buying some? If so, do you want to do a set of performance runs and we can see what result we get as a team?


I'm going to get some and will post results here.

I already bought some. problem is I have to wait for December to get my RDNA3 card. and my cpu is a kf

so i am using mx-6, cause why not, perfect timing for my new build. but i am waiting until december for tests sadly

 

[maxfly]

Are you buying some? If so, do you want to do a set of performance runs and we can see what result we get as a team?


I'm going to get some and will post results here.

Let me know what you want to run and I'll do a comparison with the pk3/TR peerless assassin 120se that's on my test rig.

 

[lexluthermiester]

I already bought some. problem is I have to wait for December to get my RDNA3 card. and my cpu is a kf

so i am using mx-6, cause why not, perfect timing for my new build. but i am waiting until december for tests sadly

Fair enough. I'm ordering a tube and am going to do my standard run of tests. Will compare MX-5, NT-H1, CMCFV and the Dragon stuff against MX-6 in my T3500. Results might be interesting or perhaps boring. We shall see.

Let me know what you want to run and I'll do a comparison with the pk3/TR peerless assassin 120se that's on my test rig.

See right above. We both have MX-5 and we both have NT-H1 so we can compare those directly. My testing system is a Dell T3500 with a W3680 which I can softOC with ThrottleStop. Maybe we can do some comparisons with and without an OC? Do you still have the 9900KF stated in your specs?

 

[Space Lynx]

Fair enough. I'm ordering a tube and am going to do my standard run of tests. Will compare MX-5, NT-H1, CMCFV and the Dragon stuff against MX-6 in my T3500. Results might be interesting or perhaps boring. We shall see.


See right above. We both have MX-5 and we both have NT-H1 so we can compare those directly. My testing system is a Dell T3500 with a W3680 which I can softOC with ThrottleStop. Maybe we can do some comparisons with and without an OC? Do you still have the 9900KF stated in your specs?

I'm going with Freeagent, I think mx-6 is going to surprise us and get about 3 celsius win over the next best one. I hope so anyway, cause if its all 1 celsius or 2 max difference, you are right, im starting to get a little too bored of this game, lol

 

[lexluthermiester]

I'm going with Freeagent, I think mx-6 is going to surprise us and get about 3 celsius win over the next best one.

That would be kinda cool(no pun intended) given how well MX-5 performed.

EDIT:
Order placed. MX-6 is on the way. Question for everyone, besides Prime95, what other CPU-centric bentchmark/stresstest would you like to see? I'm open to suggestions but please keep in mind I'm a busy guy and really only have time for one additional test. If we all can arrive at a consensus, I'll give it a shot.

 

[phanbuey]

That would be kinda cool(no pun intended) given how well MX-5 performed.

EDIT:
Order placed. MX-6 is on the way. Question for everyone, besides Prime95, what other CPU-centric bentchmark/stresstest would you like to see? I'm open to suggestions but please keep in mind I'm a busy guy and really only have time for one additional test. If we all can arrive at a consensus, I'll give it a shot.

pump out for one... I hate having to repaste my gpu / cpu every few months because the temps start to creep. Maybe a 1-2 week temp comparison after initial pasting.

 

[lexluthermiester]

pump out for one... I hate having to repaste my gpu / cpu every few months because the temps start to creep. Maybe a 1-2 week temp comparison after initial pasting.

This has been enough of a discussed point, and I don't have enough time to do weeks long testing. I'm willing to look for it and take photos to show what each looks like post-testing. Honestly, I'm not expecting to find it, but I will look. The testing I did recently in the other TIM thread gave me chance to see the MX-5 after it had been on that CPU for more than a year. It looked fine. That PC, even though older, still gets frequent use as a gaming system for the younger kids in my family.

 

[Mussels]

pump out for one... I hate having to repaste my gpu / cpu every few months because the temps start to creep. Maybe a 1-2 week temp comparison after initial pasting.

Do you use the bare minimum amount of paste, or add extra?
I'm curious if my application method having 'extra' is why i suffer less from that?

(I do a perfectly flat thin spread using plastic, and then add a small blob to the center)

 

[Space Lynx]

(I do a perfectly flat thin spread using plastic, and then add a small blob to the center)

interesting, never heard of that one. lol

ever since 2500k days I just do a big pea size blob in middle and then tighten the screws slowly taking turns on w.e heatsink using. never had any issues myself.

i been considering trying a different method for this build though.

 

[phanbuey]

Do you use the bare minimum amount of paste, or add extra?
I'm curious if my application method having 'extra' is why i suffer less from that?

(I do a perfectly flat thin spread using plastic, and then add a small blob to the center)

I usually use an X on the CPU and a spread on the gpu die just to make sure i get everything covered. I've had good luck with tfx but other pastes not as much - even MX-5 pumped out on the 3080 after about 2 months.

 

[Space Lynx]

I usually use an X on the CPU and a spread on the gpu die just to make sure i get everything covered. I've had good luck with tfx but other pastes not as much - even MX-5 pumped out on the 3080 after about 2 months.

have you considered the LGA1700 bracket from TG or thermaltake? apparently LGA1700 socket really benefits from this contact frame bracket. that might stop that problem too.

i'm still considering it myself. not sure yet.

 

[Mussels]

I usually use an X on the CPU and a spread on the gpu die just to make sure i get everything covered. I've had good luck with tfx but other pastes not as much - even MX-5 pumped out on the 3080 after about 2 months.

The X is extremely hard to get an even line across the whole length, and might be part of the issue
If you have thinner/less paste at the top, gravity is going to slowly send it down and leave you with the pump out effect, aka gravity and heat slowly gave you an air pocket

you should definitely try other methods and see if it helps, any uneven spreading is going to make the pump out effect happen a lot faster

 

[phanbuey]

have you considered the LGA1700 bracket from TG or thermaltake? apparently LGA1700 socket really benefits from this contact frame bracket. that might stop that problem too.

i'm still considering it myself. not sure yet.

Actually I have one - works pretty well. the CPU with the X is actually the best performing one. The pump out happens on the GPU, which is sad because it's the biggest pain to replace...

 

[maxfly]

Fair enough. I'm ordering a tube and am going to do my standard run of tests. Will compare MX-5, NT-H1, CMCFV and the Dragon stuff against MX-6 in my T3500. Results might be interesting or perhaps boring. We shall see.


See right above. We both have MX-5 and we both have NT-H1 so we can compare those directly. My testing system is a Dell T3500 with a W3680 which I can softOC with ThrottleStop. Maybe we can do some comparisons with and without an OC? Do you still have the 9900KF stated in your specs?

I do have the 9900kf, but it's under water with hard tubing, so yeah...pita.
My test rig is an i7 8086k with Thermalright peerless assassin 120se (I may swap it out for noctua nh u12a for simplicity)and Asrock z390 phantom gaming 9 mb. I can clock the 8086k to whatever is needed.
I generally only run aida64 mostly because I've grown incredibly lazy in my old age but I'm open to whatever is agreed upon.

Regarding my most recent opportunity to observe pump out (after several years), using MX-4. I just swapped the rtx2070 (3 or 4 weeks ago)from my wced main rig to the air cooled test rig. I pulled the block and reinstalled the stock cooler. I didn't see any pump out on the wb after roughly, ugh, guesstimating 3- 3 1/2 years possibly longer. Coverage was uniform, the paste was evenly dispersed with only a very thin amount around the perimeter from the initial application. Paste was still spreadable and in great condition. It was easily removed with a wash cloth and 91% alcohol.

I have had the TR pa 120se installed on the test rig for roughly 4 weeks so I will also check that when I pull it (I'll take some pics as well, be warned I'm a terrible photographer). The pa 120se has prolimatech pk3 applied.
I apply a short thin line on one side of the die or ihs depending on the product. Then use a spatula to spread it as thinly and evenly as possible. I leave about 1/16th of an inch of the perimeter bare. Allowing for the inevitable excess to finish the job. Any small amount of tim left on the spatula, I apply to the middle of the block or heatsink. This generally results in a very thin layer with a small amount of tim being squeezed out around the edges. I've found that excess paste around the die/ihs edge is virtually impossible to avoid but it can be mitigated.

Edit- so sorry for the book.

 

[Pictus]

pump out for one... I hate having to repaste my gpu / cpu every few months because the temps start to creep. Maybe a 1-2 week temp comparison after initial pasting.

A PCM thermal paste will have longer lifespan for your GPU, it has long molecular chains, so it will not pump-out easily.
People have being using Honeywell PTM7950 with laptops/notebook with great results.
But I would try the new Halnziye HY-P16 because it has better proprieties.
TIP, because the Halnziye HY-P16 has solvent, apply it a night and in the morning it will be dry and ready to mount
the cooler, apply a paper thin film over the GPU die.
The PCM is some sort of polymer that melts at +-+45ºC and do some "magic".

 

[Count von Schwalbe]

@lexluthermiester and @maxfly maybe try Cinebench? It seems to be accepted as the "worst-case real-world" situation for temps.

 

[P4-630]

If it can beat Syy or Kryonaut I'll get some.

The Kryonaut Extreme pink bubblegum then.

 

[Space Lynx]

Actually I have one - works pretty well. the CPU with the X is actually the best performing one. The pump out happens on the GPU, which is sad because it's the biggest pain to replace...

would you recommend I get the lga 1700 contact frame? i been on the fence about it for awhile. i feel like its main benefit is AIO's, with air coolers being a secondary thought. not sure though.

 

[phanbuey]

would you recommend I get the lga 1700 contact frame? i been on the fence about it for awhile. i feel like its main benefit is AIO's, with air coolers being a secondary thought. not sure though.

check how the paste pattern is on the chip and cooler. If you're seeing uneven spread like I was then it's worth it.

if it looks pretty even then it won't do all that much:

 

[freeagent]

have you considered the LGA1700 bracket from TG or thermaltake

You mean Thermalright! They are not in the same lowly league as Thermaltake...

 

[Space Lynx]

You mean Thermalright! They are not in the same lowly league as Thermaltake...

oops, yes that is what i meant. same thing happens to me with seasonic and seagate all the time, lmao

check how the paste pattern is on the chip and cooler. If you're seeing uneven spread like I was then it's worth it.

View attachment 268546

if it looks pretty even then it won't do all that much:
View attachment 268547

i am going to keep an eye on this. its annoying we have to do this.

 

[phanbuey]

IMO Thermalright is bascially noctua with no marketing dept. All of their products are ridiculously good but no one knows about them.

Actually my favorite paste right now is thermalright tfx

 

[Nike_486DX]

A PCM thermal paste will have longer lifespan for your GPU, it has long molecular chains, so it will not pump-out easily.
People have being using Honeywell PTM7950 with laptops/notebook with great results.
But I would try the new Halnziye HY-P16 because it has better proprieties.
TIP, because the Halnziye HY-P16 has solvent, apply it a night and in the morning it will be dry and ready to mount
the cooler, apply a paper thin film over the GPU die.
The PCM is some sort of polymer that melts at +-+45ºC and do some "magic".

to avoid all that nonsense you can just use liquid metal, its a superior thing if you know how to use it. 1 year on a rx 580, still same temps as day 1

 

[freeagent]

IMO Thermalright is bascially noctua with no marketing dept.

Actually, it's the other way around, Thermalright was first

I remember the first Noctua review that I saw and thought they totally ripped off Thermalright

I tried their D14 and thought it was overhyped.

But that's what happens when you don't put a patent on anything, you get robbed.

 

[phanbuey]

I just did a budget build with a peerless assassin 120 for $40 and the included stock thermal right paste (TF8 or 9)... thing was keeping a HOT 10850K easily under 80C in Cinebench- the guy i built it for is happy as a clam. Was really worried about the heat of that chip, but no issue.