Overclock.net banner
Cancel
Create thread New Forums

Arctic MX-6 disappointed

Tags
arctic mx-6 disappointed
1 reading
Jump to Latest
21 - 38 of 38 Posts
Discussion starter · #21 · (Edited)
leonman44 said:
MX6 is incredibly good thermal paste especially for the price. Huge pro that it’s thick as cement so pump out effect gets minimized , con is that you need like 10+ minutes to spread it with a spatula because of how hard it is.

From my testing on my 4090 it’s only hardly 1c hotter than the god tier KPX paste that may be one of the very best but after only 2-3 months it needs to be changed as it degrades too fast. So all in all MX6 is a wonderful paste that should also last long term.

So there are only 2 things that may be happening to you:

-the factory pasting was perfect so you can’t get anything noticeably better than that.

OR

-factory pasting job is bad but yours was bad too. 😅


I want to believe that it’s the first one.
Click to expand...
acoustic said:
I've also never seen KPx "degrade" in 1-3 months. After a year or so, temps will go up, but not by a ton.. but that's standard for any paste outside of the phase-change materials. There is a rumor that newer batches of KPx are not lasting as long, but I haven't seen that verified with any real science, just talk.

I used KPx exclusively for years -- I've only just started moving over to PTM7958-SP instead. Currently have it on the 4090 + 14900K. It's the future, imo.
Click to expand...
Great recommendations guys!
Well hopefully this MX-6 lasts me a year on the GPU and CPU.
Next I think I am going to try the Honeywell PTM 7958 paste, then KPX paste.
I am going to screen them out, by thermal performance and if they can last atleast a year on CPU/GPU.
 
#22 ·
BlueBase said:
Great recommendations guys!
Well hopefully this MX-6 lasts me a year on the GPU and CPU.
Next I think I am going to try the Honeywell PTM 7958 paste, then KPX paste.
I am going to screen them out, by thermal performance and if they can last atleast a year on CPU/GPU.
Click to expand...
Kpx is the best paste ive used on a cpu. Arctic mx6 was def worse temps. I have mx6 on other computer with air cooler and temps were same as when first applied a year ago.
i found kpx after 6 months temps were coming up. But Just get whatever is cheapest and run with it. On a gpu die paste seems to work just fine. I had an old 1070 msi card that was never opened up for 6-7 years and new paste dropped temps 10c. I wonder what they used back then. It lasted a long time.
 
#23 ·
I find things that are mainstream such as TIM usually are not that good.
 
  • Rep+
  • Haha
Reactions: Mergatroid and austin86
Save
Reply Quote Rep+ Rep+
#27 ·
Meh, just solder the heatsink to the IHS directly with a blowtorch and some copper pipe solder right?
 
  • Rep+
  • Haha
Reactions: Owterspace and acoustic
Save
Reply Quote Rep+ Rep+
#28 ·
Would love to see the spread of the thermal paste. I've found the biggest improvement in temps isn't the paste but how flush the cooler sits on the chip. If you're able to lapp the cooler you'll see a massive drop in temps.
 
#29 ·
Thermal paste alone is not going to make a huge difference unless you use liquid metal. If the paste they used was half decent, and applied well, repasting isn't going to do much.
If you look to thermal paste reviews and comparisons, you're looking at maybe 3 to 5c in difference (if that). The point of repasting a video card that's not 10 years old is to remove poorly applied paste, replace it correctly and remount the cooler. Just using a different paste on a cooler that was already done well is hardly going to make any difference.
 
Save
Reply Quote Rep+ Rep+
#30 ·
If you get it right, SYY-157 is probably one of the better ones that I have used. I buy it all the time because I swap CPU's so often. My temps are usually pretty good compared to others.
 
Save
Reply Quote Rep+ Rep+
#31 · (Edited)
austin86 said:
I have not used mx-6 but some TIM's have a burn in period where it takes time to preform its best, that being sad I like Phobya's nano grease extreme, it has a thermal conductivity of 16W/mK!
Click to expand...
High thermal conductivity doesn't automatically mean that it performs equally well.

For instance:
MX-5 - Thermal Conductivity: 6.0 W/mK
CoolLaboratory Liquid Pro - Thermal Conductivity: 79.0 W/mK

The difference in performance is quite small.

Owari said:
Just use a Kryosheet and be done with paste.
Click to expand...
Kinda pricy almost 6$ per cm²
 
#32 ·
eNdEmiOn said:
High thermal conductivity doesn't automatically mean that it performs equally well.

For instance:
MX-5 - Thermal Conductivity: 6.0 W/mK
CoolLaboratory Liquid Pro - Thermal Conductivity: 79.0 W/mK

The difference in performance is quite small.


Kinda pricy almost 6$ per cm²
Click to expand...
TBH most people don't know how to apply and cure LM so most tests of it are misleading. But yeah Thermal Conductivity is not 1:1 but if its higher it tends to preform better, how much depends on a lot of other factors too like the gap between the hint sink and chip, how its applied, how much mounting presser is use and so on.
 
#33 ·
It took quite a while for the LM (Conductonaut) to properly set and cure on my 4080 Super but it's great now. Never seen such a small hotspot delta with paste and overall about 8-9c lower than MX6/PK-3 on the core and 12-13c lower hotspot.

I pretty much only see LM as useful on direct die applications. GPU's, consoles, laptops, delidded CPU's.. it makes way less sense and difference on a heat spreader CPU. And working with LM and unplated copper is not great, takes a LOT of time and remounts to get it proper soaked in and especially on heatpipe direct touch coolers it can do some real damage to the heatpipes. Don't do it.
 
  • Rep+
Reactions: Mergatroid
Save
Reply Quote Rep+ Rep+
#34 ·
Imprezzion said:
It took quite a while for the LM (Conductonaut) to properly set and cure on my 4080 Super but it's great now. Never seen such a small hotspot delta with paste and overall about 8-9c lower than MX6/PK-3 on the core and 12-13c lower hotspot.

I pretty much only see LM as useful on direct die applications. GPU's, consoles, laptops, delidded CPU's.. it makes way less sense and difference on a heat spreader CPU. And working with LM and unplated copper is not great, takes a LOT of time and remounts to get it proper soaked in and especially on heatpipe direct touch coolers it can do some real damage to the heatpipes. Don't do it.
Click to expand...
Good find. That should work better than the the one he has now.
 
Save
Reply Quote Rep+ Rep+
Discussion starter · #35 ·
leonman44 said:
MX6 is incredibly good thermal paste especially for the price. Huge pro that it’s thick as cement so pump out effect gets minimized , con is that you need like 10+ minutes to spread it with a spatula because of how hard it is.

From my testing on my 4090 it’s only hardly 1c hotter than the god tier KPX paste that may be one of the very best but after only 2-3 months it needs to be changed as it degrades too fast. So all in all MX6 is a wonderful paste that should also last long term.

So there are only 2 things that may be happening to you:

-the factory pasting was perfect so you can’t get anything noticeably better than that.

OR

-factory pasting job is bad but yours was bad too. 😅


I want to believe that it’s the first one.
Click to expand...
Mergatroid said:
Good find. That should work better than the the one he has now.
Click to expand...

Would a 10 minute temperature test be long enough?
I am a noob and unsure.
I used a 10 minute test on Idle and Loaded
 
#36 ·
BlueBase said:
Would a 10 minute temperature test be long enough?
I am a noob and unsure.
I used a 10 minute test on Idle and Loaded
Click to expand...
For aircooling no , I would leave it stress testing something really heavy around an hour.
Also the paste will reform after some on/off cycles , you may have different resulted after a week , month etc…

You don’t have to go crazy with it just check your temps and hotspot to see if it’s healthy , you want the hotspot to be 15c maximum warmer than your core temp under full load ~500watts. The better the pasting and the mounting contact the lower the hotspot delta will be.
 
Save
Reply Quote Rep+ Rep+
#37 ·
BlueBase said:
Would a 10 minute temperature test be long enough?
I am a noob and unsure.
I used a 10 minute test on Idle and Loaded
Click to expand...
You need to test long enough to saturate the metal in the cooler, or the water in a water cooler. I would expect about 20 minutes for an air cooler, and maybe a half hour for an AIO. I usually run my tests for a half hour.
 
Save
Reply Quote Rep+ Rep+
21 - 38 of 38 Posts
About this Discussion
37 Replies
20 Participants
Last post:
eNdEmiOn
Overclock.net
posts
28M
members
562K
Since
2004
The premier forum for overclocking experts and enthusiasts. Discuss hardware optimization, custom builds, benchmarking, cooling solutions, and pushing the boundaries of computing performance. From beginner guides to extreme overclocking, join our technical community to master system tuning.
Full Forum Listing
Explore Our Forums
Intel CPUs Hardware News NVIDIA PC Gaming Water Cooling