Thermal Grizzly Kryonaut Extreme thermal grease has thermal conductivity of 14.2 W/mk

Discussion in 'Frontpage news' started by Hilbert Hagedoorn, Dec 3, 2020.

  1. Krizby

    Krizby Ancient Guru

    Messages:
    3,067
    Likes Received:
    1,744
    GPU:
    Asus RTX 4090 TUF
    I used the Thermalright TF-X on my new Acer Triton 500 and it should be the best TIM ever for laptop. Being viscous AF give TFX the advantage of never pumping out.
    1st repaste on my laptop with the old Kryonaut and it pumped out after a month, 2nd repaste with TFX and the result has been fantastic for over 6 months already.
    Being viscous AF also means the TFX is impossible to spread, I think some reviewers (like Luumi) judged TFX wrongly by not heating the paste up with a hair dryer before mounting the cooler, leading to a bigger gap between CPU and heatsink because of how viscous TFX is. That or TFX is not supposed to perform that well on CPU because it cannot get as thinly spread as other paste and only perform its best on GPU and Laptop.
     
  2. Mufflore

    Mufflore Ancient Guru

    Messages:
    14,691
    Likes Received:
    2,673
    GPU:
    Aorus 3090 Xtreme
    From what you say, TF-X might be just what I need for my 6700K delid.
    When paste is first applied directly to the core, all core temps are within a few degrees C of each other.
    Just weeks later its around 6C difference and after 6 months is 11C, it knocks 100 to 200MHz off the overclock. Even with Gelid Extreme GC3, takes a bit longer but it still happens.
    Depending which way round I put the lid determines which cores are affected.
    The inside surface of the lid must be none flat causing the pump effect.

    The huge amount of pressure applied to the lid on a very small area should squeeze a very thick paste out to as thin as it can get once heated a few times, no need to spread it before putting the lid on.
    What do you think?

    On the other hand I have a direct die water block on its way here finally, only 2.5 years late lol. (a Kickstarter that had a few issues)
    But I'm not too bothered about putting this machine on water now because I just bought a 10700K.
    It would be nice to fix the delid instead so I can pass it on in the family at max speeeed.
     
  3. insp1re2600

    insp1re2600 Ancient Guru

    Messages:
    2,315
    Likes Received:
    1,091
    GPU:
    4080 FE
    I'd certainly use liquid metal on a 6700k delid, best thing I ever did when I had a 7700k a few years ago. The drop was excellent and managed to get up to 5.2ghz
     
  4. Mufflore

    Mufflore Ancient Guru

    Messages:
    14,691
    Likes Received:
    2,673
    GPU:
    Aorus 3090 Xtreme
    I read it deteriorates after a few years and needs to be re-applied.
    And there is slight damage to the surface each time which is probably why it needs to be re-applied, it reacts with certain metals, some very slowly (ie copper).
     

  5. insp1re2600

    insp1re2600 Ancient Guru

    Messages:
    2,315
    Likes Received:
    1,091
    GPU:
    4080 FE
    Didn't have this issue, but then again cant say I kept it for a few years.
     
  6. Krizby

    Krizby Ancient Guru

    Messages:
    3,067
    Likes Received:
    1,744
    GPU:
    Asus RTX 4090 TUF
    Yeah you should apply Conductonaut between the die and the lid, normal TIM between the lid and the cooler for best result.
    TG Conductonaut eats away aluminum and react slowly with copper, however Intel IHS is nickel plated and safe to use with Conductonaut. You can easily shave 10C off CPU temp with Conductonaut.
    My delidded 8700K has been running well with Conductonaut for 3 years already.

    You can put some nail polish on the SMDs beside the die to void Conductonaut dripping out and short circuit them. It's also not recommended to use Conductonaut outside of the lid for the reason that it might drip down. Though I use conductonaut on my GPU and it lowers the temp by 4C compare to Kryonaut (horizontal GPU placement is safe)
     
  7. Silva

    Silva Ancient Guru

    Messages:
    2,048
    Likes Received:
    1,196
    GPU:
    Asus Dual RX580 O4G
    I did a quick search to see what was around compared to what I used.
    Thermal conductivity of liquid metal is astonishing, but that CoolerMaster paste is a great deal price/performance wise! Might not be on par at sub zero temps but we don't need to!
     
    -Tj- likes this.
  8. chainy

    chainy Active Member

    Messages:
    68
    Likes Received:
    17
    GPU:
    970gtx
    Actually when u delidd u do need to apply liquid metal, but it's just a very small amount so you can prolly use that 1g of liquid metal as many times as your 4 gram. For 11$ it's the best stuff for dropping temps..
     
  9. c0rrupt

    c0rrupt Guest

    Messages:
    4
    Likes Received:
    1
    GPU:
    Red Devil RX 480
    I wonder if this batch of Kryonaut will damage the cpu like the other one does. They made a bad batch and sold it to customers because they "couldnt" track it. frack them
     
  10. Silva

    Silva Ancient Guru

    Messages:
    2,048
    Likes Received:
    1,196
    GPU:
    Asus Dual RX580 O4G
    Deliding is a very small case scenario, I'd never do that to a CPU. If you want to OC everything counts but I just want to under-volt and live happy with what I got.
     

  11. Falkentyne

    Falkentyne Master Guru

    Messages:
    544
    Likes Received:
    79
    GPU:
    Sapphire HD 7970 Ghz Ed.
    Not really supposed to spread TFX.
    You should apply it in a large edge to edge diagonal X pattern on a GPU, large CPU die, or IHS top.
    Kryonaut Extreme's packaging also mentions this same application pattern.

    Try not to apply too little just in case you wind up with a few areas not covered. That can also be improved by applying a small drop in each "leftover" quadrant, after the large "X" is completed.

    For rectangular CPU dies, just apply a straight line through the long middle to each edge.

    -----

    So far no one here seems to know if Kryonaut Extreme lasts a long time.
    One thing is clear is TFX / ZF-EX lasts for ages, like the old Arctic Ceramique paste.
     
  12. Falkentyne

    Falkentyne Master Guru

    Messages:
    544
    Likes Received:
    79
    GPU:
    Sapphire HD 7970 Ghz Ed.
    The deterioration issue can be fixed by one of two ways:

    1) On IHS delids, apply LM on the underside of the IHS, spread it like normal, put IHS into an oven at 100C, leave it for 2 hours. (100C=212F).
    Then let it cool, then clean the LM off the heatsink with light wipes, leave the silver stain (Only clean the liquid part). Do not use alcohol, just a lint free cloth.
    Then apply a new layer of LM on top of the silver stain, spread it, then mount the IHS.

    This also works for laptop heatsinks also. I don't know how this would work for a desktop heatsink or heat block, but definitely don't do this if any liquid or wires are involved anywhere.

    2) This can work for IHS, heatsinks, heatblocks, anything that isn't aluminum, obviously, and this is what I use myself.

    a) Get some 1500 to 2000 grit sandpaper.
    b) sand the underside of the IHS by wiping it with your fingers (do not use a sanding kit or glass--you are not trying to flatten it!). Your goal is to make the surface ROUGHER with micro-scratches, not smoother! This GREATLY improves LM's wetting ability and ability to spread!
    c) clean the entire sanded surface of all particles with alcohol
    4) Apply LM to the buffed surface. Use a lint free swab--not a Q-tip. I recommend THESE personally--Falkentyne approved.

    https://www.amazon.com/gp/product/B085XXXPMM/

    5) Now--the IMPORTANT part. Spend about 10 boring minutes constantly spreading the LM around the entire surface it is being applied to. Do this for both the underside of the IHS (for delids and relids) and if you are applying LM to a copper or nickel plated heatsink, or a heat block/pump, do the EXACT Same thing to that also. Just keep spreading it around over and over. Do NOT apply downwards pressure--you don't want micro particles to be scraped out from the pressure!

    What you will notice is, eventually, the LM will start becoming 'thicker', almost like it's acting like a thin paste. This means some of the gallium is properly being absorbed into the micro-fissures in your surface. This was the entire point of the sanding--to create surfaces where the LM could bind to, and the gallium could be absorbed (like the baked silver stain).

    6) Apply a new layer of LM on top of the old layer. Spread it around and then mount your IHS/heatblock/cooler.

    This should give you outstanding longevity. My laptop application has remained stable for far over a year, after doing this method, with core temp deltas on my 7820HK just 1C after all this time.
     
  13. Mufflore

    Mufflore Ancient Guru

    Messages:
    14,691
    Likes Received:
    2,673
    GPU:
    Aorus 3090 Xtreme
    Do you have evidence this doesnt cause worse damage to the surface and deeper?
     
  14. Krizby

    Krizby Ancient Guru

    Messages:
    3,067
    Likes Received:
    1,744
    GPU:
    Asus RTX 4090 TUF
    I used the Noctua's 5 dots method on the 2070 Super MQ die and a straight line on the 10875H, heated up the TIM with the hair dryer and mount the heatsink, heated up the heatsink some more and tightern the screws until they come to complete stop. This method yielded much better result than the old Kryonaut.

    So far the TFX on my laptop doesn't show any sign of pumping out after 6 months, with the CPU constantly in the high 80C and GPU in the low 70C, normally at these temps the old Kryonaut would have pumped out after a month.

    I don't think the Kryonaut Extreme will last that much longer than the old Kryonaut, with the Extreme only slightly more viscous than the old.

    This thermal result is probably the best for any thin and light 15in laptop :D, ambient temp at 27C
    [​IMG]
     
    Last edited: Dec 9, 2020
    Falkentyne likes this.
  15. Falkentyne

    Falkentyne Master Guru

    Messages:
    544
    Likes Received:
    79
    GPU:
    Sapphire HD 7970 Ghz Ed.
    How am I supposed to give evidence of this?
    My laptop having 1C core deltas after a year and a half isn't enough evidence for you?
    Liquid metal doesn't destroy copper. It forms an alloy with it. Copper isn't aluminum. Once enough gallium is absorbed, the battery effect (which causes GA to be absorbed into Cu) is removed and the LM application remains stable. The whole point of the buffing with sandpaper early, is to make LM much easier to spread (try spreading it on a mirror finish surface and see how hard it is) and to give it something to adhere to, instead of it pooling to itself.

    The 10 minutes of spreading it (LIGHTLY) after the sanding, is to help accelerate absorption of the gallium. That's why you apply a new layer after the LM starts feeling "thicker."

    It was discussed on notebookreview forums that LM hates polished surfaces. That's where I learned this information, so I came up with that buffing technique with 1500 grit sandpaper.

    It worked for both my 9900k delid +LM+relid long term as well as my 7820k BGAbook.

    Before I used that technique (e.g. when I used to do what everyone else did and polished the heatsinks to a mirror finish), it gave TERRIBLE results. For example, the very first time I used Conductonaut, on my laptop, after 2 weeks, core temp deltas were almost 10C apart, and when I disassembled it, there was literally no liquid left, and what was left was a complete hardened layer that I had to actually -sand- off. (using napkins and alcohol took over 30 minutes).

    And on my 9900k, I did the same thing. What the other "me-toos" said--polished the underside of the IHS with fitz polish then lapped it to a mirror finish.
    And what I noticed was, first, it was very difficult to break the LM out of its ball shape.
    Finally, when it did spread, when I started experimenting with making a very thin layer, I noticed that it started "pooling" and streaking back to itself, leaving exposed gaps, because the surface tension between the LM was stronger than the mirror finish!

    When I applied more and got it to 'stop' streaking to itself, my temps were great at first in my 9900k. But after a few days of Prime95 and realbench stuff, I saw my temps start rising and my deltas increase. Once they got to 12C (originally was 6C), I knew something was wrong.

    I took apart the CPU IHS and looked under it, and there were spots on the CPU and underside of the IHS that had NO LM there at all! Even though it was sanded flat!
    The LM had 'pooled' to itself in several areas along the chip and the IHS, after thermal stresses, because there were no good surfaces for it to stick to!

    After i started experimenting on a spare laptop heatsink, I realized what was going on. When I finally perfected my 'buffing' technique, I found my core temp deltas remained stable for months (maybe losing like 1C over 4 months) and then stabilizing. The last time I checked my 9900k (delid +LM+relid) still intact, core temp deltas were still 7C a year later. I call that a win.

    I'm sorry if you don't believe me but I can't give you any more information. I did my time put in my hard work. If you don't want to profit from it, i'll just keep profiting from it myself.
     

  16. Falkentyne

    Falkentyne Master Guru

    Messages:
    544
    Likes Received:
    79
    GPU:
    Sapphire HD 7970 Ghz Ed.
    How did you determine that ZF-EX is repackaged TFX? How did Thermagic get access to it? Are they licensed with Thermalright?
     
  17. suty455

    suty455 Master Guru

    Messages:
    579
    Likes Received:
    250
    GPU:
    Nvidia 3090
    Happy with the Noctua H2 not too runny spreads well and coats well and keeps the Temps well under control
     
  18. jura11

    jura11 Guest

    Messages:
    2,640
    Likes Received:
    707
    GPU:
    RTX 3090 NvLink
    I only assume its repacked Thermalright TFX there, I tried both like TFX and ZF-EX on same GPUs and CPUs and difference between them has been minimal there

    Hope this helps

    Thanks, Jura
     
    Falkentyne likes this.
  19. Falkentyne

    Falkentyne Master Guru

    Messages:
    544
    Likes Received:
    79
    GPU:
    Sapphire HD 7970 Ghz Ed.
    Yeah I got some ZF-EX from the slow boat from China.
    I just wasn't sure because the specs on Thermalright's website don't exactly match the specs written on ZF-EX package, but the paste acts a LOT like TFX, including becoming "putty" consistency after it cures for a week (with equal to or better performance than new application just like TFX). So you're probably right! I had run out of TFX so I had to use ZF-EX and it does seem to perform and act, look and feel exactly the same.
     
    Last edited: Jan 2, 2021
  20. Neo Cyrus

    Neo Cyrus Ancient Guru

    Messages:
    10,780
    Likes Received:
    1,393
    GPU:
    黃仁勳 stole my 4090
    Is there any non-scalper that sells this stuff in Canada? I can only find regular Kryonaut from scalper 3rd party sellers in joke quantities of 0.5g/0.27ml. Yes 0.5g of the stuff is 0.27ml.

    I used to use Mastergel Maker but that was $22 for 1.5ml on Amazon and now it's $31.33... which is also taxed so that'd be $35.40, for 1.5 stinking milliliters. In this case the scalpers are Amazon themselves, it's not a 3rd party seller.

    Is there ANY very high end thermal compound I can buy right now that's not being scalped? Best I can find at near normal prices is MX-4, which I have anyway, but according to Guru3D's charts it's not quite up to par with the best stuff.
     

Share This Page