Kaby Lake crap TIM , WHY ?

Discussion in 'Intel Processors' started by gregorvich, Dec 18, 2016.

  1. Nimisys

    Nimisys Likes To Play With Trannies

    Messages:
    17,062
    Joined:
    Jun 20, 2000
    are those cracks i see?
     
  2. FrgMstr

    FrgMstr Just Plain Mean Staff Member

    Messages:
    48,386
    Joined:
    May 18, 1997
    Yeppers. Delidding 13 years ago....before it was cool. ;)
     
    Jspr, razor1 and Peter2k like this.
  3. FrgMstr

    FrgMstr Just Plain Mean Staff Member

    Messages:
    48,386
    Joined:
    May 18, 1997
    Good eye.
     
    Deleted member 184142 likes this.
  4. gregorvich

    gregorvich Limp Gawd

    Messages:
    130
    Joined:
    Mar 6, 2004
     
  5. Folterknecht

    Folterknecht n00b

    Messages:
    31
    Joined:
    Nov 28, 2016
  6. ccityinstaller

    ccityinstaller [H]ardness Supreme

    Messages:
    4,230
    Joined:
    Feb 23, 2007
    I know this post is nearly a month old, but I have to share some of my results on this whole " you cant use LM or LMU on a vertical setup" nonsense..I have used it on well over ~15 de-lidded direct die mounts in a standard ATX MB, and at least 10 re-lidded setups, going back to 2012...I haven't had a single failure, and we are knocking on the door of 5 years of use here...

    The stuff doesn't run like water guys..I use a cheap throw away water color brush, and lightly coat the entire die before mounting the water block or re-lidding..
     
    Jspr likes this.
  7. Shintai

    Shintai [H]ardness Supreme

    Messages:
    5,691
    Joined:
    Jul 1, 2016
    So you also create a setup that cant be used in a commercial product. Either no lid or relidded directly on the die using a razor thin layer. You have to think on what would work for everyone in any situation. Including lasting for 10 years+.
     
  8. ccityinstaller

    ccityinstaller [H]ardness Supreme

    Messages:
    4,230
    Joined:
    Feb 23, 2007
    Where in my OP did I say anything about creating a "setup that cannot be used in a commercial product?" I simply stated that the claim you and others have made about CL and CLU "leaking" and ruining systems is FUD. Period. I offered my experience on 25 systems that use this very setup without a problem. I think you like to argue just for the sake of arguing..Thank God I didn't mention AMD in my post :rolleyes::rolleyes::rolleyes::rolleyes:.........
     
    c3k, SomeGuy133 and thesmokingman like this.
  9. Shintai

    Shintai [H]ardness Supreme

    Messages:
    5,691
    Joined:
    Jul 1, 2016
    So what happens when you cant use a razor thin layer? Either it runs or is solidifies doesn't it?
    https://forums.anandtech.com/thread...ne-through-the-roof-after-few-months.2419707/
    http://www.coollaboratory.com/pdf/manual_liquid_pro_englisch.pdf
     
    Last edited: Jan 17, 2017
  10. -PK-

    -PK- [H]ard|Gawd

    Messages:
    1,798
    Joined:
    Aug 6, 2004
    Last edited: Jan 17, 2017
  11. Peter2k

    Peter2k Limp Gawd

    Messages:
    309
    Joined:
    Apr 25, 2016
    You know it's funny that it's actually mentioned that the ultra solidifies in the manual

    at the end of the second page


    at 48 hours it consolidates

     
  12. FrgMstr

    FrgMstr Just Plain Mean Staff Member

    Messages:
    48,386
    Joined:
    May 18, 1997
    Vote with your wallet. I don't think anyone at Intel made you buy anything.
     
    Last edited: Jan 20, 2017
    ZodaEX likes this.
  13. olDirdey

    olDirdey n00b

    Messages:
    4
    Joined:
    Jan 13, 2005
    Thank you very much Folterknecht (y)

    It´s me, Stullen Andi from Hardwareluxx Forum. The TIM I´am using to apply on the core is Phobya Liquid Metall or Thermal Grizzly Conductonaut. The Coollaboratory stuff I only used once cause it´s not as good as the phobya or conductonaut stuff.

    Here you can see how easy it is to apply it to the die, but I am not using the brush that comes with the tim.

     
  14. SomeGuy133

    SomeGuy133 2[H]4U

    Messages:
    3,447
    Joined:
    Apr 12, 2015
    So your also saying and would agree a Titan Pascal that throttles due to shit heatsink is working as advertised and totally cool?
     
  15. SomeGuy133

    SomeGuy133 2[H]4U

    Messages:
    3,447
    Joined:
    Apr 12, 2015
    it throttles at stock boost.

    A product made cutting corners is not something consumers should find acceptable. period.
     
  16. SomeGuy133

    SomeGuy133 2[H]4U

    Messages:
    3,447
    Joined:
    Apr 12, 2015
    being advertised as an overclocking chip (K) and cutting edge while cutting corners that cripples its performance is not something you should accept so stop spewing your crap
     
  17. Peter2k

    Peter2k Limp Gawd

    Messages:
    309
    Joined:
    Apr 25, 2016

    mmm

    that technique to stretch it out in a line and then spreading it to the sides, nice
    gonna remember that whenever I have to redo mine (if I have to that is, happy with my OC)

    first time on my air cooler and on Skylake I used liquid ultra

    but some time back I ordered LM from Phobya and with my Kaby and an AiO I used the Phobya this time around (between DIE and IHS and the IHS and AiO)

    I'm sure you get asked a lot
    any long term experience you've come across, or heard regarding the coollaboratory or Phobya stuff?

    though seeing on the Hardwareluxx Forum youre rather experienced in delidding :ROFLMAO:
    any first hand knowledge / opinion / experience would be appreciated (as youre test samples seem rather high :D )
     
  18. olDirdey

    olDirdey n00b

    Messages:
    4
    Joined:
    Jan 13, 2005
    I tried a few ways to apply the tim on the die and spreader and for me it was the easiest way to do it with a "wattestäbchen", I think you call it cotton bud.

    The very first application I used the coollaboratory stuff too, but after this I only used the phobya and conductonaut lm, cause it you got a few degrees better result.

    As far as I know the delidding starts in 2012 in hwluxx forum and there are cpus I know that running more than 3 years without a problem. My first cpu is working 1 and a half year without any difference in temp.

    I think the key is to apply the lm to a very clean surface and leave a gap while glueing. This one here had no gap and I think not enaugh pressure while remounting the IHS and this cpu starts to raise temp after a few months. I got it for reapplying LM.

    But even this is fixable with some cleaning. The Liquid Metall on the IHS was dry and needed some heavy force to remove.

    [​IMG]

    [​IMG]

    [​IMG]
     
  19. FrgMstr

    FrgMstr Just Plain Mean Staff Member

    Messages:
    48,386
    Joined:
    May 18, 1997
    Actually it dates back to 2003. Ask me how I know. ;)
     
  20. olDirdey

    olDirdey n00b

    Messages:
    4
    Joined:
    Jan 13, 2005
  21. SomeGuy133

    SomeGuy133 2[H]4U

    Messages:
    3,447
    Joined:
    Apr 12, 2015
    cotton tip/q-tip i think is what you mean above.

    He showed a picture further above of him delidding a 478 socket chip IIRC

    I like the idea of using 5-10K grit paper/cloth to get better contact.

    I have 1-3/5K paper and a 8K IIRC cloth. The cloth is for diamond polishing.

    6/8K is like 1 micron particle size.

    Problem is applying even pressure and not making it worse.
     
    Last edited: Feb 2, 2017
  22. FrgMstr

    FrgMstr Just Plain Mean Staff Member

    Messages:
    48,386
    Joined:
    May 18, 1997
    Always had luck with putting the sand paper down on a thick glass mirror and then lapping the IHS on that. Nice and flat.
     
  23. FrgMstr

    FrgMstr Just Plain Mean Staff Member

    Messages:
    48,386
    Joined:
    May 18, 1997
    I know what you are saying. Physics...I get it.

    "Lapping" is the process we use to flatten out IHS parts. And yes, while it is impossible to get "flat" a lot of times in the past the IHS was nowhere close to being flat.

    This is exactly how we used to do it.



    I would however suggest using a circular motion to wet sand however...

    I did address the flatness of the KL IHS today in our review. http://www.hardocp.com/article/2017/02/02/intel_kaby_lake_i57600k_cpu_delid_relid_temp_results
     
  24. FrgMstr

    FrgMstr Just Plain Mean Staff Member

    Messages:
    48,386
    Joined:
    May 18, 1997
    Also, if you lap it down to a mirror finish, you can then use a piece of graph paper to check for flatness as well.

    1351693272ZeNpxmp4Kx_2_5_l.jpg
     
    Peter2k likes this.
  25. FrgMstr

    FrgMstr Just Plain Mean Staff Member

    Messages:
    48,386
    Joined:
    May 18, 1997
    Pretty much. Although they well never be "straight."
     
  26. FrgMstr

    FrgMstr Just Plain Mean Staff Member

    Messages:
    48,386
    Joined:
    May 18, 1997
    OK, thanks for the insight.
     
  27. FrgMstr

    FrgMstr Just Plain Mean Staff Member

    Messages:
    48,386
    Joined:
    May 18, 1997
    Sorry, missed the question.

    Yes, that is a picture we took. No, we do not used bubbled up paper for that test. The paper is flat on a desktop.
     
  28. olDirdey

    olDirdey n00b

    Messages:
    4
    Joined:
    Jan 13, 2005
    It´s a bit confusing to me. I think it´s just a 479 socket but the CPU is socket 478. I don´t know a single socket 479 CPU with heatspreader.
     
  29. FrgMstr

    FrgMstr Just Plain Mean Staff Member

    Messages:
    48,386
    Joined:
    May 18, 1997
    Put a piece of graph paper down on a flat surface. Look at the reflection in the mirror finish. Are the blocks in the reflection still square? If yes, it is "flat." If no, then it is not "flat." Does that help you understand better? I am not sure of what you are confused about. Sorry for my snapping at you. Please post your question here and I will try to answer it.
     
    ZodaEX likes this.
  30. FrgMstr

    FrgMstr Just Plain Mean Staff Member

    Messages:
    48,386
    Joined:
    May 18, 1997
    I really do not know brother. When we had wavy mating surfaces it was an easy way to show it to our readers.
     
  31. SvenBent

    SvenBent 2[H]4U

    Messages:
    3,035
    Joined:
    Sep 13, 2008
    So the lines are only good if they are straight.. bunch of homophobics around here. gay cobber-bases are cobber-bases too you know.
     
    FrgMstr likes this.
  32. exlink

    exlink [H]ardness Supreme

    Messages:
    4,312
    Joined:
    Dec 16, 2006
    So what do you guys think are the chances that the newly rumored i7-7740K and i5-7640K have better TIM?
     
  33. PontiacGTX

    PontiacGTX Gawd

    Messages:
    695
    Joined:
    Aug 9, 2013
    Intel will probably keep using epoxy for Enthusiast/HDET CPUs and still giving bad Thermal interface to rmainstream
     
  34. damstr

    damstr [H]ard|Gawd

    Messages:
    1,362
    Joined:
    Jun 26, 2005
    I doubt they would use higher quality TIM on the lower end CPU's.

    I think there is also an issue with the Z height which is the distance of the die to the IHS. Removing the glue or whatever they use helps the IHS sit closer to the die.
     
  35. SomeGuy133

    SomeGuy133 2[H]4U

    Messages:
    3,447
    Joined:
    Apr 12, 2015
    silicon lottery said its merely a lack of application pressure which is why it seats too high.