Which Thermal Paste to buy and apply (Traditional and Liquid Metal)

Discussion in 'Hardware Components and Aftermarket Upgrades' started by Vasudev, Jul 11, 2017.

  1. seanwee

    seanwee Notebook Deity

    Reputations:
    100
    Messages:
    975
    Likes Received:
    343
    Trophy Points:
    76
    Is a mirror finish heatspreader beneficial of you're using liquid metal? Or is it just better if you're using paste?
     
  2. hmscott

    hmscott Notebook Nobel Laureate

    Reputations:
    7,079
    Messages:
    20,412
    Likes Received:
    25,193
    Trophy Points:
    931
    I'd be interested to see if the high polish would seal against migration reducing draining the LM over time into the metal. That's where someone could innovate - make a surface that doesn't melt or migrate with LM.
     
  3. seanwee

    seanwee Notebook Deity

    Reputations:
    100
    Messages:
    975
    Likes Received:
    343
    Trophy Points:
    76
    I doubt it, since the migration is a chemical process.

    And there is a way of protecting against LM corrosion - nickel plating
     
    hmscott likes this.
  4. hmscott

    hmscott Notebook Nobel Laureate

    Reputations:
    7,079
    Messages:
    20,412
    Likes Received:
    25,193
    Trophy Points:
    931
    Then you answered your own question - polishing the surface to a mirror finish - removing the plating corrosion barrier - would be a bad idea for using LM over leaving it sealed and protected.

    That is unless you polished it - and then plated it, again. :)
     
  5. seanwee

    seanwee Notebook Deity

    Reputations:
    100
    Messages:
    975
    Likes Received:
    343
    Trophy Points:
    76
    I was talking about unplated copper heatsinks. Like on the GT76.

    Obviously you don't sand plated surfaces XD
     
    hmscott likes this.
  6. hmscott

    hmscott Notebook Nobel Laureate

    Reputations:
    7,079
    Messages:
    20,412
    Likes Received:
    25,193
    Trophy Points:
    931
    It's also obvious you'd want to lap a surface over leaving it raw - unless it came pre-lapped.

    If your question is - would making the surface mate more closely by lapping it squeeze out the LM causing it to migrate out from between the mating of the CPU / heatplate, then that would be just as true before lapping - which is why you use barriers to stop leakage of LM.

    A better mating across the surface improves heat transfer, so yeah, using LM doesn't make it any different than regular paste, except LM leaks with nothing to "hang on to".

    That's why I don't recommend using LM at all, it's an unnecessary risk for leakage and damage on the motherboard - and anything else it drizzles onto.

    Try lapping + a good enthusiast paste instead of LM, and if see for your specific usage going without LM is good enough. :)
     
    Last edited: Nov 28, 2019
  7. seanwee

    seanwee Notebook Deity

    Reputations:
    100
    Messages:
    975
    Likes Received:
    343
    Trophy Points:
    76
    I sorta see where the confusion stems from.

    Let me rephrase

    How much would a lapped surface reduce the temp difference between using LM and paste?
     
  8. hmscott

    hmscott Notebook Nobel Laureate

    Reputations:
    7,079
    Messages:
    20,412
    Likes Received:
    25,193
    Trophy Points:
    931
    It depends on the surface(s) before and after lapping. That's why I suggested trying it out and seeing your results. I've seen LM give 0c difference and 10c+ - a lot depends on pressure on the surfaces as well. It also varies according to thickness of the paste.

    I would suggest trying LM first (if you must) before lapping - a couple of pastes you like as well - then lap and repeat - come back and let us know.

    @der8auer did this a few times already and shows similar results for lapping improvements reducing paste improvements - it's a process of catching up from both ends, better thermal transfer from different means.
     
    seanwee likes this.
  9. Falkentyne

    Falkentyne Notebook Prophet

    Reputations:
    7,765
    Messages:
    5,679
    Likes Received:
    7,989
    Trophy Points:
    681
    Mirror finish is terrible. Nothing to do with absorption. Mirror finishes stop LM from even adhering to the surface thus making spreading it almost impossible. Then when it is spread finally, when the surface gets squeezed thin, LM starts adhering to itself rather than the surface, leaving completely exposed streaks or spots, because there is absolutely no grip from the surface whatsoever. This is horrible for temps.

    Buffing a surface with 1500 grit sandpaper makes LM love the surface, makes spreading easy and helps it adhere to the surface far better.
     
    electrosoft and hmscott like this.
  10. seanwee

    seanwee Notebook Deity

    Reputations:
    100
    Messages:
    975
    Likes Received:
    343
    Trophy Points:
    76
    Yeah I've heard of that, but doesn't it work just fine on the cpu and gpu dies which are already mirror finish?
     
    hmscott likes this.
Loading...

Share This Page