Arctic Silver Ceramique thermal paste...Results! (Pax please read)

Discussion in 'MacBook' started by AlphaDogg, Jun 9, 2010.

  1. AlphaDogg, Jun 9, 2010
    Last edited: Jan 16, 2011

    AlphaDogg macrumors 68040

    AlphaDogg

    Joined:
    May 20, 2010
    Location:
    Boulder, CO
    #1
    I went into my local microcenter and I purchased 2.5g of Arctic Silver Ceramique (really good price, $3.99). I would not dare reapply already have reapplied thermal paste on my brand new MBP (specs in sig), but I have an old (late 2007 model) white MacBook which was in desperate need of a reapplication.

    Before the reapplication of thermal paste, it would idle between 55-65ºC (131º-150ºF). While watching flash videos, it would hover around 91ºC (195ºF)

    After the reapplication of the thermal paste, it idles between 39º-53ºC (102º-128ºF). Now, it never goes over 71ºC (160ºF) while watching flash videos.

    If you can, I would recommend reapplying thermal paste. I would also recommend Arctic Silver Ceramique. I would also like to report how easy it was, if you follow ifixit's online guides on how to remove the heatsink and how to apply thermal paste. I have never done this before, and when following these guides, it was very easy.

    A good way to keep track of your screws from your MacBook is to take a piece of duct tape, and flip it over so that the sticky side is exposed. Put the screws on the duct tape, and you will find that they stay where you put them and it is easy to organize the screws.

    Note: I am not responsible for any damage done to your property resulting from this post/thread.
     
  2. ClassicII macrumors 6502

    ClassicII

    Joined:
    Sep 21, 2009
    Location:
    Chicago, IL
    #2
    190! hot damn.

    When you were inside the laptop was it really dusty ?
     
  3. AlphaDogg thread starter macrumors 68040

    AlphaDogg

    Joined:
    May 20, 2010
    Location:
    Boulder, CO
    #3
    The fan was slightly dusty, but there was no dust whatsoever, anywhere on the logic board or components. The only variable here is the thermal paste.

    Right now, while just reading forums, my internal CPU temperature is at 117ºF and my fans are at 1794rpm. This thing is actually bearable to have sitting on my lap! Because the fans aren't spinning up as high, my battery life has also increased dramatically.

    I also plan on doing this to my aging iBook G4 (14" 1.33GHz model) that was given to me, because the previous owner had no use for it. My mom uses it now, and she complains of heat. Does anybody know about any problems that I would encounter? I know that the iBooks had graphics issues. Will removing the induce invoke any graphics issues?
     
  4. nph macrumors 6502a

    Joined:
    Feb 9, 2005
    #4
    Did you notice if the case got any warmer after the re-application?
     
  5. #5
    I used Arctic Silver #5 on my Black Macbook.

    To make a long story short, I spilt an entire, huge glass of coke on my Macbook a while ago, so I had to disassemble the entire machine and clean it out. I basically gave it an alcohol bath.

    When I removed the heat-sink, I saw that whoever had applied the original thermal paste at the factory did one hell of a sloppy job, because there was too much of it and it was sloppily applied.

    I picked up a tube of Arctic Silver #5 when I was putting the heat-sink back on the CPU, I got my machine up and running again, and it's working better than ever!
    :D
     
  6. AlphaDogg thread starter macrumors 68040

    AlphaDogg

    Joined:
    May 20, 2010
    Location:
    Boulder, CO
    #6
    That is very unfortunate, but at least you knew exactly what you were supposed to do. I am glad that your machine is up and running again! Arctic Silver 5 dries out much quicker that Arctic Silver Ceramique, so you should be wary of the heat and if it gets past a certain temp, shut it off and reapply the paste.

    before when it was on my lap, when i was surfing the interwebs, it would feel pretty hot. Now when its on my lap, when i am surfing the interwebs, it feels like it is powered off.
     
  7. #7
    I didn't take or monitor any CPU temps before my machine died, but I've had my machine running for about 6 months now, since I applied the new paste, and it doesn't seem to run any hotter than before at least.

    I pretty much average around 60-61 C, that's what temp-lite is telling me right now as I am typing this. It rarely goes above 70 C though, so I guess that's good enough for me. My Black Macbook is a late '06 model.
     
  8. AlphaDogg thread starter macrumors 68040

    AlphaDogg

    Joined:
    May 20, 2010
    Location:
    Boulder, CO
    #8
    When thermal paste dries out, the thermal contact is nonexistent. Your processor can overheat and you can lose your processor. If you're not careful, this could happen to you.

    Because the paste has started curing/setting, my temps hover around 115ºF when just reading forums. This thing runs like it should. Right now, my internal CPU temp is 112ºF
     
  9. #9
    But is the Arctic Silver #5 only meant to last for as little as six months?

    I was reading a shootout recently between a whole bunch of different thermal pastes and the Silver #5 scored pretty good in that test.
     
  10. AlphaDogg thread starter macrumors 68040

    AlphaDogg

    Joined:
    May 20, 2010
    Location:
    Boulder, CO
    #10
    I don't know how long Arctic Silver 5 is supposed to last, but all I know is that it dries out quicker than Arctic Silver Ceramique. I do not know how much quicker it dries that Ceramique, but all I know is that it is not one of the longest lasting thermal compounds.
     
  11. #11
    I was just reading some big thread on an overclocking forum and some of the people there said that yeah, it will dry out or harden eventually.

    I don't think I have anything to worry about just yet, but it's not that big of a deal for me to just reapply some again in a few months time, as I still have the tube lying around, and it probably won't take more than 5-10 minutes to do.

    I also used that same paste when I made my hackintosh a couple of weeks ago, and CPU temps on that average in the upper 30's C, though that is a desktop of course (with a third party heatsink) and not a laptop with a tiny heatsink.
     
  12. pastrychef macrumors 601

    pastrychef

    Joined:
    Sep 15, 2006
    Location:
    New York City, NY
    #12
    I applied Arctic Silver #5 in my MacBook a while back. My results are similar to Black MacBook's. I remember having difficulty removing the temp sensor cable that was attached to the heatsink. Fortunately, I managed to put it all back together.

    I didn't know about Arctic Silver drying out. I've had machines where I applied the stuff once and continued to use the machines for years without ever re-applying. My Mac mini which was upgraded from C2D 1.83GHz to 2.33GHz has been running 24/7 for about 1.5 years since the upgrade and application of Artic Silver. The only time I saw temps increase over that time was due to dust getting in to the fan/heatsink area. Once I opened it up and used compressed air to clean it out everything is running fine again.
     
  13. AlphaDogg thread starter macrumors 68040

    AlphaDogg

    Joined:
    May 20, 2010
    Location:
    Boulder, CO
    #13
    Like I said, I do not know how long it is supposed to last. It also depends on your usage, IIRC.

    Sorry for the late reply, and I inadvertently bumped this thread :eek:.
     
  14. Pax macrumors 6502a

    Joined:
    Dec 12, 2003
    #14
    AlphaDogg, can you post the temp difference between CPU heatsink and CPU at idle, and (ideally) under full load eg running 2 x yes > /dev/null. If you have the same figures from before paste renewal that would be great.

    I'm very interested in this - obviously the paste forms the thermal interface between CPU and heatsink. "Bad" paste will result in a high temperature difference between the two.

    I think Apple's paste application on the newer Macs is much better:- on my 2009 C2D Unibody the temperature difference is about 5 C at idle and just over 20 C under full load. Which I think is not bad. And it seems to be very consistent from Mac to Mac - I have seen loads of posts on here with people with very similar values

    I wonder if the "bad paste" phenomenon was only limited to early pre-unibody models.

    Hope you can find time to post these numbers.

    Cheers
     
  15. AlphaDogg thread starter macrumors 68040

    AlphaDogg

    Joined:
    May 20, 2010
    Location:
    Boulder, CO
    #15
    I do not believe that the thermal paste application on newer macs is better, but I believe that the way the heat is routed out of the machine is better, ie better thermal ducts, or whatever. I will post the temps after 2x yes>/dev/null soon-probably tomorrow.

    Unfortunately, I do not have any of this data from before the reapplication. Here are my temps while running 2x yes >/dev/null processes after 10min.

    CPU-193ºF
    Heatsink A-158ºF
    Heatsink B-155ºF
    Enclosure Base-81ºF

    However, immediately after closing the yes >/dev/null processes, the CPU temp dropped down to 130ºF.

    I will allow it some time to cool down, and then I will post the idle temps.

    These numbers seemed high, so i ran the same tests on my 13" MBP. I only monitored the CPU temps on my MBP, but you will get the point. In the same test, my 13" MBP got to 198ºF and when the processes were closed, it dropped down to 155ºF. So obviously, my thermal paste application is superior to Apple's. Btw, my MBP is unmodified.

    On my white MB, while running idle, after about 15min, the cpu is at 106ºF, and both heatsinks are at 106ºF. The enclosure base is 81ºF.

    sorry for the bump! I did not know it was against forum rules!
     
  16. Pax macrumors 6502a

    Joined:
    Dec 12, 2003
    #16
    Hi AlphaDogg, thanks very much for these results. Very interesting. I think they confirm a theory I have had for a while, which is that older Macs did have problems with thermal paste, but newer ones do not. Let me explain, see what you think.

    The junction between the CPU die and the heatsink causes thermal resistance. Moving heat energy over this junction causes a temperature difference. One way to measure the thermal resistance is in ºC/Watt (or ºF/Watt). So a given heat flow expressed in Watts will cause a temperature difference between the CPU and heatsink. Bad paste will cause a higher thermal resistance, say 10 ºC/Watt, good paste a lower value, say 1 ºC/Watt.

    Now the CD and C2D CPUs (which we both have) had a Thermal Design Power of 25 Watts. So at full output they emit around about 25 Watts of heat. Nearly all of that goes through the thermal paste.

    Your excellent analysis shows that with well-applied quality paste, when your CPU is at full load (about 25 Watts) there is a 20 ºC temperature difference between CPU and heatsink, so the thermal resistance is about 0.8 ºC/Watt.

    Now my C2D unibody with its original paste shows just a sightly higher value, about 0.9 ºC/Watt. I have seen a few other unibodies and they have similar values. (I bet yours does too)

    The new i5 and i7 CPUs have a 35 W TDP. This thread shows some i5/i7 users' CPU and heatsink temperatures under full load: http://forums.macrumors.com/showthread.php?t=908210. The temperature differences are....
    Post 8 -> 33 ºC
    Post 22 -> 27 ºC
    Post 24 -> 30 ºC
    Post 34 -> 33 ºC
    Post 36 -> 44 ºC

    Average thermal resistance is about 0.95 ºC/Watt, just a little more than you have got with your top of the range Arctic Silver Ceramique, and about the same as I have on my C2D. (Post 36 is a little funny, maybe something odd was going on there, but I included it in the average)

    So I think this shows that the paste which Apple applies in its new products is pretty good, nearly as good as "The Best Paste In The World" carefully applied.

    cheers
     
  17. ouimetnick macrumors 68020

    ouimetnick

    Joined:
    Aug 28, 2008
    Location:
    Beverly, Massachusetts
    #17

    Here is a picture from iFixit of the 2009 MacBook Pro 13" logic board with Apple applied thermal paste. I've seen :apple: computers from 2008 like the Unibody Aluminum MacBook that had the paste applied correctly from Apple. Seems like a hit and miss. They didn't completely fix the issue, as some MacBook Pro Core i5 and i7s have bad thermal paste too.
     

    Attached Files:

  18. Pax macrumors 6502a

    Joined:
    Dec 12, 2003
    #18
    I just spotted a couple of "oops" in my post above

    1) The CoreDuos had a 31 W TDP, at least that's what it says herehttp://ark.intel.com/ProductCollection.aspx?familyId=22731. That makes AlphaDogg's 20 C / 31 W = 0.65 ºC/W, somewhat better than I estimated above

    2) there are some small-ish differences in die size, about 10%, which also affect the results a bit

    So maybe I should modify what I said before - the paste in Apple products is not "nearly as good as the best paste in the world" - it could be better. How much? I don't know.
     
  19. Pax macrumors 6502a

    Joined:
    Dec 12, 2003
    #19
    Hi Ouimetnick. Yes I was aware that the paste splats don't look ideal, I was interested to know whether the bad application makes any significant difference to the real world thermal performance. I guess all my waffle up above is inconclusive.

    My 2006 1.83 MB shows 17 C difference @ 31 W so that's about 0.55 C/W, slightly better than AlphaDogg's after its Arctic Silver. But my MB had a new logic board in late 2006, so I've no idea how "good" the paste is. Must be pretty good though.
     
  20. AlphaDogg thread starter macrumors 68040

    AlphaDogg

    Joined:
    May 20, 2010
    Location:
    Boulder, CO
    #20
    You are mistaken. I have a 2.2GHz C2D.
     
  21. Pax macrumors 6502a

    Joined:
    Dec 12, 2003
    #21
    Interesting. So my original estimate of 20 C / 25 W = 0.8 C/W was correct. Not vastly different to my "stock" paste value of 0.9 C/W.

    thanks for all the useful info you've posted.
     
  22. AlphaDogg thread starter macrumors 68040

    AlphaDogg

    Joined:
    May 20, 2010
    Location:
    Boulder, CO
    #22
    I got bored after I reapplied paste on my 2010 MBP and 2010 MB, so I reapplied paste on my 2007 MB once again. The numbers are astonishing!

    Here is a screenshot of idle (ºC):[​IMG]

    Here is a screenshot while running 2x "yes > /dev/null" processes (ºC):
    [​IMG]

    On idle, the differences between the CPU and heatsink A is only 2ºC, and on full load, the difference between the CPU and heatsink A is only 21ºC!
     
  23. Pax macrumors 6502a

    Joined:
    Dec 12, 2003
    #23
    Nice one
    On my 1.83 CD 2006 Macbook at idle the temp difference is 3-4 C. So yours is similar to, a bit better than mine. Excellent stuff.

    Interesting how the old Macbooks seem to have a lower idle difference than the newer C2D models. Some possible reasons for this.
    - the CDs have a lower idle power consumption than the C2Ds
    - the CDs have a larger area in contact with the heatsink (more area for heat dissipation means lower temperature difference)

    I suspect it's the latter, the CDs were made on a different process technology so they may be physically larger than the C2Ds. I'm sure some Intel geek could tell us for sure.

    Anyway thanks for giving me more insight into thermal paste stuff.

    I think from now on my answer to people who come on these forums and say "Should I change my paste?" will be:- "if your idle temperature difference is less than about 5 C, and your full load difference is less than about 25 C, don't bother, you won't see much difference".

    Also when people come on and say "OMG Apple's stock paste is garbage" I think the answer is "no it's not, there might be a few lemons with Friday afternoon paste jobs, but for many (the vast majority?) of Macs the stock paste is nearly as good as well applied Arctic Silver."

    cheers now, enjoy your Macs.
     
  24. AlphaDogg thread starter macrumors 68040

    AlphaDogg

    Joined:
    May 20, 2010
    Location:
    Boulder, CO
    #24
    Pax.... It is a 2.2ghz C2D Processor, not a CD processor...
     
  25. Mdv2 macrumors regular

    Joined:
    Feb 26, 2008
    #25
    I've also recently reapplied the thermal paste (Arctic Silver 5) and I get the following temperatures.

    [​IMG]

    Are these idle temperatures normal (especially CPU and heat sinks)?

    It's a Macbook unibody 2008 13''
     

Share This Page