Rev.B re-applying thermal paste

[tsubikiddo]

Warm greetings to other MBA users

This is probably my first thread in MR
I have a MBA rev.B base model, it is a lovely machine.
Its operating temperature was alright (certainly much cooler than the rev.A)
~43C from cold start, and working at ~70C with Acrobat9 Pro, OutlinerPro, Fx3, Word, Mail + WiFi (802.11g)
~77C + max spin fan with YouTube video

But as my interest lies in maximum battery endurance, my focus quickly turn towards the spinning fans.
(My concept was, slower spin = longer battery life)

Having heard the general comments of AAPL's thermal paste,
I bought a tube of AS5 and did this quite a few months ago.

The temp now:
37~39C from cold start
~57C with all the aforementioned apps running, or watching YouTube video

and, as expected, lengthened battery life
(I have also installed smcFanCtrl a few weeks later and locked the fan @1800rpm)

Hopefully this is just another thread that get you MBA users actively thinking about re-applying thermal paste ahead of summer

hope this may help a little

 

[tsubikiddo]

oh, and I used the blob method when I applied the AS5

for this reason:
I am sort of expecting the thermal paste will kind of 'melt' by the CPU/GPU heat and gradually spread across the die surface
(and in fact, it did! The photos in #1 only show my 1st attempt at re-applying AS5. I did re-reapply the AS5 a couple of times later on that week to get the thermal paste quantity right)

whereas in contrast, if the thermal paste is all spread out,
I kind of have a worry that there is always a chance of trapping a blanket of air in the middle when the heatsink is reinstalled onto the CPU/GPU

The reason that I think the blob method is better is, the paste spread itself across the die surface and so it should be pushing the air away while it spreads.

am I understanding the thing correctly? or do I simply think/worry too much?

Thanks in advance

 

[nph]

This looks cool

It seriously makes me think about doing the same on my MBA.
Heat is energy and if temperature can drop that much then it would be worth it.
What is the improved battery time you see now vs before?

Thanks

 

[tsubikiddo]

This looks cool

It seriously makes me think about doing the same on my MBA.
Heat is energy and if temperature can drop that much then it would be worth it.
What is the improved battery time you see now vs before?

Thanks

Hey thanks
Yea, I was kinda going to the extreme, trying to maximize my MBA's battery life
I mean, hey, who knows how long this task/session/meeting is gonna last before you can recharge your MBA?
With the already impressive battery life of the MBA, a few people would have actually carry the AC/DC adapter with them into the meeting/session if they are sort of forecasting a duration that MBA's battery is gonna 'outlast' it, rite?

To be honest, I have never dried my MBA's battery once.
Going into the red zone, yes, but never completely drained it to the point of forced shut down.

So, Sorry, I simply just don't have the numbers.
(Cuz I re-apply the thermal paste fairly early once I received my MBA, before I actually have a chance to test drive the battery in all its stock settings)
However, I can share some of my previous exp, and you can make an informed, intelligent guess. Feel free to do so.
(all the numbers and figures are post-reapplying thermal paste)

I was once in a meeting for ~210mins (~3.5hrs),
by the end of the meeting,
battery indicator (the one on the menu bar) shows 25% remaining and forecast 1:13 remaining battery life.

My settings in that meeting: (pretty crazy I'd say)
Brightness: 4 blocks
Keyboard Bklit: Off
Sound: Mute
Bluetooth: Off
Battery/Performance profile: default (in fact, I have no idea where do I adjust this)
WiFi (802.11g): Occasionally switched on, just to look up the GOOG results or catch that email.
Apps (all launched): Word'08, Outliner Pro, Acrobat 9, iCal, Address Book, Fx3, TextEdit
Exposé and Spaces were also left on, pretty much in their default status.

I have never had any other productive session longer than this.

Hope this exp will be helpful to you and other MBA owners

 

[SeanU]

I've been thinking about doing this on my rev-a. It runs fine (67deg under load) but I like things to be as optimized as possible as well. One thing I might consider doing is getting some thin copper pads to remove some of the gap between the chips and the heatsink. It should be more efficient than glopping on the AS-5. The only downside is that there would need to be thermal paste on both sides of the copper. Any body think this will improve things... or is it not worth trying?

 

[spencers]

I really need to do this.

 

[tsubikiddo]

I've been thinking about doing this on my rev-a. It runs fine (67deg under load) but I like things to be as optimized as possible as well. One thing I might consider doing is getting some thin copper pads to remove some of the gap between the chips and the heatsink. It should be more efficient than glopping on the AS-5. The only downside is that there would need to be thermal paste on both sides of the copper. Any body think this will improve things... or is it not worth trying?

Hi there, thank you for your reply.
As you can see, I am just such a n00b in this amazing MacWorld & Thinnovation

Would you mind elaborating on, how is the copper padding method a better, efficient way of managing the heat?

My personal opinion, like hifi audio, the system is as good as the weakest link.
Although copper is considered as the superior heat conductor, nevertheless, you still need thermal paste to provide a tight, sealed contact between the CPU&GPU die surface and heatsink.

In fact, I'd argue the whole setup is a worsen form of heat management solution, as there will be more (/thinker) thermal paste involved into dissipating the heat.

-----
The situation that I can imagine:
CPU surface(1) --thermal paste(2)-->copper piece(3)--thermal paste(4)-->Alu heatsink piece(5)
-----
(2) + (4) = more thermal paste needed than necessary
superior heat conductivity @(3), but still need to go through (2),(4)
-----
my rationale: Less is more, simple things work better.

Love to hear from you.

 

[SeanU]

Hi there, thank you for your reply.
As you can see, I am just such a n00b in this amazing MacWorld & Thinnovation

Would you mind elaborating on, how is the copper padding method a better, efficient way of managing the heat?

My personal opinion, like hifi audio, the system is as good as the weakest link.
Although copper is considered as the superior heat conductor, nevertheless, you still need thermal paste to provide a tight, sealed contact between the CPU&GPU die surface and heatsink.

In fact, I'd argue the whole setup is a worsen form of heat management solution, as there will be more (/thinker) thermal paste involved into dissipating the heat.

-----
The situation that I can imagine:
CPU surface(1) --thermal paste(2)-->copper piece(3)--thermal paste(4)-->Alu heatsink piece(5)
-----
(2) + (4) = more thermal paste needed than necessary
superior heat conductivity @(3), but still need to go through (2),(4)
-----
my rationale: Less is more, simple things work better.

Love to hear from you.

Well... it's just a theory... but my thinking is *much* less thermal paste could be used if the copper thickness was a good match... like you said, copper is a much better thermal conductor. But, also as you say, it would create more thermal "interfaces" so I'm not sure if it would help or hurt. How much thermal paste is required to fill the gap to the heatsink?

It's probably more effort than its worth... but I like to tweak stuff.

* Edit : I found this link discussing similar issues with the eeepc. Interesting read : http://forum.eeeuser.com/viewtopic.php?id=20683