Rev B cooler after new Thermal Paste?

[nph]

I have seen a lot of posts about new Thermal Paste (mostly Arctic Silver 5) helping with the heat on MBA, most being positive. I have a rev B and have just bought the AS 5 but before I apply and risk the warranty I would like to see how many actually saw lower temperatures after applying it.

So if you have a Rev B and you have re-applied the Thermal Paste did you temperature under idle and load decrease?

Thanks

/Peter

 

[Scottsdale]

I helped a friend do it, and we reduced his CPU temp by 8-9 degrees C (we argued on what before and I thought it was 9 but he said 8 degrees difference)!

I plan to do this to my new MBA. I will post the results on Monday with my full review versus the 13" MBP. UNLESS, I get hung up on working on the display issue the whole weekend.

EDIT >>> ADD>>> Maybe it was F??? I don't remember which it was??? That really makes a big difference, and it was an original MBA and NOT a rev B. BUT, I would still do it if I were you.

 

[h1d]

Posted on the other thread, but on rev C 2.13, I didn't see much difference if not any reapplying paste, stays good from 50c to 80c and no kernel_task gone crazy.

 

[iaymnu]

Posted on the other thread, but on rev C 2.13, I didn't see much difference if not any reapplying paste, stays good from 50c to 80c and no kernel_task gone crazy.

That doesn't sound right. After reapplying and still hitting 80c. Is your room exceptionally warm? (room temp) . What are the fan speeds?

 

[dudeitsjay]

Man I've written this too many times

I've reapplied with AS5 and saw a decrease in temperatures ranging from 5C on idle to over 12+ C on loads. It'll still hit 80+ C regardless if you don't leave it on a flat surface or take the means to provide free airflow where the vents are located. But if you use it on a table, then you can see the decrease in temps. It's pretty dramatic for such a little mod. I'm tempted to just reopen mine and do a proper DIY on it, since my last one wasn't really good. Its easy if you've done it before like on desktop cpus.

Understand that thermal paste is not a coolant itself, it is a means of transferring heat. What essentially you're trying to do with a reapplication is transfer the heat away from the cpu and gpu onto the heatsink, which then acts as a larger surface area where the heat can transfer to the air which is being blown by the fan.

Thus you are simply doing two thing. 1) Getting rid of the crappy paste that apple put in that was overdone which ironically insulates the heat rather than transfers, and 2) Transferring the heat to someplace else. In my case, when it gets hot playing flash/netflix, the heat would slowly leak onto my aluminum underbody as the cold air intake and exhaust design isn't quite good enough to properly alleviate the heat from the heatsink before it spills onto adjacent components (Aluminum isn't too bad of a conducive material itself... yay for lapburns). It's not that bad as it sounds--I just had a bad time where it felt like a jacuzzi. But your cpu temps will definitely see improvements, and with that you'll have a longer lifespan.

People who are getting over 80C could definitely use this, and even if you're still getting that, it is most likely you are not giving clear space for your vents. It kinda blows since I generally want to watch something in bed, so I'd have to lift it up or prop it on a book.

 

[h1d]

Well, let's be a bit clearer.

I put my MBA on a flat wooden desk, so vents and everything are supposed to be good. I live in Japan and these days the room temperature is going up to 25c and a bit more soon enough with some nice humidity. (Not using air conditioning yet, so the room temp is at its natural state).

What I did with the 2.13 MBA is, when I got it, I run HD vids without tuning anything and hit 80c+. I figured I might give it a try and reapplied paste (Wiped it off completely and put some steady amount. This is 3rd time total, doing this on every MBA I owned, so I hope I know what I'm doing) And then, just bought another CoolBook just to see how it fairs. And, CB keeps reporting just about same temp... Maybe I should buy a good paste? I kept using something that was just found in house...

Also, I've read on another thread but iStats menu and CoolBook reporting different temperature when high, but which app should I use to get the actual good measurement?

Fans are moving at 2k when cool, 6.2k when hot.

 

[dudeitsjay]

Well, let's be a bit clearer.

I put my MBA on a flat wooden desk, so vents and everything are supposed to be good. I live in Japan and these days the room temperature is going up to 25c and a bit more soon enough with some nice humidity. (Not using air conditioning yet, so the room temp is at its natural state).

What I did with the 2.13 MBA is, when I got it, I run HD vids without tuning anything and hit 80c+. I figured I might give it a try and reapplied paste (Wiped it off completely and put some steady amount. This is 3rd time total, doing this on every MBA I owned, so I hope I know what I'm doing) And then, just bought another CoolBook just to see how it fairs. And, CB keeps reporting just about same temp... Maybe I should buy a good paste? I kept using something that was just found in house...

Also, I've read on another thread but iStats menu and CoolBook reporting different temperature when high, but which app should I use to get the actual good measurement?

Fans are moving at 2k when cool, 6.2k when hot.

IC, well to be frank, it's not the quality of the thermal paste that will make the biggest difference, though that is obviously relative to the quality of the paste you'll be replacing the oem with. Re-applying is more concerned with undoing the crappy insulating job that was done, rather than replacing it for a better conducive material. Be sure you're not putting on too much--I used 1/3 a grain of rice worth of AS5 on GPU and CPU each.

Guessing on the iStats, but the Coolbook refreshes every second, while the minimum you can refresh the temp on iStats is 2sec. Considering this, I'd say that its just a matter of difference in sample sizes that can lead to differences in average temperature readings--though 8C up is pushing it (maybe they're different temperature reading nodes...). Stick to a good subjective estimate of the temperature through the CB readings if the differences between the iStats and CB readings are dramatic.