[Kraetos]

So the thermal paste salesman is workin' overtime again, because Apple is buying about 3x as much as they need!

New MacBook Pros still overheating due to misapplied thermal grease?

What's the deal with this? Surely Apple engineers know that too much thermal paste is actually bad for a computers cooling mechanisms? Surely they inform the assembly workers and robots that they need to lay off the paste? Surely, after this has happened on previous models and been blogged about all over the intertubes, Apple would realize that this shit doesn't fly?



Apple isn't supposed to cut corners like this. I don't know if this is mismanagement, confusion, or idiocy, but I expect better from Apple.

[datapusher]

yikes. that's a huge deterrant to buying a MBP, but i wonder how widespread this problem really is. when things are published on the internet (especially on Engadget), it seems to put the issue under a high powered microscope to the point where you have these applecondriacs complaining of a "hotter" MBP as well.

for old times sake, i have to wonder if this is what we would've seen if the G5 powerbook ever saw the light of day.

[Kraetos]

Originally Posted by datapusher yikes. that's a huge deterrant to buying a MBP, but i wonder how widespread this problem really is. when things are published on the internet (especially on Engadget), it seems to put the issue under a high powered microscope to the point where you have these applecondriacs complaining of a "hotter" MBP as well. for old times sake, i have to wonder if this is what we would've seen if the G5 powerbook ever saw the light of day.

Yeah, and this has been a problem for since February 2006. What the hell, man?

Christ. G5 PowerBook? It's not like I really need my sperm.

[datapusher]

i have to wonder why this issue has been neglected. it seems like most mobile computing manufacturers have this down pat with their C2D offerings, I don't understand why Apple has been coming up more often besides the fact that they're Apple and their name alone gets bloggers page views and MS fanboys fodder.

I hope increasing coverage will make them evaluate this gross overuse of thermal paste.

[julesstoop]

This design is by choice, actually.

If Apple were to use the "appropriate" amount of thermal paste, the heatsink temperature would change much more quickly with short bursts of activity from the CPU (or GPU). This would make the fans kick in much more often for short amounts of time. Most people would not prefer this behaviour.

So the excess paste could be considered a thermal buffer absorbing some of the heat and slowing down the heat transfer somewhat.

You have to think of the system CPU - thermal paste - sensor+fan as a mass-spring-damper system. If you were to remove the paste, you would remove the damper introducing the risk of oscillation (fan kicking in every few seconds under constant load).

You might ask: why use thermal paste at all?
The answer is pretty simple. It ensures that the whole surface area of the CPU is in contact with the heatsink. If you were to press metal against metal, the microscopic roughness would keep certain small spots without contact, thus introducing the risk of micro-scale failure (single or small groups of transistors overheating).

[curiousuburb]

Informative answer. </thumbsup>

[Kickaha]

Interesting - I hadn't considered before that it might not be optimal to ZOMG GET ALL THE HEAT OUT NOW but you're right, if it were hyper-efficient, the fans would be kicking on all the time, on high, then dropping off completely over and over again. This way it moderates the heat to a slightly slower ramp up, moderated level, and then back down. One smooth curve instead of a series of spikes. Hmm. Must ponder the effects of a longer operation at a higher temperature versus thermal fatigue brought on by periodic heating and cooling cycles.

[kscherer]

I don't know what all the fuss is about. Hell, all that extra heat can be put to good use:

Bed warmer;

Space heater;

Seat heater for December football games;

Etc.

But seriously, the guys in our service department have seen little issue with it. Makes one wonder…

I think Apple has a handle on this and, if it is consistent, then I bet their engineers have a reason for it.

[Dave Hagan]

I wonder about this. I am reading reports that they are getting hot to the point of just shutting off.

[julesstoop]

Can you point us to a (credible) source for these reports?

[Windowsrookie]

You can't get much more creditable than Engadget.

[kscherer]

Straight from Engadget:

"Reports of MacBook Pros getting crazy-hot because of misapplied thermal grease have been floating around for a year now, and it doesn't look like the recent bump to Santa Rosa has changed anything. A reader at the MacRumors forums noticed his week-old MBP getting a little hot, so he bravely decided to pull the unit apart to check things out and found what he estimated to be forty times too much paste applied to the logic board. There are also a few threads on the Apple support site with similar pics of MBP logic boards slathered in grease, so this doesn't appear to be an isolated issue. Apple hasn't commented yet, but seeing as it managed to cool things down last time with a firmware update, we wouldn't expect too much fanfare when this finally gets resolved."

Two questions:

1) Why are the vast majority of reviews so positive if this is such a HUGE issue? Heat, of course, is a problem with all modern laptops.

2) A "reader" at MacRumors estimated that there was 40 times too much paste. I wonder about his qualifications on the matter. Is he qualified to make that kind of an assessment?

Hmmmm…I guess I'm not sold on the paste thing quite yet. I'll wait and see what comes out of the service department.

[Banana]

Originally Posted by julesstoop So the excess paste could be considered a thermal buffer absorbing some of the heat and slowing down the heat transfer somewhat. You have to think of the system CPU - thermal paste - sensor+fan as a mass-spring-damper system. If you were to remove the paste, you would remove the damper introducing the risk of oscillation (fan kicking in every few seconds under constant load).

At first it made sense to me, but then I read Kickaha's post and realized that it surely could be possible to use less thermal paste for hyper-efficient dissipation of heat; you just need to make sure that fans will only start spinning only when the temperature has sustained at a certain point for long enough time, which would accomplish basically same thing, no? And of course it'd start spinning automatically whenever temperature exceeds a certain higher point to avoid crucial hardware failure.

[kscherer]

Originally Posted by Banana At first it made sense to me, but then I read Kickaha's post and realized that it surely could be possible to use less thermal paste for hyper-efficient dissipation of heat; you just need to make sure that fans will only start spinning only when the temperature has sustained at a certain point for long enough time, which would accomplish basically same thing, no? And of course it'd start spinning automatically whenever temperature exceeds a certain higher point to avoid crucial hardware failure.

And that's where the problem (or not) comes in. Apple is so fickle about noise that they don't want the fans blasting away. Noise is always an issue. Perhaps the better fix would be a more efficient fan design that moves more air with less noise? Am I a fool to think that Apple didn't already think of that?

[Banana]

No, it's always a good idea to start with the assumption that someone has already thought of it. Rarely do we think anything truly original.

I believe that the noise generated by fans is simply unavoidable, but only has became a big deal because everything else is now much quieter. (Is that right?)

So they've got a feat on their hand; how to move hot air from one place to another without creating a vacuum, which would in turn create wave which we commonly associate with as sounds. Electrostatic? Gravity? Magnetism?

They could have made it much easier if they just went with passive liquid cooling, but nooooo....

[Windowsrookie]

Let's slap a 12" x 5" copper Heat sink on the top of all the MacBooks.

[Banana]

Yeah, and watch various courts' backlog spill over with suits of 3rd degree burns and skin cancer to palms, fingertips and wrist from the proud owners of CopperBook!

That's a solid thinking there, WR.

[Windowsrookie]

Originally Posted by Banana Yeah, and watch various courts' backlog spill over with suits of 3rd degree burns and skin cancer to palms, fingertips and wrist from the proud owners of CopperBook! That's a solid thinking there, WR.

Why would you put your palms, and finger tips on the top of the CopperBook? That's where the plasma lasers are. Duh.

[kscherer]

I wonder why there are no vents in the front of the system. Seems like airflow, front to back, would be more efficient than the current back to back thing.

[kscherer]

Originally Posted by Windowsrookie Why would you put your palms, and finger tips on the top of the CopperBook? That's where the plasma lasers are. Duh.

Yeah! Duh!

[julesstoop]

Originally Posted by Banana At first it made sense to me, but then I read Kickaha's post and realized that it surely could be possible to use less thermal paste for hyper-efficient dissipation of heat; you just need to make sure that fans will only start spinning only when the temperature has sustained at a certain point for long enough time[...]

It doesn't work like that, really.
During normal operation, with a few do a dozen apps open and some hundred processes running, actual processor load changes from millisecond to millisecond. So when heat transfer were to be as efficient as possible, there would be very little dampening and the temperature will exceed your preset safety threshold very oftenly.

At the moment the fan kicks in though, a few milliseconds later, core temperature could already have dropped all by itsself due to lower processor load in the past few frames. This holds especially true for modern speed throttling notebook CPU's. The net effect is fans kicking in and switching off much more often and more vigorously. This would cause a much larger mechanical stress on the fans themselves while introducing more noisy and 'nervous' operation.

I'm pretty sure Apple's has take all of these considerations into account and decided that a fairly large heat buffer is the best solution. Under normal use it evens out the heat dissipation allowing for the fans to operate at a more or less constant 1000 rpm thus minimizing mechanical stress, noise and 'nervousness'. Under heavy load the fans throttle up - quite gently - to dissipate the larger heat.

The only real solution to the hot MacBook problem would be to use more efficient fans that dissipate more heat at the same rpm (noise/stress). Then Apple could either opt for slower fans or a slightly colder notebook (on average).

[Banana]

Julesstop, perhaps I wasn't clear in my first post;

I suggested that fans only kick in only if the temperature stays high for long enough time. You say there's spikes of high temperature in milliseconds? Let's make fans turn on *only* when the temperature stay above X for 2 seconds, but not if it's less than that. The key here is sustained elevated temperature.

Or is it still infeasbile?

[julesstoop]

I'll stop

I understand what you mean, and I believe it is exactly what Apple does at this moment. However, instead of using your proposed timing 'rule' in sofware or on a microcontroller, they use a few microliters of thermal paste. Much safer. If you were to remove this physical buffer and trust on your solution, they would have to ensure at the same time that the circuits controlling the thermal system operate within much smaller margins of tolerance.

Besides this, the main disadvantage I mentioned still stands. If you can't absorb enough heat ( = if your dissipation system has too small of a heat capacitance) Small spikes - due to fact that CPU-load often changes very quickly - still would more often push the temperature over your proposed 'override threshold' thus triggering the fans to kick in to save the processor from damage.

[kscherer]

It wouldn't seem that processor fluctuations over the course of milliseconds would have an affect on the fans. I'm no engineer so I don't know, but I know how long it takes for a stove burner to heat up, and then how long it takes to dissipate that heat. I can put more or less energy through the burner and it's response is not immediate, regardless of how quickly I adjust the knob. It seems that a heat sink would act in a similar fashion. The processor gets warmer and transfers its heat to the sink, which fluctuates at a slower pace than the processor itself. Am I wrong?

[Banana]

Julesstop- Cool. Makes sense. Thanks for taking time to explain it.

[julesstoop]

Originally Posted by kscherer It wouldn't seem that processor fluctuations over the course of milliseconds would have an affect on the fans. I'm no engineer so I don't know, but I know how long it takes for a stove burner to heat up, and then how long it takes to dissipate that heat. I can put more or less energy through the burner and it's response is not immediate, regardless of how quickly I adjust the knob. It seems that a heat sink would act in a similar fashion. The processor gets warmer and transfers its heat to the sink, which fluctuates at a slower pace than the processor itself. Am I wrong?

No. You are right. The idea is though that if your heat dissipation from your cooling system is (much) too large and the heat capcaitance of your cooling system (much) too small, they (processor fluctuations over the course of milliseconds) would.
On the other end of the spectrum you have a system consisting of a huge (much too heavy for a laptop) heatsink that just very slowly heats up a (condirably) large mass and is cooled by radiation. In between there are all sorts of more practical solutions.

I'm just saying that it's all about striking a balance and Apple made some conscientious design choices in finding theirs.