Cooling the RMBP. A Mission.

[Ryan1524]

I've been giving this some thought myself.

My 2008 MBP used to be able to do 3-4 Safari Windows with 10 tabs each, while playing iTunes, with Word running in the background, and a few other misc tasks running.

Now, on Lion 64bit, it starts bogging down with ONE chrome window with a handful of tabs while playing a a single video.


Is this just the OS progression going too fast for the hardware? Seems odd since I'm doing non-complex tasks. All I know is that my MBP has been constantly HOT.

 

[whitedragon101]

I use the zalman ZM-NC3000S cooler and its awesome. My 17" MBP lives on it. It has a single huge 120mm fan so it is quiet but pushes a good amount of air. Even when you turn the fan off it keeps the bottom vented and the laptop at a nice angle for typing and using the trackpad. Also you can put it on your lap without cooking yourself.

http://www.zalman.com/ENG/product/Product_Read.asp?idx=398

 

[chevalier433]

If you work in a desk go for Mac Pro if you are on the field go for RMBP and dont worry a friend have an old MBP core duo for Avid pro video editing and after effects all day long and still running strong.

 

[MCAsan]

Operate it within design targets and it should last for years.

Line voltage: 100V to 240V AC
Frequency: 50Hz to 60Hz
Operating temperature: 50° to 95° F (10° to 35° C)
Storage temperature: –13° to 113° F (–24° to 45° C)
Relative humidity: 0% to 90% noncondensing
Maximum operating altitude: 10,000 feet
Maximum storage altitude: 15,000 feet
Maximum shipping altitude: 35,000 feet

 

[monksealpup]

IR Images

I just got my rMBP in yesterday, so I took some IR photos today at work; this seems like a good thread to share them in.

This was taken with the computer sitting more or less at idle, with Parallels running Windows 7 (also at idle, fwiw). That's important to note since Parallels fires up the discrete graphics chip.

http://imgur.com/I90ko

http://imgur.com/yRO6i

http://imgur.com/0gW8l

 

[RealEyes]

She's on fire!

My rMBP stays around 50~55C on the GPU diode while connected to a simple external display with HDMI...

What gives?

 

[Halon X]

Throw a desk inside a walkin freezer and you'll be good to go!

 

[w00t951]

First, heat will decrease the lifespan of computer components. It's not a matter of decades, but of years. Proven fact, especially if the user is taxing their components to the limit every single day.

My old 2008 MacBook Pro was replaced in early 2011 by the Mac in my signature. I'm a heavy gamer (at least an hour a day, tenfold that during the summer), so I'm taxing everything a video editor would be (CPU, GPU, RAM, and HDD/SSD). I can tell you that my Mac (not Retina) gets extremely hot while gaming. An ordinary laptop cooler will not do much, but the one I bought (Cooler Master Infinite Evo) makes a huge difference. It blows air onto the CPU and GPU of the computer, and can result in a 7C difference.

The computer will run games at 60FPS for the first 5-10 minutes, then it throttles down. Running Windows software like ThrottleStop results in temperatures of 110C, which is absolutely ridiculous.

I suggest running your video editing machine on a cooling pad with an open back (check out the one I have) and in a cool room. The A/C actually makes the biggest difference in the temperature - more than 10C.

Finally, the thermal paste. Despite what posters say about Apple picking the best paste for their machines, this is definitely false. In all of my PC building experience, I know that you are supposed to apply a very thin veneer of paste over the CPU die. I use Arctic Cooling MX-4, while my friends prefer Arctic Silver Formulae 5. Anyway, the stuff that Apple uses is like Larry King shavings, and is dolloped on in massive globs that actually insulate the CPU/GPU, not assist in heat dissipation. This is probably one of the best things you can do for your Mac, as 99% of Apple computers come with terrible paste application. Check out the Retina teardown guide here. The Retina Mac is much easier to re-paste than the normal Mac, as the heat sink is directly accessible. My Mac requires a full logic board teardown, which takes an hour to do properly.

Good luck!

 

[AzN1337c0d3r]

First, heat will decrease the lifespan of computer components. It's not a matter of decades, but of years. Proven fact, especially if the user is taxing their components to the limit every single day.

If it's a fact that you should cite your sources. Electromigration doesn't set in for decades at elevated temperatures and that is the primary form of heat-induced damage in modern silicon. But what do I know, I just have a MSEE specializing in VLSI design :\

An ordinary laptop cooler will not do much, but the one I bought (Cooler Master Infinite Evo) makes a huge difference. It blows air onto the CPU and GPU of the computer, and can result in a 7C difference.

I find it hard to believe that blowing air over a flat piece of metal that is not even in contact with the heatpipes will provide any significant cooling.

The computer will run games at 60FPS for the first 5-10 minutes, then it throttles down. Running Windows software like ThrottleStop results in temperatures of 110C, which is absolutely ridiculous.

Umm, the 2011 Macbook Pro which have Sandy Bridge processors which have a TjMax of 100C. I dont know how you got readings of 110C but something is definitely very wrong. Your processor should stop turbo and possibly even clock itself down as it hits 100.

I agree with the bit about thermal paste though. Apple does a really poor job of assembling them properly.

 

[w00t951]

1) ....

2) Your MSEE in VLSI doesn't mean you're an expert on thermal design - ergo, it matters not one iota what you believe. Anyway, have you ever wondered why the entire bottom of the MacBook Pro turns into a skillet under load? Maybe because that's where a lot of heat is stored? Perhaps it would benefit the user to relieve the bottom plate of that heat?

3) ThrottleStop is software that bypasses Intel thermal management and allows the manual setting of CPU clocks, regardless of temperature.

 

[auhlixer]

I just got my rMBP in yesterday, so I took some IR photos today at work; this seems like a good thread to share them in.

103F isn't bad at all

 

[AzN1337c0d3r]

2) Your MSEE in VLSI doesn't mean you're an expert on thermal design - ergo, it matters not one iota what you believe.

We're talking about silicon failure processes, not thermal design, which does make me an expert compared to you, unless your credentials are significantly better than mine.

And you still haven't cited the source of your "fact".

Anyway, have you ever wondered why the entire bottom of the MacBook Pro turns into a skillet under load? Maybe because that's where a lot of heat is stored? Perhaps it would benefit the user to relieve the bottom plate of that heat?

The bottom of the Macbook Pro hardly turns into a skillet, my morning cup of coffee gets hotter than that. If it actually approached anywhere near dangerous temperature, you'd get a serious 3rd degree burns.

ThrottleStop is software that bypasses Intel thermal management and allows the manual setting of CPU clocks, regardless of temperature.

You can't disable the throttling imposed by Intel with respect to TjMax by modifying software. The processor automatically asserts PROCHOT# when distance to TjMax reaches 0. It is a NEVER EXCEED operating range which cannot be disabled by software. Thus you getting 100C+ temperatures is [...].

 

[KevinC867]

The battery will definitely age faster if subjected to too much heat. (http://batteryuniversity.com/learn/article/how_to_prolong_lithium_based_batteries)

The processor and GPU generate the most heat and they are only spec'ed to operate up to certain temperatures. However, there are plenty of heat sensors inside the Mac and if the system sees the temps going too high, it will scale back the operating frequency to reduce the heat generated. (http://www.anandtech.com/show/6023/the-nextgen-macbook-pro-with-retina-display-review/12) Lower performance is definitely a possible result of temperatures that are getting too high.

The best thing to do is install software like HardwareMonitor and see what kind of temperatures you are getting at the CPU and Battery. That will also allow you to see how much useful improvement you get from an external cooling system.

 

[AzN1337c0d3r]

The battery will definitely age faster if subjected to too much heat. (http://batteryuniversity.com/learn/article/how_to_prolong_lithium_based_batteries)

OP is using machine as desktop replacement. Battery life isn't important to him. I should add that the extra draw on the machine if you're using a USB-powered cooler is also detrimental to battery life through extra cycling.

That will also allow you to see how much useful improvement you get from an external cooling system.

Several posters here seem to be claiming that an external cooler will yield significant differences in cooling perhaps due to heat conduction and radiation to the laptop chassis. I should point out the heatsinks in the Macbook Pro are only connected through the chassis through some very thin screws that make a poor thermal conductor.

The primary means (>95% for systems of this caliber) of cooling in systems of this size is through forced convection.

Since laptop coolers do not force more air through the heatsinks, why should the temperature lower?

 

[KevinC867]

OP is using machine as desktop replacement. Battery life isn't important to him. I should add that the extra draw on the machine if you're using a USB-powered cooler is also detrimental to battery life through extra cycling.

I see your point, but who wants a weakened battery in a laptop, even if it is *primarily* used while plugged in. I think an external cooler might be of *some* help with this.

Several posters here seem to be claiming that an external cooler will yield significant differences in cooling perhaps due to heat conduction and radiation to the laptop chassis. I should point out the heatsinks in the Macbook Pro are only connected through the chassis through some very thin screws that make a poor thermal conductor.

The primary means (>95% for systems of this caliber) of cooling in systems of this size is through forced convection.

Since laptop coolers do not force more air through the heatsinks, why should the temperature lower?

I agree for the most part. However, there is a clearly some heat conduction path which causes the bottom plate and chassis to get toasty. Reducing the temperature of those components will cause some increase in the amount of heat drawn out that path. But, you are right - the primary cooling path for the CPU and GPU through the heat sinks and radiators won't be improved much at all. That's probably what will be seen if the temperatures are monitored.

 

[w00t951]

We're talking about silicon failure processes, not thermal design, which does make me an expert compared to you, unless your credentials are significantly better than mine.

And you still haven't cited the source of your "fact".




The bottom of the Macbook Pro hardly turns into a skillet, my morning cup of coffee gets hotter than that. If it actually approached anywhere near dangerous temperature, you'd get a serious 3rd degree burns.



You can't disable the throttling imposed by Intel with respect to TjMax by modifying software. The processor automatically asserts PROCHOT# when distance to TjMax reaches 0. It is a NEVER EXCEED operating range which cannot be disabled by software. Thus you getting 100C+ temperatures is donkey sh|t.

[...]

Look up ThrottleStop. You can do whatever the hell you want with it. Of course, I'm under the assumption you can't figure out how to utilize Google Search, so here it is.

 

[AzN1337c0d3r]

3) Look up ThrottleStop. You can do whatever the hell you want with it. Of course, I'm under the assumption you can't figure out how to utilize Google Search, so here it is.

And nowhere in that article does it claim the ability to disable TCC or PROCHOT#, just modify the voltages and frequency WITHIN the thermal limits.

I suggest you learn to Google Intel's datasheets, which are a much more definitive source on what you can or can't do.

P.S. Still waiting for the source of your "fact".

 

[w00t951]

And nowhere in that article does it claim the ability to disable TCC or PROCHOT#, just modify the voltages and frequency WITHIN the thermal limits.

I suggest you learn to Google Intel's datasheets, which are a much more definitive source on what you can or can't do.

P.S. Still waiting for the source of your "fact".

I can't attach a file from my phone, so here's a link to my Dropbox.

By the way, the inside of a Mac isn't vacuum - the heat doesn't travel through just the screws. I would have thought someone like you would have guessed that.

 

[AzN1337c0d3r]

I can't attach a file from my phone, so here's a link to my Dropbox. I guess you should give up that prized MSEE in VLSI, huh? Suck it, you arrogant prick. That's right, give me a warning. Report this post. He deserved it.

By the way, the inside of a Mac isn't vacuum - the heat doesn't travel through just the screws. I would have thought someone like you would have guessed that.

You're an idiot. HWinfo64 is KNOWN to be inaccurate with reading the DTS.

A little clue: Intel spec sheets say to calculate temperature by subtracting the value in DTS register from TjMax. Since that register can't ever be negative, the temperature can't ever get above TjMax.

Try CoreTemp64. It is the only program that I know of that sticks to Intel's specs when reading the DTS.

Oh and for all intents and purposes of calculating heat dissipation, you can treat the inside of a Mac as a vacuum except the heat pipes and the heatsink, if you've taken any basic physics at all.

Here's a short lesson: There's 4 methods of heat transfer:

1. Conduction
2. Convection
3. Radiation
4. Advection

Still air is terrible thermal conductor, so 1 is basically 0.
The object is not glowing, so 3 is basically 0.
Since we are not dealing with air rising or moving between temperature zones on it's OWN, advection is synonymous with convection. We pretty only talk about advection when it comes to climate (or other phenomenon related to horizontal movement of mass.

So that leaves us with convection. What does convection come from? Moving air. ie. fans blowing air through heatsinks.

Insert more coins to continue.

 

[w00t951]

1) Your sample size of 1 (or was it 2?) definitely means that those temperatures are incorrect. Also, that's my friend's MacBook Pro (it's noticeably different from the one in my signature). I'll ask him to install CoreTemp64.

2) Anyone who pushes their Mac knows that the bottom gets ridiculously hot. Plus, if you, the genius, has ever popped the bottom of your Mac (or done a simple Google search), you would know that there is perhaps 0.25mm of "air" between the bottom plate and the CPU/GPU. The fans are actually touching the bottom plate, and the back portion of the bottom plate acts as the final stretch of exhaust pipe.

 

[AzN1337c0d3r]

1) Your sample size of 1 (or was it 2?) definitely means that those temperatures are incorrect. Also, that's my friend's MacBook Pro (it's noticeably different from the one in my signature). I'll ask him to install CoreTemp64.

2) Anyone who pushes their Mac knows that the bottom gets ridiculously hot.

Hot doesn't mean it dissipates heat well. The space shuttle tiles reach up to 1260C upon reentry, yet the thermal conductivity of such a tile is basically 0. You could be holding the other side in your hand.

Plus, if you, the genius, has ever popped the bottom of your Mac (or done a simple Google search), you would know that there is perhaps 0.25mm of "air" between the bottom plate and the CPU/GPU.

If you had 0.25mm of air between your CPU and your heatsink, your processor wouldn't operate and if you disabled the safety mechanism it would burn. That's how bad a thermal conductor air is. Remember that Space Shuttle tile I linked above? Did you read it? What is 90% of it made of? Oh that's right, air!

Btw, the gap is more like, 5mm. You couldn't move enough air to cool the processor if you only had a 0.25mm gap between the heatsink and the plate.

The fans are actually touching the bottom plate

No they dont. Your entire bottom plate would act as a resonant structure for the fan and make it loud as hell if it did

the back portion of the bottom plate acts as the final stretch of exhaust pipe.

But that doesn't say anything about it actually getting rid of heat, just that it's absorbing exhaust heat, which doesn't do ANYTHING for the temperature of your processor. In fact it maybe contributing to the problem because it preheats incoming air.

 

[w00t951]

1) Ceramic tiles are very different from aluminum.

2) Look at this iFixit teardown. It's more like 0mm of air, and whatever thickness of motherboard there is. This also applies to the fans - if none of the computer is touching the bottom, why did they put insulation on the inside of the bottom plate? Besides, it's obviously not the same scenario as a heatsink. A heatsink is required to dissipate immense amounts of heat - not so with the bottom of a MacBook Pro.

 

[AzN1337c0d3r]

1) Ceramic tiles are very different from aluminum.

Can you read?

Composition: 90% air, 10% silica fibers.

None of which qualify as "ceramic".

2) Look at this iFixit teardown. It's more like 0mm of air, and whatever thickness of motherboard there is.

You didnt link to any so I can't see what you're referring to.

But I did take the liberty of looking at the teardowns for the mid-2010 and late-2011 MBP and none of them show with any evidence at all the gap is 0mm. Furthermore, having owned both those models and taken them apart, my recollection is that I dropped them a couple of times and made a nice dent in the bottom plate. If there was no gap, the corresponding dents should have shown up in the PCB too.

This also applies to the fans - if none of the computer is touching the bottom, why did they put insulation on the inside of the bottom plate?

Perhaps it's not insulation? Perhaps it's vibration dampening material? Maybe to stop the parts of the PCB and fan from touching the plate and making a nice flapping sound?

Besides, it's obviously not the same scenario as a heatsink. A heatsink is required to dissipate immense amounts of heat - not so with the bottom of a MacBook Pro.

No, moving air AND a heatsink is required to dissipate immense amount of heat. Isn't that the definition of convection I made 2 posts ago? Moving air?

Just having a heatsink results in conduction, which is complete **** with still air, as in the space shuttle thermal tiles.

The bottom plate of a Macbook Pro qualifies as neither.

P.S. Still waiting for the citation of the source of your so-called "fact" that electronics only last years under elevated temperatures.

P.S.S. While we're at it, Google (with the world's largest farm of disk drives) found that failure rates have no correlation to temperature, and what do hard drives contain? Electronics.

 

[leman]

So what you're saying is that working with the rmbp day in, day out on heavy HD projects, sometimes in prores 422, sometimes 4444, rarely even 4/5k just sitting on a desk would be cool enough? Do you use the computer for such intense activities as a main computer?

I have been abusing my 2009 MBP like that for more then 2.5 years now. Seems to do fine still. And the rMBP has much better cooling.