How to Improve MacBook Pro Performance and Thermals
bsid.io
bsid.io
I have fixed multiple overheating Macbook pro's by simply cleaning out the dust buildup underneath the rubber pad, between the fan housing and the heatsink. The airflow compresses the dust into an almost solid piece (like felt), which blocks the airflow. You can usually pull it out as one piece. All you need is a pentalobe screwdriver, a pair of tweezers and maybe some compressed air.
Especially if your laptop often gets into contact with fabrics (on your lap or in bed), you should do this cleaning once a year.
TLDR: Remove the screws on the bottom. You need a special screwdriver and keep in mind that (at least on the <=2015 models afaik) there are different size screws used around the hinge.
I've used mine to take apart every conceivable consumer electronic, and through social bits for Nintendo, Apple, etc are priceless.
I also got the magnetic mat, a Christmas gift from my wife, and when disassembling things with lots of screws is useful for drawing out the device for their location.
P5 Pentalobe screwdriver: https://en.wikipedia.org/wiki/Pentalobe_security_screw
1) Suction cup.
2) Screwdriver kit that contains pentalobe and torx bits. You don't need to buy some expensive kit as you can get them cheaply from any hardware store.
I think Apple finally won with making their computers impossible to open.
My MBP seems to always have the fans on. I do occasionally run compressed air across the vents (from the outside) but decided to open it up thanks to this thread.
As I was cleaning, I noticed that when blowing compressed air into the vents, one of the fans was not turning (indicating no air reaching?) I grabbed some precision tweezers and pulled this out: https://i.imgur.com/aIPSRdi.jpg
That said, blowing compressed air at a fan will spin the fan much faster than what it was designed for. So you still risk breaking the fan mechanically. Always hold the fan when blowing compressed air at it.
Enjoy your now silent and non-thermal-throttling laptop =)
Another thing also worth looking at is the process list and the Chrome process manager, sorted by CPU. I have no idea what, for example, sueddeutsche.de does - leave a page of them open in the background and the ads will run amok.
Another thing that can cause high CPU load is memory pressure resulting in massive swapping. If 2GB RAM are already set aside for the Docker VM and another 2GB for the core OS, you're left with 12GB RAM and that is not very much.
I can see how it might be problematic in a headless environment, but for most people most of the time it's probably a useful feature.
omg
For people working on 64GB+ workstations this is no problem but Apple only supports 32GB in the newest lineup of MBPs.
IMHO you are wildly overestimating the RAM use of those programs, aside from the others, as they normally run (sure - I can get Excel over a GB running scripts or manipulating a large file).
- Slack 1.08GB
- Safari (11 tabs) 4.50GB
- Chrome (5 tabs) 2.17GB
- Outlook 1.13GB
These are ludicrous numbers and, frankly, the only reason why I need 16GB+ machines.
German leather?
http://attach.mobile01.com/attach/200605/mobile01-514da8fc7f...
http://attach.mobile01.com/attach/200605/mobile01-e89d3804df...
Here's a picture from Intel's site showing the correct amount:
https://www.intel.com/content/dam/support/us/en/images/proce...
That's for a processor with a large heatspreader on top, so a small bare-die one like those in a laptop should have proportionally less.
Exactly. The thermal conductivity of air is 0.026 (W/m/K), your average thermal paste (non-metal) is at 2. For manufacturers it's way better to guarantee uniformly meh conductivity by using too much thermal paste than risking air bubbles by using too little. Most manufacturers will lean this way when they don't just use those POS thermal stickers.
Same with average versus silver: silver provides a gain but a limited one, and it's conductive so you have to make very sure not to leak any onto circuits or you risk a short. That's why manufacturers don't bother.
What about any excess TIM off the edges drawing heat away from the heat pipes/cold plate?
We never could figure out why they had so much in their manuals.
I believe the first two come from the MacBook (not pro) early 2006 service manual, since the colour of the paste is white instead of the grey of later manuals which show (only slightly) smaller amounts.
So who do companies send to defend themselves?
Caveat: at least in the US. What shenanigans manufacturers can and can't pull in other countries is beyond my knowledge.
The problem was that to remove the bottom, I'd have to destroy a warranty void sticker.
I emailed them and they replied in no time at all confirming that removing the sticker to perform the upgrade was not an issue and would not affect my warranty.
I'm in the UK.
It is my understanding that here in Europe, the same rules will be honoured with respect to warranty void stickers so long as you don't cause damage while voiding the sticker to perform whatever it is you're doing.
It is significantly more complex than on previous generations. There are a lot of very tiny, very fragile cables which are very easy to break.
It gets a 1/10 rating for repairability for a reason.
In my experience HP pushes the thermal boundary more than most (they generally offer higher TDP options for custom builds than their competitors without tailoring for it).
Some people take this to the next level and decap the processors and then use liquid thermal compound instead of the paste. That probably voids your warranty.
Those are nice if you develop a spreadsheet with some VBA now and then, but once you actually need the power if you're developing enterprise applications you can forget it. Under any serious load they overheat and clock the processor down from 3Ghz+ to 2Ghz.
They're probably deliding the CPU (removing the integrated heat spreader (IHS) to put the compound on the die directly), not decapping it. I can't imagine a CPU surviving the latter.
The CPU in the article is lid-less so that's not a concern.
And some CPUs have a soldered heatspreader, this allows for much better heat transfer to the IHS but means delidding is destructive (though way less necessary to be fair). Ryzen chips are soldered.
> This only works efficiently, though, if a very thin layer of thermal paste is applied between CPU and heatsink in such a way that minimises the chance of creating “air bubbles” (air has a bad thermal conductivity).
The issue here isn't really air bubbles. It's true that air has very poor thermal conductivity, but the real issue is that the thermal paste itself has very poor thermal conductivity compared to the solid copper heatsink. The paste is really just supposed to fill in any microscopic pits in the surface to get the copper in good contact. People go to great lengths to make the surfaces as smooth as possible, going as far as to hand lap their heatsinks to a mirror polished finish.
Apply too much thermal paste and it'll just squeeze out of the edges, which is merely inconvenient as long as the paste is reasonably non-conductive. Apply too little paste and you'll get air gaps, which have terrible thermal conductivity. If in doubt, just spread the paste to ensure that it fully covers the contact area.
https://www.gamersnexus.net/guides/3346-thermal-paste-applic...
Even putting way too much makes 0 difference.
OP's problem was probably dust and ageing of paste. Plus interestingly there was less power used in test AFTER so I guess that contributed most.
Without going these exocitic routes, the goal of using TIM is to fill up gaps (due to surfaces not being ideal) to replace no thermal conductivity in those places with some thermal conductivity. What you do not want to do, however, is put TIM where there was already good metal-on-metal contact. That’s why too much TIM is definitely bad.
I'm skeptical of this, I'd be surprised if it's physically feasible to get literally perfect contact between the surfaces, micro-gaps are too real.
If the author has results from after the cleaning, they would help paint a clearer picture.
I have the same model. Had overheating problems. Took it to the Apple store. They fixed it and refurbished it for me. Even replaced the bottom case because of a couple of dents. It’s like a brand new machine. Charged me $0. This is why I love Apple and will never switch.
I don’t think that thermal grizzly products are a bad place to start looking. I haven’t tried them myself, but GamersNexus are advertising/recommending them and I do trust them.
One should be very careful if that stuff can leak.
[0]: https://www.youtube.com/watch?v=k919f7Qi4es&vl=en <--random video
you should. kryonaut and conductonaut are both great. However I'd not recommend liquid metal (gallium) for daily use in a laptop with low tension force. I use liquid metal on delidded ones and kryounat elsewhere.
Most Apple spec intel CPUs have no die-side caps, you’re going to be fine.
Of course I don't remember where I heard this, nor do I have any sources...
The gain is low, it increases costs, and better thermal pastes are electrically conductive so applying them industrially is a more complex endeavour (as you absolutely can't risk leaking/spreading them to any component). The risk/reward just isn't there for manufacturers.
Or at least they last longer than 2 - 3 years?
Copper is OK, but Gallium does alloy slightly and you may need to reapply after a year or so, not a problem with nickel.
Normal shock levels range in 2-5g maximums when you are carrying around your laptop in a briefcase, backback, by hand, etc. This is tested on a shaker table in the reliability lab at 10ms classical half-sine shock. They also do RMS, sine-burst, etc waveforms to ensure that the laptop is going to survive normal use.
The capillary action between gallium liquid sandwiched between the copper heatsink and the die is higher than any of the shocks or external forces you can experience in normal use. Higher by an order of magnitude. Also, the die surface has extremely high surface energy - it wants to wet badly.
It is not going to spill all over the motherboard.
UNLESS - 1) You apply more than necessary. 2) The heatsink is not secured correctly 3) During the application process, you spill all over the motherboard (what Linus @ Linus Tech Tips did in one of his videos).
There is literally no need to worry - I wish I can disclose where I work to give some credibility.
I have used liquid metal thermal compound before. It is not hard to apply correctly and works great.
The vast majority of thermal paste is non-conductive, including AS5.
Clear coat around the die covering any components (resistors and caps) and less is more. Seriously, a tiny drop is all you need.
I understand that it doesn't make the system look better to a low-information buyer. So Lenovo aren't doing it. But Apply seem motivated to add small costs in exchange for moderate (if invisible) benefits.
Now imagine doing this in a factory pipeline. I don't think this would scale.
Print shops can do this kind of work, too (precise amounts, sub-millimeter accuracy, irregular surfaces, almost any kind of coating, etc.).
https://www.electronics-cooling.com/2007/11/reliability-test...
https://scotchi.net/2017/03/04/considering-reapplying-the-th...
With that, my 2011 2.7 GHz i7 Macbook Pro continues to be completely usable even for C++ development (also has a 1 TB SSD and 8 GB of RAM), and hopefully will get me through to a point that Apple releases a laptop that I actually want to buy. If they haven't released a Macbook with a better keyboard by the time this thing dies, it'll be my last Apple laptop.
But that should not be case for idle, so lower fan speeds (or no fan) seems logical
http://en.wikipedia.org/wiki/Poole–Frenkel_effect
https://forums.anandtech.com/threads/effect-of-temperature-o...
The effect is so pronounced that a watercooled computer running under full load will typically use less power than an identical aircooled computer, because the lowered silicon current leakage more than makes up for the power consumption of the pump and additional fans.
Not on this MacBook Pro.
This reminds of Canadian CBC National investigation into how instead of identifying problems, Apple was caught misdiagnosing and pushing purchasing new products. https://youtu.be/_XneTBhRPYk
I don’t know if you’d have to pay, I suspect you would, but an independent repair shop is likely going to care more about you and your nuanced issue.
(also, does it apply to a MacBook Air?)
Or better, find a friend to do it for you for a beer. This is just basic maintenance, so very low skill kind of thing.
What other possible ways to clean up the dust without opening my machine?
Would Apple reapply the paste and clean up the dust if I had other form of repair, assuming I am still under AppleCare+?
Except those new fancy graphite thermal pads ( thermal grizzly sells some ). Those are generally worse than good paste but are pretty good and they have the added pro of being reusable and keeping their thermal property forever unlike paste. In a laptop that is subject to tension stress they may provide better cooling over time.
Read the warnings and disclaimers and inspect the source code:
https://github.com/sicreative/VoltageShift
newer MBPs (2016+) appear to be locked down.
works like a charm, even with multiple editing sessions on many monitors.
edit: i misread the price of the screwdrivers. I still recommend a cooling pad. High power draw is usually when I'm using extra monitors at a desk anyways.