What's the point if I may ask? A useless number?
Gaudeix del teu ordinador i oblidat de fer un milió de benchmarks. Simplement no val la pena ...
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
2020 Air Heatsink Modification Thread
- Thread starter ilikewhey
- Start date
- Sort by reaction score
What's the point if I may ask? A useless number?
Gaudeix del teu ordinador i oblidat de fer un milió de benchmarks. Simplement no val la pena ...
Isn't 1080p also using the same codec?
So if the processor load is caused by higher resolution then VP9 codec is not a factor. So why do people say YouTube in Chrome on Mac is causing problem but not Safari?
[QUOTE = "KPOM, post: 28381730, member: 505893"]
1080p is fine in Chrome. Remember, 4K uses 4x as many pixels as 1080p.
[/ QUOTE]
So if the processor load is caused by higher resolution then VP9 codec is not a factor. So why do people say YouTube in Chrome on Mac is causing problem but not Safari?
[QUOTE = "KPOM, post: 28381730, member: 505893"]
1080p is fine in Chrome. Remember, 4K uses 4x as many pixels as 1080p.
[/ QUOTE]
[QUOTE = "alstrike, publicación: 28381784, miembro: 640697"]
¿Cuál es el punto si puedo preguntar? ¿Un número inútil?
Gaudeix del teu ordinador i oblidat de fer un milió de benchmarks. Simplement no val la pena ...
[/CITAR]
si que val la pena !! tinc una mica mes de potència per tenir menys temperatura !! Ademas m'ho él passat pipa !!! jajaja
[automerge]1587325274[/automerge]
Guys final conclusions of the 2019 MBA:
The best solution is to add those copper plates to better dissipate the heat to the refrigerator and cover all the holes where the air can escape. It is not worth transmitting heat to the back plate, nothing is gained and it gets hot. The last thing I have done has been a rubber semicircle around the turbine so that no air passes behind the fan.
Now seriously it goes fast of noses! not only because the performance tests say so, but also because it works cooler with all applications. Right now without a fan and writing this marks me 38º, I just finished seeing 90º or 100º ...
I think that more can not be removed at least without breaking something and losing the warranty.
I leave you the last test.
A hug to all! and lots of health !!!
PS: you can compare the data that it gives of origin since I don't have it.
¿Cuál es el punto si puedo preguntar? ¿Un número inútil?
Gaudeix del teu ordinador i oblidat de fer un milió de benchmarks. Simplement no val la pena ...
[/CITAR]
si que val la pena !! tinc una mica mes de potència per tenir menys temperatura !! Ademas m'ho él passat pipa !!! jajaja
[automerge]1587325274[/automerge]
Guys final conclusions of the 2019 MBA:
The best solution is to add those copper plates to better dissipate the heat to the refrigerator and cover all the holes where the air can escape. It is not worth transmitting heat to the back plate, nothing is gained and it gets hot. The last thing I have done has been a rubber semicircle around the turbine so that no air passes behind the fan.
Now seriously it goes fast of noses! not only because the performance tests say so, but also because it works cooler with all applications. Right now without a fan and writing this marks me 38º, I just finished seeing 90º or 100º ...
I think that more can not be removed at least without breaking something and losing the warranty.
I leave you the last test.
A hug to all! and lots of health !!!
PS: you can compare the data that it gives of origin since I don't have it.
Attachments
Last edited:
If you put a better heat sink on the CPU, I’d worry about those mosfets beside it. All the heat from the CPU flows over those mosfets on the way to the fan. I think a better heat sink will wick away more heat into the case, while also keeping fan speeds lower. This could put more heat over those mosfets. Those mosfets power the CPU so they already run hot when in turbo boost. I bet this is why there seems to be turbo boost limits that aren’t temperature dependent.
I've been following this thread for a couple days now and you inspired me to do something similar.
What were the dimensions of the copper shim you used for the recessed part of the heatsink? I'm guessing it was 20x20mm but how deep was the recess on the heatsink and/or how thick was the shim you used? 0.5mm?
And what do you think would be the best thermal paste to use for this mod? Arctic seems like a popular brand, but is there a specific kind that's best/recommended?
It seems clear now that Apple (inadvertently..?) made a subpar heatsink for the CPU/GPU. This project looks like an inexpensive and easy way to boost performance 10% while also protecting the internals from heat wear!
Just so I understand properly, you're doing this on a 2019 Macbook Air with the i5-8210Y processor, not the 2020 with the dual core i3-1000NG4, the quad core i5-1030NG7 or i7-1060NG7? I assumed because of the thread title you were doing these experiments on the 2020 MBA heatsink.
I've got the Macbook Air 2019. It a) doesn't have the same heat sink design as the 2020, and b) doesn't have the same TDP vs. cooling issues some users are experiencing with the 2020 i5 and i7 models.
I'm glad your experiments have led you to the conclusion that a copper shim and thermal heat paste is the best way to go, and I admire your efforts, experimentation and skill.
But for everyone reading this wanting to replicate it, this mod was performed on a 2019 MBA, not a 2020. The dimensions of the 2020 MBA heatsink are different, as are all three CPU options (the i3, i5 and i7 are all built on a different architecture, with different very Thermal Design Power benefits and constraints). The results you get will not be the same.
I'm all for modding and getting the most out of computers, but the 2019 Air didn't have any issues with thermals or emissions that I've been aware of in my day-to-day use, or that I've read about in reviews. This mod for the 2019 Air heatsink looks great, and I might try it myself out of warranty, but... only if I was really bored.
My 2019 idles at 30-40C and takes far longer for the CPU to hit 100 (because it's not using the 10th generation Ice Lake chip's boost parameters). It took an hour's videoconferencing yesterday for my 2019 MBA's fans to hit 8000rpm, vs less than five minutes on my 2020 i5 MBA. They are very, very different.
It'd be interesting to see results from people who try it with 2020 models to see if they're able to improve temperatures and emissions.
Thank you for all of your efforts and updates on the 2019 MBA srkirt, and good health to you from the other side of the world!
Last edited:
If two 0.25mm sheets. I took them out of a refrigerator in an old pc and it is pure copper.
The dimensions I can not tell you because I have not measured them ...
0.50mm of thermal paste is outrageous. Pasta Artic MX4 2019.
[automerge]1587336246[/automerge]
Thank you, if I entered this group it was because I found it interesting because of the high temp.
I would bet anything that although the heatsink is a little bigger it also has that 0.5mm gap between the chip and the heatsink ... I don't know ... carefully and slowly it's easy to remove and so step change the paste for one like the one I use in pc overclocking which is the Artic MX 2019. Although it seems silly to go down 10º to the laptop is a lot !! and covering all the gaps with rubber makes the output air flow higher.
I am not an engineer but I think that here in 2019 they wanted to cut benefits and the cheapest way is to dissipate the heat a little.
Last edited:
So all this copper coins, soda cans and thermal paste fun ride was done on a 2019 MBA after all?
It was, yeah, but Srkirt's experiments have provided a good proof of concept for the 2020 in my view. The next step would be for someone who i) has a 2020 MBA (preferabbly an i5 or i7) and ii) isn't happy with the thermal performance to give it a go.
Ideally, they'd provide precise mesaurements of the heat sink design - as the 2019 and 2020 are different - so that the modification could be repeatable if people want to have a go themselves. Although with 11 months remaining on most owner's warranties, it'd have to be someone who really, really can't put up with the heat and the noise because they value silent/cool running above other factors (someone like me, but I just dropped my 2020 i5 off for return).
As long as I can watch porn on this machine and pull it up quickly I’m happy. Ideally I will use Deut to connect to my 2020 MacBook Air to my 5th gen iPad so I can have porn in 2 screens. Other than that all my i5/8gb/256 needs to is help me shop online.
Oh and I’d bet money this machine WILL NOT get as hot or loud as my 1st get Lenovo Helix does. You could light a cigarette off that machine.
If I ever decide I care about cooling I will buy a liquid cooled desktop.
It’s simple really if it gets too hot and becomes a fire hazard or another other type of hazard (looking at you Galaxy Note) and I get hurt because of it I will sue that pants off Apple.
Oh and I’d bet money this machine WILL NOT get as hot or loud as my 1st get Lenovo Helix does. You could light a cigarette off that machine.
If I ever decide I care about cooling I will buy a liquid cooled desktop.
It’s simple really if it gets too hot and becomes a fire hazard or another other type of hazard (looking at you Galaxy Note) and I get hurt because of it I will sue that pants off Apple.
I have the 2020 i5 MacBook Air. The only heat “issue” I’ve had was when I was running it at max for maybe 30 min and had the backside of my left bare hand directly underneath the cpu. It became uncomfortably hot in that one spot in the center (even thought I had a minor burn at first, but I didn’t lol).
For everything else it’s cool to the touch, and presumably if you’re doing something that uses the full power of it you’d be using a desk/table.
I’m mostly interested in doing this copper shim mod just for even more power and to extend the life of the cpu/gpu.
For everything else it’s cool to the touch, and presumably if you’re doing something that uses the full power of it you’d be using a desk/table.
I’m mostly interested in doing this copper shim mod just for even more power and to extend the life of the cpu/gpu.
Novella incoming:
I was aware that srkirt used a 2019 MBA, however nothing changed other than a *slightly* larger heatsink, which would not make a large difference especially when the new chips also sip a bit more power, including the improved graphics. Still the same concept of thermal management.
I see two groups of people (from youtube to forums, etc) regarding apple's thermal management: Those who are trying to bring down the temps, for whatever reason. To avoid being gaslit, they provide pictures of said temps, fan speeds, etc. The next group are those who say, "you are using it wrong" and whose argument at best is, "who cares if the fans run at 5k RPM to keep my temps at 99* during a 30 minute zoom meeting?". Truth is, I hear it on the other end of the meeting - the fan noise - it's not fun to listen to. Somebody asked if they were running a fan in the background, lol.
If covid wasn't a thing, I'd go to Apple right now and test it. But for now I see actual evidence from one side and words from the other. A videochat is not a heavy workload. Nor are youtube videos, even at 4k, but if you only have a 2k screen you might as well just use 2k. The workloads that this notebook appears to handle without high temps or high fanspeed are the same ones that a $300 ipad could handle smoother with less heat and no fan noise.
I'm typing this on a 4 year old XPS that weighs the same as the MBA in a physically smaller package. It doesn't get hotter than 65* for a 4k video. It doesn't get hot at all for a video chat. 20 memory-hogging chrome tabs, 40*, no fan spin, old dual cores sitting at 15% usage. I wasn't even aware that web browsing could possibly make a machine buckle as it does with this - RAM should be the limiting reagent in that equation. This is with old crusty paste and a fan that is probably full of dust, yet the chip and the metal body stay cool. It's limited only by my growing hatred for windows and the 8GB of RAM, which I constantly grab 90+% of. It is the cheaper model - the XPS lineup's equivalent to the MBA. Integrated graphics, dual core, the works.
Coming from the other side and glancing in to see the offerings, and willing to keep using an old XPS if those offerings appear to be a sidegrade instead of an upgrade, Apple has no excuse for the temps of this thing. Their iphones and especially their ipads are (imo) so superior to any other in terms of power and longevity that I'm surprised that same company produces these machines.
I was aware that srkirt used a 2019 MBA, however nothing changed other than a *slightly* larger heatsink, which would not make a large difference especially when the new chips also sip a bit more power, including the improved graphics. Still the same concept of thermal management.
I see two groups of people (from youtube to forums, etc) regarding apple's thermal management: Those who are trying to bring down the temps, for whatever reason. To avoid being gaslit, they provide pictures of said temps, fan speeds, etc. The next group are those who say, "you are using it wrong" and whose argument at best is, "who cares if the fans run at 5k RPM to keep my temps at 99* during a 30 minute zoom meeting?". Truth is, I hear it on the other end of the meeting - the fan noise - it's not fun to listen to. Somebody asked if they were running a fan in the background, lol.
If covid wasn't a thing, I'd go to Apple right now and test it. But for now I see actual evidence from one side and words from the other. A videochat is not a heavy workload. Nor are youtube videos, even at 4k, but if you only have a 2k screen you might as well just use 2k. The workloads that this notebook appears to handle without high temps or high fanspeed are the same ones that a $300 ipad could handle smoother with less heat and no fan noise.
I'm typing this on a 4 year old XPS that weighs the same as the MBA in a physically smaller package. It doesn't get hotter than 65* for a 4k video. It doesn't get hot at all for a video chat. 20 memory-hogging chrome tabs, 40*, no fan spin, old dual cores sitting at 15% usage. I wasn't even aware that web browsing could possibly make a machine buckle as it does with this - RAM should be the limiting reagent in that equation. This is with old crusty paste and a fan that is probably full of dust, yet the chip and the metal body stay cool. It's limited only by my growing hatred for windows and the 8GB of RAM, which I constantly grab 90+% of. It is the cheaper model - the XPS lineup's equivalent to the MBA. Integrated graphics, dual core, the works.
Coming from the other side and glancing in to see the offerings, and willing to keep using an old XPS if those offerings appear to be a sidegrade instead of an upgrade, Apple has no excuse for the temps of this thing. Their iphones and especially their ipads are (imo) so superior to any other in terms of power and longevity that I'm surprised that same company produces these machines.
Apple basically mailed it in when upgrading the 2020's internals. Seems the majority of the r&d was spent on the keyboard. The larger heatsink is a pittance to what they should have done with the cooling. The thing is Apple's macOS hardware r&d has been dwarfed by iOS hardware r&d for years and it shows. They're just following the money.
OTOH, here's my 8 year 2012 rMBP 15 that's still chugging along...
The 2020 MBA i5/i7 is basically neck and neck vs. my ancient MacBook Pro while weighing almost 40% less and superior battery life. Even the GPU is superior as I had to hack the rMBP 15 to disable the defective Nvidia dGPU as it's no longer working...
OTOH, here's my 8 year 2012 rMBP 15 that's still chugging along...
The 2020 MBA i5/i7 is basically neck and neck vs. my ancient MacBook Pro while weighing almost 40% less and superior battery life. Even the GPU is superior as I had to hack the rMBP 15 to disable the defective Nvidia dGPU as it's no longer working...
I just have to laugh when I read these replies. Apple did exactly what they wanted. They want to provide a machine for someone like me who and who wants to keep it for the next 5-6 years. I came from a MacBook Air 2011. I honestly could have probably kept it for another few years and used it. Word, Excel, Powerpoint, email, light photo editing, light PDF manipulation, some Zoom (Microsoft Teams), messages, Dropbox, FaceTime, some video consumption, some music consumption, and then a couple proprietary work software packages that are no issue.
Gotta love the myopic views that if it can't handle every single thing someone throws at it then it is faulty.
If the Air does not work for your workload then they make a laptop for you called the Pro.
Gotta love the myopic views that if it can't handle every single thing someone throws at it then it is faulty.
If the Air does not work for your workload then they make a laptop for you called the Pro.
So your better-suited-for-a-MBP workload doesn't do so well on a MBA?
Big surprise.
Wrong tool for the job.
[automerge]1587391739[/automerge]
Fixed that for you.
Start *moving* the air and it's pretty good at cooling.
Kinda what a fan does... Hmm...
Last edited:
I think the reason Apple used such a bad thermal solution is that their goal was to differentiate between the Air and the MBP 13. It was not about trying to perform the best possible way. Cause let's agreed on that - it's technically possible to do much better, and it's pretty obvious as well.
If the Air had better cooling, lots of people (including me) would buy it instead of a Pro model. Lots of people would be satisfied with the Air's performance, as long as it could keep it for longer than a minute without melting down...
I'm sad for all those great engineers at Apple who had to stand down and accept a decision from the product team that's technically abhorrent.
P.S. Sorry if someone suggested this before, I wasn't able to read the whole thread.
If the Air had better cooling, lots of people (including me) would buy it instead of a Pro model. Lots of people would be satisfied with the Air's performance, as long as it could keep it for longer than a minute without melting down...
I'm sad for all those great engineers at Apple who had to stand down and accept a decision from the product team that's technically abhorrent.
P.S. Sorry if someone suggested this before, I wasn't able to read the whole thread.
LOL. Game set match. By you saying this it pretty much shows that Apple engineers nailed the "thermal solution"
So you're saying there's something inherently wrong with using a light laptop for watching 4K YouTube videos or editing a video clip of your cat? Cause these tasks are quite non-Pro, and yet the non-Pro MacBook model suffers performing them.
Or did I get it wrong? I'd appreciate it if you actually said what you mean instead of implying it.
Or did I get it wrong? I'd appreciate it if you actually said what you mean instead of implying it.
There is no wrong or right. There simply is. What you do with your laptop is up to you. If that is the right tool then it is the right tool. For my workload as well as dozens of other here (and thousands of others not here) the MacBook Air is flawless or you could say the right tool. I have no cooling issues, no noise issues, and no performance issues. If your workload cannot be completed with this tool then it is not the right tool for you.
To use a car analogy. A Honda Accord is a fine automobile. If you try to use it to race on a track it will be horrible. It is not the right tool for that use. It doesn't somehow make it a poor automobile now.
If consuming 4K video or editing 4K clips of your cats pushes this thing beyond its limits and that is specific to your workload then this is not the right laptop for you. I don't do either of those things, nor do tens of thousands of others.
What you are basically saying (another analogy) is that a car that costs $100,000 and is built for speed can go 150mph, but my car that I bought which was 1/2 the price and not built for speed cannot go 150mph therefore that car is flawed.
These posts are all the same which is "This laptop is not the right laptop for me therefore it sucks and is flawed."
Last edited:
No it is just simple for most to understand. Apparently they cannot fathom how there are laptops that exist that can't handle every single thing thrown at it.
Apparently people can understand lower cost cars (same brand) do not have the performance of more expensive cars but cannot seem to piece together that a lower price laptop does not have the performance of a more expensive laptops (same brand)
Is everyone who uses a Mac a 4k video editor? I work in IT and I don't think I've ever edited 4k video.
This is why many of us oldsters are upset with this 2020 model. When it was announced it would have quad core some of use were VERY excited as circa 2015, when you upgraded the 2015 MBA to i7, it would come very close to the base MBP 13 of 2015
See here for notebookcheck.net's Cinebench 15 results:
I had that MBA and it was awesome. Within 10% of the MBP while being lighter and much cheaper
Compare to 2020 MBA vs 2019 MBP 13
A whopping 40% slower when pushed. (I'm comparing 363 vs 611.7 when the MBA is heat soaked from it's less than adequate cooling system
I have the i7 2020 MBA and it's only barely 7% faster than the i5 in Cinebench. Not even close...
We did not see this market segmentation back then. But it's definitely being done today.
No, my test was done on a base model 2020 i3, it made a significant difference: https://forums.macrumors.com/posts/28373009/