2020 Air Heatsink Modification Thread

[throAU]

and you know what else desktops got? Fans on the heatsink
(jk here)

Not always, mac pro does not for instance.

Neither does my desktop, but it does essentially have a heat pipe and a radiator (liquid cooled - with 5x140mm fans)

 

[rba1989]

Great new video on testing the new MacBook Air (i3 and i5) in real world tasks in comparison with 2018 MacBook Air. It is opposite to Maxtech videos...

 

[jgorman]

I've got a heat sink question or two:

If you look at the MacBook Air teardown on iFixIt - https://www.ifixit.com/News/36480/theres-something-new-in-the-macbook-air - They show the design changes from 2018 to 2020 - both only have a heat sink (not looking at the 2015 version). The 2020 heat sink is even a bit bigger.

So was the 2018 MBA a bad design? Or was that adequate for a 7W chip (2018) compared to a 9W chip (2020)?

Is it wise to compare these to the 2015 MBA that had a 15W chip?

Instead of a heat pipe - can't someone design a thin add-on heat sink? One that someone else would work the kinks out of before I ever tried to use one?

I'm sure the heat sink has thermal paste underneath - Has anyone (youtuber?) tried to take a look and determine the sloppiness of it?

I'm a vim user over Visual Studio - but I'd like to know what the heat/fans are like for VSCode 'power user' - please let us know.

Yes, there are a few differences in the thermals between the 2020 model and the 2018 model.

Also, as you probably know, VS Code shares the same engine and framework as Chrome. Chrome seems to run hotter on these machines than Safari.

 

[jgorman]

Great new video on testing the new MacBook Air (i3 and i5) in real world tasks in comparison with 2018 MacBook Air. It is opposite to Maxtech videos...

This video seems consistent with what other reviews say. MaxTech said in a video that the CPU temperature goes to 100 degrees C in Chrome on launch, scrolling through a page and switching tabs. This video also seems to show that.

That MaxTech video tested a 4K60 video, and this video tested a 4K30 video. Another review did check temperatures after five minutes of playing 1080p60 video in Chrome (He is speaking Japanese, but the Intel Power Gadget is in English). It does appear Chrome runs hotter.

It also looks like the i3 model is running cooler in a lot of operations than the i5, which should be helpful for people trying to decide between them.

 

[peter.black]

this is absolutely a deal breaker for me. reason why i held off on air so far for personal use is cause of lack of heatpipe

Yep, definitely a deal breaker for me too. Decided not to buy the MBP and I've just bought a bit of heatpipe instead. Completely silent in operation and shows no sign of overheating.

 

[ilikewhey]

Yep, definitely a deal breaker for me too. Decided not to buy the MBP and I've just bought a bit of heatpipe instead. Completely silent in operation and shows no sign of overheating.

you use that heat pipe to connect that dry sense of humor of yours 😂

 

[bill-p]

I'm a vim user over Visual Studio - but I'd like to know what the heat/fans are like for VSCode 'power user' - please let us know.

I just returned my 2020 MacBook Air since it's not what I was looking for, but... VS Code wasn't bad at all. Chrome's high resource use has to do with stray JavaScript on a lot of pages these days. VS Code itself is pretty well optimized and I don't think I noticed much of a difference between my MacBook Pro and the MacBook Air running VS Code.

Then again, VS Code is not my main IDE. I only use it for my website. Android Studio and Xcode are more my thing, and let's just say... Android Studio was not kind to the MacBook Air at all.

 

[intelligence]

Baffled that a lot of people seem to be fine with a machine running warmer than it has to, just because it's marketed towards casual users. Would it hurt you to have a slightly cooler running machine, or what?

Truth is that Apple cheaped out on the thermals here, either because they wanted to save a buck or they didn't wanna cannibalize 13" Pro sales. Whatever it is, it's hurting the customers and for no good reason. This machine could be great, now it's OK.

 

[nylon]

Baffled that a lot of people seem to be fine with a machine running warmer than it has to, just because it's marketed towards casual users. Would it hurt you to have a slightly cooler running machine, or what?

Truth is that Apple cheaped out on the thermals here, either because they wanted to save a buck or they didn't wanna cannibalize 13" Pro sales. Whatever it is, it's hurting the customers and for no good reason. This machine could be great, now it's OK.

I guess its subjective. Though I would argue that lowering the price to $999 ($899 student) and doubling the storage is also having an eye on saving the consumer some money and increasing the value proposition. So it really depends from where you come from.

 

[Bolanders]

I can't believe Lexus doesn't put their best engine in every single car. I can't believe car manufacturers limit the speed on certain models. I can't believe they don't put every single option (or many options) in a car when it would (might)add minimal cost for some things. The list goes on and on.

 

[RegularGuy09]

MacBook Air's CPU seems to be one that works in bursts, and not one that is designed to work for sustained loads. I'm guessing(and correct me if I'm wrong), adding a heatpipe or whatever, wouldnt make a HUGE difference to its performance in anyway.

 

[deeddawg]

Baffled that a lot of people seem to be fine with a machine running warmer than it has to [...]
This machine could be great, now it's OK.

Lots of folks simply don't bother getting worked up over stuff they cannot control.

They vote with their wallet and move on.

 

[Saul Giordani]

I'm writing from my new i7/16/512: a 4k video playing, using whatsapp, mail, writing here and surfing Facebook. Temp at 53 deg average, no fan. It goes up to 80 for 1 second when I start new apps and then it comes back to 53. I'm using safari.

Same operations with the video played on chrome and the cpu temp goes up to 75-80 and the fan starts.

I left it open for a couple of hours today, to index and everything. For now I'm not seeing overheating behaviors, but it's only my first hour of usage.

 

[KPOM]

MacBook Air's CPU seems to be one that works in bursts, and not one that is designed to work for sustained loads. I'm guessing(and correct me if I'm wrong), adding a heatpipe or whatever, wouldnt make a HUGE difference to its performance in anyway.

That’s likely. These chips are binned and tested for burst speeds, but not for sustained output. Even the YouTube poster who connected a liquid cooler observed that Apple appears to limit the voltage. That’s likely for several reasons including reliability and power consumption.

 

[intelligence]

Lots of folks simply don't bother getting worked up over stuff they cannot control.

They vote with their wallet and move on.

I'm a bit miffed. Product looked great on paper and geekbench scores looked amazing for an ultrabook, but there are some concerns with how hot this thing get. I'm glad it seem to perform better with Safari, but that's not an option for me as a web developer unfortunately.

I'd gladly wait and opt for 13"/14" MBP with better cooling, but that one will most likely be outfitted with the non optional touch bar. So yeah, stuck between a rock and a hard place.

 

[jgorman]

That’s likely. These chips are binned and tested for burst speeds, but not for sustained output. Even the YouTube poster who connected a liquid cooler observed that Apple appears to limit the voltage. That’s likely for several reasons including reliability and power consumption.

That YouTube video not only tested a water cooler, but also what happens if a cooling gel pack is under the MacBook Air's bottom case. Here are the differences in performance and temperature. Even though that video was just for fun, it would be nice to have some of these performance and temperature changes. The effectiveness of a heat pipe connected to fan likely falls somewhere between these two cooling methods, even with no change in firmware.

 

[bill-p]

Baffled that a lot of people seem to be fine with a machine running warmer than it has to, just because it's marketed towards casual users. Would it hurt you to have a slightly cooler running machine, or what?

Well, honestly, I tried, but some won't see the point, and I'm starting to not see the point myself. Even if you manage to convince everyone in the entire world that the MacBook Air 2020 needs a heatpipe and unlocked voltage, someone at Apple does not see that as a necessity, and that's where we are.

Also, I realized I need a more robust machine that can handle my... let's say "advanced text editing needs", and the MacBook Air 2020 is just not that machine. I'm moving on. Rumors of a new MacBook Pro 13" coming out next month or by WWDC basically bolsters the point home: Apple knew what they were doing, and they knew people like me were looking for a replacement, they'll have that replacement out soon.

The MacBook Air is just not for me.

 

[KPOM]

That YouTube video not only tested a water cooler, but also what happens if a cooling gel pack is under the MacBook Air's bottom case. Here are the differences in performance and temperature. Even though this video was just for fun, it would be nice to have some of these performance and temperature changes. The effectiveness of a heat pipe connected to fan likely falls somewhere between these two cooling methods, even with no change in firmware.

Interestingly I didn’t notice much of a difference with a gel pack (I have a 2019 1.4GHz MacBook Pro and a 2020 i7 Air). On the Pro, there is no difference at all in performance, but the fan shuts off.

On a side note, I’ve been running the Folding at Home COVID-19 routines on my MacBook Pro (part of the AnandTech/Tom’s Hardware friendly “contest”). That’s really taxing on the CPU. It basically maxes it out (so I’ve been running it a few hours a day). The Pro is about 75% faster in that ”test” because it is able to sustain very high CPU speeds even when all 4 cores are running (2.8-3 GHz). The Air settles around 1.5-1.6GHz. With the gel pack, the Air bumps up to about 1.7-1.8GHz. The Pro shuts off the fan, but the speed is the same.

When I‘m not running Folding at Home (i.e. normal use), both handle running Quicken for Windows under WINE (Crossover 19), Parallels Desktop running the latest Windows 10 build, etc. I don’t do any heavy video or photo editing, but do touch up photos occasionally.

I had a hard time recommending the 2018 or 2019 MacBook Air since the Pro was just $200 more and, particularly since the 2019 Pro went quad-core, offered a lot more performance. The 2020 Air is easier to recommend for now. With the i5, it’s $400 less with comparable RAM and storage. The CPU is still slower, but it is very workable for most people. Maybe that changes in a few weeks when the new Pro comes out, but even then, the Air is a very usable machine.
[automerge]1586396632[/automerge]

Well, honestly, I tried, but some won't see the point, and I'm starting to not see the point myself. Even if you manage to convince everyone in the entire world that the MacBook Air 2020 needs a heatpipe and unlocked voltage, someone at Apple does not see that as a necessity, and that's where we are.

Also, I realized I need a more robust machine that can handle my... let's say "advanced text editing needs", and the MacBook Air 2020 is just not that machine. I'm moving on. Rumors of a new MacBook Pro 13" coming out next month or by WWDC basically bolsters the point home: Apple knew what they were doing, and they knew people like me were looking for a replacement, they'll have that replacement out soon.

The MacBook Air is just not for me.

But that’s fine. Apple will soon have a suitable computer to replace your 2018 MacBook. The 2020 Air is a significant upgrade from the 2019 Air. Heck, I still am hopeful they resurrect the 12” MacBook (ARM or otherwise). The 2017 MacBook was my favorite Mac ever. It was so small and light I didn’t mind the 5W fanless CPU. The 2020 Air is “good enough” for the 30-40% of the Mac buyer population who will buy it.

I do think that Apple could have been more aggressive with the Air’s design. Perhaps with a heat pipe, they could have made it a bit smaller and lighter to differentiate it from the 13” Pro. The tapered design is somewhat more ergonomic than the 13” Pro, but I agree that the 13” Pro is just as portable. If the 2020 base Pro is the same size, then it will still likely be my recommendation for many users, but the 2020 Air is good enough for me to recommend to the average user. By contrast, I did NOT recommend the 2019 Air since the 2019 Pro was so much more powerful for $200 more.

 

[jgorman]

Interestingly I didn’t notice much of a difference with a gel pack (I have a 2019 1.4GHz MacBook Pro and a 2020 i7 Air). On the Pro, there is no difference at all in performance, but the fan shuts off.

On a side note, I’ve been running the Folding at Home COVID-19 routines on my MacBook Pro (part of the AnandTech/Tom’s Hardware friendly “contest”). That’s really taxing on the CPU. It basically maxes it out (so I’ve been running it a few hours a day). The Pro is about 75% faster in that ”test” because it is able to sustain very high CPU speeds even when all 4 cores are running (2.8-3 GHz). The Air settles around 1.5-1.6GHz. With the gel pack, the Air bumps up to about 1.7-1.8GHz. The Pro shuts off the fan, but the speed is the same.

When I‘m not running Folding at Home (i.e. normal use), both handle running Quicken for Windows under WINE (Crossover 19), Parallels Desktop running the latest Windows 10 build, etc. I don’t do any heavy video or photo editing, but do touch up photos occasionally.

I had a hard time recommending the 2018 or 2019 MacBook Air since the Pro was just $200 more and, particularly since the 2019 Pro went quad-core, offered a lot more performance. The 2020 Air is easier to recommend for now. With the i5, it’s $400 less with comparable RAM and storage. The CPU is still slower, but it is very workable for most people. Maybe that changes in a few weeks when the new Pro comes out, but even then, the Air is a very usable machine.
[automerge]1586396632[/automerge]

But that’s fine. Apple will soon have a suitable computer to replace your 2018 MacBook. The 2020 Air is a significant upgrade from the 2019 Air. Heck, I still am hopeful they resurrect the 12” MacBook (ARM or otherwise). The 2017 MacBook was my favorite Mac ever. It was so small and light I didn’t mind the 5W fanless CPU. The 2020 Air is “good enough” for the 30-40% of the Mac buyer population who will buy it.

I do think that Apple could have been more aggressive with the Air’s design. Perhaps with a heat pipe, they could have made it a bit smaller and lighter to differentiate it from the 13” Pro. The tapered design is somewhat more ergonomic than the 13” Pro, but I agree that the 13” Pro is just as portable. If the 2020 base Pro is the same size, then it will still likely be my recommendation for many users, but the 2020 Air is good enough for me to recommend to the average user. By contrast, I did NOT recommend the 2019 Air since the 2019 Pro was so much more powerful for $200 more.

I understand. I am actually interested in the i3 model. My typical use falls at the ends of two extremes: abacus with a screen and high-end desktop. So, what I would do is get the i3 and use Turbo Boost Switcher to turn off turbo boost. I do not know really how much the turbo boost draws in the i3, but it can make the i5 CPU draw over 32W. Even if the temperatures do not stay high, I want to minimize swings in temperature too.

 

[Trixs]

Mine hits a 100 degrees too while browsing in Chrome (mostly hovering between 80-100). But at least the case doesn't seem to get that hot. The case on my Macbook Pro 2018 sometimes gets so hot I can hardly use it on my lap.

 

[KPOM]

Mine hits a 100 degrees too while browsing in Chrome (mostly hovering between 80-100). But at least the case doesn't seem to get that hot. The case on my Macbook Pro 2018 sometimes gets so hot I can hardly use it on my lap.

I‘ve noticed that, too. Even if Intel Power Gadget reports the CPU temperature at 100 degrees, the MacBook Air itself doesn’t seem hot.

 

[magbarn]

That’s likely. These chips are binned and tested for burst speeds, but not for sustained output. Even the YouTube poster who connected a liquid cooler observed that Apple appears to limit the voltage. That’s likely for several reasons including reliability and power consumption.

These chips are more capable than what you're giving them credit for.
It's limitation that's in Apple firmware that's keeping the chip from boosting too much for the cooling design of the MBA 2020. I've posted a link for notebookcheck on the Acer that has the i7 version of this chip. In Cinebench R 15 on the Multicore benchmark, the Acer Scores 612 cold and 553 hot. A drop of roughly 10%. The i5 2020 MBA scores only 448 cold and 363 hot. A significant drop of 19%! All the while the heatpipe equipped Acer is over 10dbA quieter!
Also, based on the cinebench scores, the i7 version of ice lake on the Acer is roughly 27% faster than the i5 chip in the MBA when cold and up to 35% faster when hot. Either the i7 ice lake is much much faster (previous i7 vs i5 comparisons have a much closer performance delta) than the i5, or Apple is severely limiting the performance of the MBA's CPU to keep it within the power/heat envelope of the MBA.

As soon as I get my i7 2020 MBA, I'm going to turn off power saving and let it stay on for 24 hours to get all the indexing done and then I'm going to run Cinebench like crazy on it to compare against the same i7 ice lake in the Acer

EDIT: Nevermind, I was assuming the Acer was using the same ice lake cpu with the same TDP, but the Acer's chip has a TDP of 15 watts, while the MBA is limited to only 9~10watts. So the Acer is able to draw 50% more power.
That's a significant difference and likely accounts for the big performance gap between the 2 machines. The MBA still drops close to 20% when running hot vs cold, so that makes it "bursty" in nature as you previously eluded to,.

 

[KPOM]

These chips are more capable than what you're giving them credit for.
It's limitation that's in Apple firmware that's keeping the chip from boosting too much for the cooling design of the MBA 2020. I've posted a link for notebookcheck on the Acer that has the i7 version of this chip. In Cinebench R 15 on the Multicore benchmark, the Acer Scores 612 cold and 553 hot. A drop of roughly 10%. The i5 2020 MBA scores only 448 cold and 363 hot. A significant drop of 19%! All the while the heatpipe equipped Acer is over 10dbA quieter!
Also, based on the cinebench scores, the i7 version of ice lake on the Acer is roughly 27% faster than the i5 chip in the MBA when cold and up to 35% faster when hot. Either the i7 ice lake is much much faster (previous i7 vs i5 comparisons have a much closer performance delta) than the i5, or Apple is severely limiting the performance of the MBA's CPU to keep it within the power/heat envelope of the MBA.

As soon as I get my i7 2020 MBA, I'm going to turn off power saving and let it stay on for 24 hours to get all the indexing done and then I'm going to run Cinebench like crazy on it to compare against the same i7 ice lake in the Acer

EDIT: Nevermind, I was assuming the Acer was using the same ice lake cpu with the same TDP, but the Acer's chip has a TDP of 15 watts, while the MBA is limited to only 9~10watts. So the Acer is able to draw 50% more power.
That's a significant difference and likely accounts for the big performance gap between the 2 machines. The MBA still drops close to 20% when running hot vs cold, so that makes it "bursty" in nature as you previously eluded to,.

The 15W TDP chip maxes out at 25W. The chips in the MacBook Air are rated 10W but max out at 12W. BIG difference.

 

[Trixs]

Mine hits a 100 degrees too while browsing in Chrome (mostly hovering between 80-100). But at least the case doesn't seem to get that hot. The case on my Macbook Pro 2018 sometimes gets so hot I can hardly use it on my lap.

Actually looks like more 60-100, mostly around 70 now. Before onedrive was going nuts which is probably why it didn't go below 80.

 

[magbarn]

The 15W TDP chip maxes out at 25W. The chips in the MacBook Air are rated 10W but max out at 12W. BIG difference.

Whelp, I guess the MBA is the true spiritual of the MB 12 with a very power constrained chip, no wonder Apple discontinued the MB 12.