M1 Air with Thermal Pad Mod

[rui no onna]

The problem is that you don’t know if it’s low risk. It could be very high risk, and that’s a oroblem. You want to assume it’s low risk because you want to do it. Not everything we want to be true, is.

The risk is you might damage the laptop taking it apart and you might lose warranty.

The fact that the M1 is running faster just means the thermal pad is doing its job at quickly taking away heat from the SoC. The system still decides if it should throttle or not and they're basing that on internal SoC temps (which it obviously thinks is acceptable enough to keep running at high clocks). If it was botched in any way, what we'd see is even quicker throttling.

I wouldn't do it myself (too clumsy) but hey, if someone has weighed the risks and rewards and still decides to do it, then by all means go for it. It's their money.

 

[downshiftdre]

Still, anyone tested their score after 11.1 update?

I reran 10 minute test and I dont see a noticeable difference in scores. What does 11.1 have that would make you think the score would change?

 

[SesameWasher]

I reran 10 minute test and I dont see a noticeable difference in scores. What does 11.1 have that would make you think the score would change?

Optimization down the road as it happened a couple of times.

 

[Leon1das]

Hi sorry - new here and love your post.

I want to do this mod but am aware of the possible warranty impact. My key Q - can you actually remove the screws without leaving any marks on them?

 

[Chozes]

Hi sorry - new here and love your post.

I want to do this mod but am aware of the possible warranty impact. My key Q - can you actually remove the screws without leaving any marks on them?

Does using the screws cause issues? Wear will prevent apple from servicing?
Keen on this ‘mod’.

 

[Leon1das]

Does using the screws cause issues? Wear will prevent apple from servicing?
Keen on this ‘mod’.

I am super interested too...esp seeing performance..

2 things that worry me about the mod (re warranty) - are:
- potential marks on the screws (If they get scratched during the opening) and
- removing sticker tape inside the cover to ensure direct adherence of thermal pad with the aluminium of the cover. (OP stated he did this - but aparently this removal is irreversible)

I am in Europe and my device has 2y warranty.

If anyone has positive experience of modding while keeping screws intact and without removal of the sticker tape - I am very interested.

 

[Tenkaykev]

I am super interested too...esp seeing performance..

2 things that worry me about the mod (re warranty) - are:
- potential marks on the screws (If they get scratched during the opening) and
- removing sticker tape inside the cover to ensure direct adherence of thermal pad with the aluminium of the cover. (OP stated he did this - but aparently this removal is irreversible)

I am in Europe and my device has 2y warranty.

If anyone has positive experience of modding while keeping screws intact and without removal of the sticker tape - I am very interested.

I did this with my 2020 intel i5. Removing the screws was straightforward. I used a good quality screwdriver of the correct size, the sort that allow you to rest your finger on the top of the screwdriver which remains static while the body of the screwdriver rotates.
Rest the top of the screwdriver on the top of the screw and gently rotate it anti clockwise. You will both hear and feel a click as the head drops into place in the screw head. When tightening up after the mod follow the same procedure and tighten until each screw starts to offer resistance then go on to the next. Once all are inserted go back and gently tighten. I didn’t remove the insulation on the base as leaving it in place prevents the base becoming too hot for comfort if used on your lap.
I’ve not opened my M1 air yet but have seen a couple of photos showing a small thermal sponge, was curious if it is fitted to all M1 airs?

 

[Leon1das]

I didn’t remove the insulation on the base as leaving it in place prevents the base becoming too hot for comfort if used on your lap.

Thanks - just to summarize: on your prev Mac - you did install thermal pad without removing insulation from the cover? If so - did thermal pad still have contact with aluminium directly? Without this contact I doubt this mod would work as needed...
I have nothing against base being super warm - as this takes the heat outside the housing...

 

[Tenkaykev]

Thanks - just to summarize: on your prev Mac - you did install thermal pad without removing insulation from the cover? If so - did thermal pad still have contact with aluminium directly? Without this contact I doubt this mod would work as needed...
I have nothing against base being super warm - as this takes the heat outside the housing...

Yes, I left the insulation in place. The pad acted as a thermal bridge to the case as it was pressing against the insulation.it was still a better heat transfer medium than air, and I wasn’t happy removing the insulation.

 

[Leon1das]

Yes, I left the insulation in place.

Sorry, call me stupid - but I dont read this as a full response...

You left insulation + you installed thermal pad mod + thermal pad did not touch aluminium but only insulation?

Edit: Many thanks for editing and clarifying...

 

[Tenkaykev]

Sorry, call me stupid - but I dont read this as a full response...

You left insulation + you installed thermal pad mod + thermal pad did not touch aluminium but only insulation?

Many thanks for clarifying...

I’d previously done the shim mod, then later went back and did the pad mod as well, I’ll dig out the before and after scores and post them shortly.

 

[Tenkaykev]

I’d previously done the shim mod, then later went back and did the pad mod as well, I’ll dig out the before and after scores and post them shortly.

OK,
Just had a look back at my Geekbench 5 scores. Original i5 2020 Air: 1111/2681
Shim mod 1157/3502
Shim and Pad mod 1123/3859
I had a couple of readings over the 4000 mark when starting from cold, but generally getting 3800/3900.

My base M1 Air has just given me 1709/7611 - no fans

 

[denzdaniel]

I am going to do this MOD as soon as I receive the thermal pad I've ordered from an online store. I bought the Fujipoly Ultra Extreme 17.0 W/mk. Let's see the score if it will change and if this really worth doing. I'm doing this to lengthen the life of my M1 MBA. If it's always warm it's ok... But when I do rendering Handbrake, FCPX ETC then I will have to make sure my machine will not keep the heat inside the beautiful M1 Chip right? I will make sure the heat will find its way out faster by doing a very good passive heat transfer/cooling.

It will be posted here and my Youtube.
Standby friends!

 

[Blasserre]

I am going to do this MOD as soon as I receive the thermal pad I've ordered from an online store. I bought the Fujipoly Ultra Extreme 17.0 W/mk. Let's see the score if it will change and if this really worth doing. I'm doing this to lengthen the life of my M1 MBA. If it's always warm it's ok... But when I do rendering Handbrake, FCPX ETC then I will have to make sure my machine will not keep the heat inside the beautiful M1 Chip right? I will make sure the heat will find its way out faster by doing a very good passive heat transfer/cooling.

It will be posted here and my Youtube.
Standby friends!

What is your YouTube channel account?. I am waiting to see more results.

 

[KnICS]

A bit late to the game, but I played around a bit with the thermal mod, and had good results.

* Without the mod, 30 minute Cinebench R23 scores averaged 7004
* With the mod, 30 minute Cinebench R23 score averaged 7749

My Air had the black thermal sponge that was pictured earlier - my guess is that Apple either added this as an early design modification, or perhaps only for higher spec’d airs (mine is 16GB/2TB)

Earlier posts about battery heat scared me a bit, so I monitored that with TG Pro, and the battery did get as high as 41C … so I reversed the mod, and it turns out that the battery still gets that hot without the mod. That worries me a bit, and is the first thing I’ve found so far that I don’t like about the M1 Air. I don’t have any hard data (unless I un-mod it AGAIN and do tests) but I think the battery actually heats up slower with the thermal mod in place — likely because the heat is dispersing through the shell rather than into the stagnant internal air mass.

Anyone have thoughts on how to insulate the batteries?

 

[denzdaniel]

A bit late to the game, but I played around a bit with the thermal mod, and had good results.

* Without the mod, 30 minute Cinebench R23 scores averaged 7004
* With the mod, 30 minute Cinebench R23 score averaged 7749

My Air had the black thermal sponge that was pictured earlier - my guess is that Apple either added this as an early design modification, or perhaps only for higher spec’d airs (mine is 16GB/2TB)

Earlier posts about battery heat scared me a bit, so I monitored that with TG Pro, and the battery did get as high as 41C … so I reversed the mod, and it turns out that the battery still gets that hot without the mod. That worries me a bit, and is the first thing I’ve found so far that I don’t like about the M1 Air. I don’t have any hard data (unless I un-mod it AGAIN and do tests) but I think the battery actually heats up slower with the thermal mod in place — likely because the heat is dispersing through the shell rather than into the stagnant internal air mass.

Anyone have thoughts on how to insulate the batteries?

That's good! Well, I will do mine and it will be a permanent so I can prolonged the life of the MacBook Air. 😉

When it comes to Thermals I wanna make sure I do it right.
You can buy a Thermal Tape some what helps with the heat and protects the surrounding components next to the M1 Chip.

 

[denzdaniel]

What is your YouTube channel account?. I am waiting to see more results.

Hi! Yes sure... My Youtube channel here...

https://www.youtube.com/c/OutrunDYC?sub_confirmation=1

 

[Tenkaykev]

A bit late to the game, but I played around a bit with the thermal mod, and had good results.

* Without the mod, 30 minute Cinebench R23 scores averaged 7004
* With the mod, 30 minute Cinebench R23 score averaged 7749

My Air had the black thermal sponge that was pictured earlier - my guess is that Apple either added this as an early design modification, or perhaps only for higher spec’d airs (mine is 16GB/2TB)

Earlier posts about battery heat scared me a bit, so I monitored that with TG Pro, and the battery did get as high as 41C … so I reversed the mod, and it turns out that the battery still gets that hot without the mod. That worries me a bit, and is the first thing I’ve found so far that I don’t like about the M1 Air. I don’t have any hard data (unless I un-mod it AGAIN and do tests) but I think the battery actually heats up slower with the thermal mod in place — likely because the heat is dispersing through the shell rather than into the stagnant internal air mass.

Anyone have thoughts on how to insulate the batteries?

Brilliant that you have done this so methodically.
With regard to the battery heat, that doesn’t sound too much of an issue to me. It’s not really getting “hot” just a bit hotter as it is supplying more energy to the processor on a continuous basis. Perhaps someone with access to the “Pro” could run the test and report back with the comparative readings.
if it is of concern, then a thermal pad on each battery would do a similar job as the one on the heatsink.

 

[KnICS]

I've done a few more tests, with the thermal mod units own and with the thermal mod + a small extra thermal pad on each battery. I'll post more later if/when I have time to do more tests, but the quick impression is:

Putting additional thermal pads on the battery actually makes the battery heat worse. I'm getting 45C after 30 minutes and 46C after 45 minutes. I only used very small pads (because I ran out), but the data leads me to conclude that the SOC's heat is transferring quickly to the case thanks to the original thermal pad mod, and then the hot case is transferring heat back onto the batteries, rather than transferring via internal air.

Sustained temps above 40C will degrade battery life, so this is a bit of a concern. However, the only way to get these levels of heat is to play a game through rosetta (I've tried Divinity Original Sin 2 at ultra settings, and SOTR at Highest) for sustained periods. For normal use, the M1 with thermal pads will give M1 Pro performance without worry (IMO).

I think this says something about poorly optimized apps, though. Games under rosetta 2 seem to really overtax the CPU and if it's a game it's going to tax the GPUs as well. While performance is great, these apps (again IMO) push the M1 beyond what it's intended. This is bad because the games play so great ... I'm going to continue to experiment because I really want the M1 Air to be a suitable (and safe!) gaming laptop.

Next up:
1) I've ordered some higher-conductivity thermal pads (15W vs 6W) - this will make the case heat up even more and help price or disprove how heat is soaking through this system
2) I'm experimenting with external laptop coolers. If the case *is* becoming the primary vector for heat transfer, a cooler should help a lot (although in past experience they usually don't do much).

 

[Tenkaykev]

I've done a few more tests, with the thermal mod units own and with the thermal mod + a small extra thermal pad on each battery. I'll post more later if/when I have time to do more tests, but the quick impression is:

Putting additional thermal pads on the battery actually makes the battery heat worse. I'm getting 45C after 30 minutes and 46C after 45 minutes. I only used very small pads (because I ran out), but the data leads me to conclude that the SOC's heat is transferring quickly to the case thanks to the original thermal pad mod, and then the hot case is transferring heat back onto the batteries, rather than transferring via internal air.

Sustained temps above 40C will degrade battery life, so this is a bit of a concern. However, the only way to get these levels of heat is to play a game through rosetta (I've tried Divinity Original Sin 2 at ultra settings, and SOTR at Highest) for sustained periods. For normal use, the M1 with thermal pads will give M1 Pro performance without worry (IMO).

I think this says something about poorly optimized apps, though. Games under rosetta 2 seem to really overtax the CPU and if it's a game it's going to tax the GPUs as well. While performance is great, these apps (again IMO) push the M1 beyond what it's intended. This is bad because the games play so great ... I'm going to continue to experiment because I really want the M1 Air to be a suitable (and safe!) gaming laptop.

Next up:
1) I've ordered some higher-conductivity thermal pads (15W vs 6W) - this will make the case heat up even more and help price or disprove how heat is soaking through this system
2) I'm experimenting with external laptop coolers. If the case *is* becoming the primary vector for heat transfer, a cooler should help a lot (although in past experience they usually don't do much).

Sterling work!
I'd not considered the possibility of heat transferring from the processor via the thermal pad on the heat sink, through the case and then through the pads on the batteries.
I did a Heath-Robinson cooling kludge and posted photos on another thread here:

2020 Air Heatsink Modification Thread

I was having a general clear out / tidy up recently and came across a set of Electronic Bathroom Scales. These were an inexpensive set we purchased years ago from Tesco. They're the sort with a LED Display mounted in a heavy Glass base, with the load sensors being mounted in the four corner...

forums.macrumors.com

I'm also curious about the ventilation in the M1 Air. With the M1 pro there is a cooling fan mounted in the upper left corner of the machine. As I understand it, this draws air in through an area near the hinge on the right hand side, over the Processor heat sink and exhausts via a gap on the left hand side. I can't see any gap but assume that it is in the area between the hinges immediately below the " MacBook Pro " logo just below the screen. I'm unsure whether this gap exists on the M1 Air allowing the internals to "breathe"

 

[KnICS]

There is no gap that I can see. I doubt it's air-tight but it definitely cant breathe. This is good and bad - good because no dust can get in, bad because hot air can't escape... it basically has to soak through the chassis. Under normal (and even heavy) use, this seems to work great. Under crazy loads like high end gaming, it's just not fast enough.

My cooler experiments are proving very effective so far, but its also super ugly at the moment. I'm working on cleaning it up a bit (in my spare time) and will eventually post pictures. I don't have the skills to machine a nice base out of metal though.

Apple Silicon is powerful enough that it should hopefully attract game developers back to Apple. Hoping that happens, and that the real "pro" MacBooks coming in 2021 are what they're cracked up to be, I expect that enthusiast gamers will want to push the limits of the MacBooks just like PC enthusiasts do. So this research is all towards that end. Anyone considering an M1 Air for normal, non-psycho use will never have to worry about heat or performance, this thing is way beyond expectations. Heck, even when it throttles, it's still way faster than my old 2016 MBP 15", which is at max spec.

A slight word of caution: I've pretty much screwed my warranty at this point - I've accidentally marked up my batteries (the tack from the pads left discolored spots), scratched the inside of the case, and lost one of the screws. While the thermal pad mod is easy, safe and reversible, if you do it like 15 times in a row late at night ... well, accidents happen

 

[Tenkaykev]

There is no gap that I can see. I doubt it's air-tight but it definitely cant breathe. This is good and bad - good because no dust can get in, bad because hot air can't escape... it basically has to soak through the chassis. Under normal (and even heavy) use, this seems to work great. Under crazy loads like high end gaming, it's just not fast enough.

My cooler experiments are proving very effective so far, but its also super ugly at the moment. I'm working on cleaning it up a bit (in my spare time) and will eventually post pictures. I don't have the skills to machine a nice base out of metal though.

Apple Silicon is powerful enough that it should hopefully attract game developers back to Apple. Hoping that happens, and that the real "pro" MacBooks coming in 2021 are what they're cracked up to be, I expect that enthusiast gamers will want to push the limits of the MacBooks just like PC enthusiasts do. So this research is all towards that end. Anyone considering an M1 Air for normal, non-psycho use will never have to worry about heat or performance, this thing is way beyond expectations. Heck, even when it throttles, it's still way faster than my old 2016 MBP 15", which is at max spec.

A slight word of caution: I've pretty much screwed my warranty at this point - I've accidentally marked up my batteries (the tack from the pads left discolored spots), scratched the inside of the case, and lost one of the screws. While the thermal pad mod is easy, safe and reversible, if you do it like 15 times in a row late at night ... well, accidents happen

I don't have the skills to machine a nice base out of metal though.

I'm retired now, but when I was working my job involved me working with copper every day, from wire to bar via foil of various thicknesses.
( as a completely off topic aside, myself and my colleagues that I worked with are now all in our 70's and we all still have a full head of hair ) I know, correlation does not imply causation 😀

Had I still been at work I could have fabricated something in copper, nothing too heavy just malleable enough to follow the contour of the external base of the Air, to fit between the rear feet. Another thought was to get a roll of heavy duty aluminium cooking foil and play around with layers of that ( externally of course )
This is all good fun and keeps me occupied as I'm having to stay home and shield due to health reasons.

 

[Mike Boreham]

Anyone considering an M1 Air for normal, non-psycho use will never have to worry about heat or performance, this thing is way beyond expectations. Heck, even when it throttles, it's still way faster than my old 2016 MBP 15", which is at max spec.

Haha....You took the words out of my mouth

My takeaway from this thread is that while it may be possible to squeeze a slightly higher Geekbench score than the already very high level, overall (unsurprisingly) Apple got the balance between performance, battery temp, case temp about right.

The other thought is a bit of puzzlement that people who enjoy this kind thing wouldn't be doing it to M1 MacBook Pros rather than MacBook Airs (perhaps they are).

Anyway I am enjoying reading about it so keep up the good work! Before M1 I was following the Intel Air cooling mod thread and could have conceived doing that myself if I had one of those, but def not an M1.

 

[Tenkaykev]

Haha....You took the words out of my mouth

My takeaway from this thread is that while it may be possible to squeeze a slightly higher Geekbench score than the already very high level, overall (unsurprisingly) Apple got the balance between performance, battery temp, case temp about right.

The other thought is a bit of puzzlement that people who enjoy this kind thing wouldn't be doing it to M1 MacBook Pros rather than MacBook Airs (perhaps they are).

Anyway I am enjoying reading about it so keep up the good work! Before M1 I was following the Intel Air cooling mod thread and could have conceived doing that myself if I had one of those, but def not an M1.

Hi Mike,

I had the 2020 i5 Air and had great fun doing the mods on that ( both shim and pad, with various pad thicknesses ) It got so I could strip the Air down on autopilot 😀
My wife now has my i5 Air, I've got the M1 base model and as you say I can't really see any real world advantage to fitting a thermal pad. Having said all that, I keep glancing over at my tech screwdriver kit which is on the back of the desk...

 

[juicy77]

Hi all i registered just to say thanks to TS i did this mod and got a significant improvement in performance and less throttling. My primary use is encoding and editing videos so this works out pretty well. Below are my results.

Time, Multi core R23 results (batt temp C)

Before
10min, 7276 (41) -> 20min 6412 (42) -> 30min 6093 (43)

After thermal pad mod.
10min, 7799 (43) -> 20min 7289 (44) -> 30min 7155 (46)

In my environment (28C ambient) the battery did heat up more, i'm not sure if that is due to the processor being able to draw more watts as there is less thermal throttling (i am still hitting 99C) or if it is heat transfer via the hot backplate. Will need to do more tests.