2020 Air Heatsink Modification Thread

[Wij]

I wonder if something like this on an angled stand would work when docked for work purposes:

Black Shark FunCooler Pro

14 ℃ Drop In 1 minute | Advanced TEC Technology | Various Devices Compatibility | iOS/Android App Management

uk.blackshark.com

Obviously similar products are probably available from other companies. No particular love for Xiaomi here. Might be better than a gel-pack.

 

[AirBud]

Still waiting on my 20x20x0.5mm shim.. 😔 The whole virus situation has really messed with shipping speeds for me. Hoping to try it next week.

 

[excelsior.ink]

So I bought a second 2020 Air base i3 for my wife. Did the heatsink mod on this one right out of the gate, and took the opportunity to get some good pics of the stock thermal paste application. Also worth noting, the paste on this thing is unlike any I've ever seen. First of all its black, and it has a gritty consistency, almost like there's a fine sand in it. You can hear it and feel it when scraping/cleaning it off. Very odd, don't know why it would have this grit added to it. Last pic shows the amount of paste I removed, minus the small amount that was cleaned off with rubbing alcohol at the end.

Wow! Very interesting. But now I have a lot of questions...
Have you just applied a different thermal paste? Or have you put also a copper shim? What are the results? Pictures?
Thank You!

 

[Robotronic]

Wow! Very interesting. But now I have a lot of questions...
Have you just applied a different thermal paste? Or have you put also a copper shim? What are the results? Pictures?
Thank You!

I used the same shim and paste as the first one I did for myself, so I didn't bother documenting the process for this one; just used the opportunity to get pre-mod pics of the stock paste. I didn't run any benchmarks on the second one either.

 

[throAU]

Not only is the copper shim helping the heat transfer to the heatsink but the 3rd party pastes we're using probably have better thermal conductivity.

Yup, maybe the paste is just crap from Apple. Given one of the above posts mentions it has some sort of grit in it.... sounds like it is...

I would argue this proves there is nothing wrong with the basic design. It's the poor execution of mating the heatsink to the CPU/GPU. That heatsink is plenty big for the job and the fan obviously moves enough air over it. There's just bad thermal conductance between the chips and sink from the factory.

Has anyone tried/compared simply replacing the thermal compound with something that isn't trash?

 

[Loog]

Has anyone tried/compared simply replacing the thermal compound with something that isn't trash?

I think @srkirt did this earlier in the thread on 2018 but there was oceans of MX4 thermal paste used. I've not seen anyone just do this on the MBA 2020. Agree would be interesting to see.

The design of both does appear to have tolerances that mean there is a quite a gap between CPU and heatsink (removed by the shim), this may account for why apple has chosen to have a compound that is solid when set.

 

[dcicer1]

I think @srkirt did this earlier in the thread on 2018 but there was oceans of MX4 thermal paste used. I've not seen anyone just do this on the MBA 2020. Agree would be interesting to see.

The design of both does appear to have tolerances that mean there is a quite a gap between CPU and heatsink (removed by the shim), this may account for why apple has chosen to have a compound that is solid when set.

I did it on a 2020 i3, I can’t see any difference at all. I’m just thinking about the gap between heat sink and die... but how much gap is there? Idk

 

[russell_314]

MacBook Air 2020: i7/16GB/1TB

Finally, here are pictures of what was underneath my heat sink!


Processor
The "2 cm" marking is aligned to the inner edge of the heatsink padding:

View attachment 910956


I should have aligned the "3 cm" marking to the inner edge of the heat sink padding:

View attachment 910957


Heat Sink
Underside of the heat sink, featuring a very shallow rounded-rectangle cutout (might be around 0.2 mm in depth):

View attachment 910959 View attachment 910960


The cutout relative to a 15x15 mm copper shim:

View attachment 910961 View attachment 910962


The "2 cm" marking is aligned to the long edge of the cutout:

View attachment 910964


The "3 cm" marking is aligned to the short edge of the cutout:

View attachment 910965


To gauge how thick the shims needed to be, I prototyped various shims with paper of different thicknesses. I found that the thickness of a business card was quite good. Unfortunately, the shims I bought were 0.6 mm thick, so I had to spend a few hours slowly filing them down 🤦🏻 In the end, they were thicker than needed, but thin enough for me to screw the heat sink down. The screws are incredibly small and the threaded ends are only around 2 mm in length.

I spread a thin layer of thermal paste onto the heat sink and "pasted" the shims on. This way, I know that the shims will be positioned properly when I lay the heat sink back onto the processor (with another layer of thermal paste).

View attachment 910966

Wow that's some awesome work. I don't think I'd have the nerve to do that kind of modification on a brand new $1000+ laptop. Taking apart my Mac mini just to add RAM made me very nervous. I hope your computer never has any issues with the motherboard though because I'm pretty sure Apple will say your modifications caused the damage.
[automerge]1588942434[/automerge]

you guys are nuts but in a good way. thoroughly enjoying watching this thread and seeing what ya come up with even though I won't attempt it myself.

Same here. If I was to buy a MBA it would be for light work so this would be unnecessary. If I wanted something more robust I would get a MBP. I am impressed with the level of bravery these people have though. I guess it depends on your financial situation but for me destroying any kind of Mac would be financially painful.

 

[srkirt]

All afternoon in bed with the MBA in my belly and the temperature does not exceed 55º. And we are now in Barcelona at 24º and the heat is beginning ...

 

[ilikewhey]

All afternoon in bed with the MBA in my belly and the temperature does not exceed 55º. And we are now in Barcelona at 24º and the heat is beginning ...

chrome with 4 tabs doesn't really tell anything, my 2009 mbp can have 5-6 chrome tabs with word and safari open and temp not exceed 55c.

 

[teksurv]

If you are thinking of repasting, and based on the earlier pic, I would, I highly recommend Kryonaut. I've used it on my minis and my Alienware, and the difference was noticeable.

 

[Loog]

I did it on a 2020 i3, I can’t see any difference at all. I’m just thinking about the gap between heat sink and die... but how much gap is there? Idk

@dcicer1, that's because (IMHO) the gap is just too much for paste alone. I estimate that 0.3mm shim is needed (possibly less but this is what we have tested as a community) to provide CPU > Heatsink physical contact. 0.5mm seems to also work but with just a little less thread contact for the screws that hold the CPU in place for the 2020 model, we've not seen what the differences are between 0.3mm or 0.5mm to fully establish if one is more preferable to the other. I doubt there is little to choose between them unless someone can prove to me otherwise.

 

[kinchee87]

Wow that's some awesome work. I don't think I'd have the nerve to do that kind of modification on a brand new $1000+ laptop. Taking apart my Mac mini just to add RAM made me very nervous. I hope your computer never has any issues with the motherboard though because I'm pretty sure Apple will say your modifications caused the damage.

Haha thanks for the wishes. I went through with the mod because it was non-invasive, in the sense that I didn't have to disassemble anything more than the bottom case and heat sink. Knowing that the mod could be "undone" if anything appeared to go wrong also helped.

 

[Loog]

Same here. If I was to buy a MBA it would be for light work so this would be unnecessary. If I wanted something more robust I would get a MBP. I am impressed with the level of bravery these people have though.

This is subjective @russell_314 ... the MBA is a wonderful platform but there are some applications that just don't suite or run efficiently. For me, Microsoft teams to being the only application in play and the CPU temp max out at 30% load? I'm sure you're not suggesting I need a Pro to video conference for more than 15 mins LOL?

This is why we've done what we have, the mods are quite safe if you take it easy and easily reversible if required. I hope you've liked our ventures, tense at times but fun and very rewarding

 

[srkirt]

[QUOTE = "ilikewhey, publicación: 28448137, miembro: 891227"]
Chrome con 4 pestañas realmente no dice nada, mi mbp 2009 puede tener 5-6 pestañas de cromo con word y safari abiertos y la temperatura no excede los 55c.
[/CITAR]

chrome with 4 tabs doesn't really tell anything, my 2009 mbp can have 5-6 chrome tabs with word and safari open and temp not exceed 55c.

My MBP early 2015 also but we talk about the MBA 2019 in this case.
[automerge]1588976718[/automerge]
I want to see how direct copper works in the CPU ... but I'm lazy ... the aluminum one I didn't do well ...

 

[BLiT_Racing]

If you are thinking of repasting, and based on the earlier pic, I would, I highly recommend Kryonaut. I've used it on my minis and my Alienware, and the difference was noticeable.

I've been lurking through this thread for a week or two and made an account to chime in here. Kryonaut is excellent paste, but degrades above 80*C. As such, I don't think the MBA is a suitable environment.

A quote (translated to English) from their website:

Kryonaut uses a special carrier structure, which does not begin drying processes even at 80 ° C. This carrier structure ensures that the nano-aluminum and zinc oxide components contained in the paste combine optimally, thus optimally compensating for unevenness in the heat transfer medium (e.g. the CPU) and a heat sink (e.g. heat sink) with excellent thermal conductivity.

Thermal Grizzly Kryonaut Wärmeleitpaste

Die Kryonaut-Wärmeleitpaste wurde speziell für die extremen Anwendungen und die höchsten Ansprüche der Overlocking-Community entwickelt.

www.thermal-grizzly.com

 

[raymanh]

MacBook Air 2020: i7/16GB/1TB with wind channel and heat sink mod

Here are more benchmark results to supplement my previous ones. I always ran the benchmarks with the MacBook Air fully charged and plugged in, and always waited for the CPU to cool down to around 42 ℃ before running the next benchmark. Each benchmark was executed at least three times and I averaged the results.


Geekbench 5
The single-core and graphics performance are only slightly better. I think it's because thermal throttling didn't happen much for those tests with the original heat sink. Multi-core performance, however, is about 20% higher. Click on a Mod result to view the best that I got.

Test Type	Original (Points)	Mod (Points)	Improvement
CPU (Single-core)	1,214​	1,271​	4.8%​
CPU (Multi-core)	2,992​	3,666​	22.5%​
Compute (OpenCL)	3,825​	3,944​	3.1%​
Compute (Metal)	4,285​	4,491​	4.8%​

Cinebench R20
Single-core and multi-core performance both improved by about the same amount, unlike the large performance improvement gap seen with Geekbench.

Test Type	Original (Points)	Mod (Points)	Improvement
CPU (Multi-core)	979​	1,110​	13.5%​
CPU (Single-core)	353​	390​	10.5%​
MP Ratio	2.77​	2.84​	2.7%​

View attachment 910875


Intel Power Gadget 3.7.0
Lastly, I did some plain-old stress tests, which show about 20% improvement in sustained CPU and GPU performance.

Test Type	Original (GHz)	Mod (GHz)	Improvement
CPU (All Thread Frequency, Scalar)	2.25​	2.70​	20.0%​
CPU (All Thread Frequency, AVX-256)​	2.12​	2.59​	22.2%​
CPU (All Thread Frequency, AVX-512)	2.00​	2.40​	20.0%​
Intel Graphics Test (Max Frequency)	0.90​	1.10​	22.2%​

View attachment 910904View attachment 910905View attachment 910906View attachment 910907

I don't remember if my MacBook Air was originally able to sustain max Turbo Boost for single-core.
View attachment 910927

Wow, great work! Surprising to see the largest percentage performance gain in a Geekbench test rather than Cinebench.

So was that just with a larger heatsink, or some kind of heat pipe too?

Have you tried reinstalling the original heatsink and testing it again? Maybe a significant part of the performance gain came just from the fact you wiped off the factory paste and reapplied some better stuff. Apple is known for poor quality thermal paste.

 

[kinchee87]

Wow, great work! Surprising to see the largest percentage performance gain in a Geekbench test rather than Cinebench.

So was that just with a larger heatsink, or some kind of heat pipe too?

Have you tried reinstalling the original heatsink and testing it again? Maybe a significant part of the performance gain came just from the fact you wiped off the factory paste and reapplied some better stuff. Apple is known for poor quality thermal paste.

Thanks! You can see photos of the mod in posts 966 and 1003. It was simply sealing up more of the wind channel and adding a copper shim to the stock heat sink. Nothing exotic

I only have Arctic MX-4 thermal paste, and as @Loog said, I think it would slowly ooze out of the gap between the heat sink and processor. Maybe another brave soul can reapply the paste and test the improvement.

 

[dcicer1]

Okay, today i did the mod by applying a little piece of copper between heatsink and die.. i didn't covered it by applying too much thermal paste but a tiny ball behind it. On cpu i did the same, a tiny ball of thermal paste. Temp seems to be decreased a little.. main problem is when you watch a film or video streaming stuffs, cpu kicks 100% of usage at 3.2 ghz and temperature obviously start getting up. This is the limit of 2020 MBA.

 

[srkirt]

Okay, today i did the mod by applying a little piece of copper between heatsink and die.. i didn't covered it by applying too much thermal paste but a tiny ball behind it. On cpu i did the same, a tiny ball of thermal paste. Temp seems to be decreased a little.. main problem is when you watch a film or video streaming stuffs, cpu kicks 100% of usage at 3.2 ghz and temperature obviously start getting up. This is the limit of 2020 MBA.

Thermal paste mark ???

 

[dcicer1]

Thermal paste mark ???

MX4

 

[Loog]

Okay, today i did the mod by applying a little piece of copper between heatsink and die.. i didn't covered it by applying too much thermal paste but a tiny ball behind it. On cpu i did the same, a tiny ball of thermal paste. Temp seems to be decreased a little.. main problem is when you watch a film or video streaming stuffs, cpu kicks 100% of usage at 3.2 ghz and temperature obviously start getting up. This is the limit of 2020 MBA.

Nice work @dcicer1, Is this the i3 ? What thickness shim did you use?

 

[IngerMan]

at @Loog , he stated 100% at 3.2 ghz so that would be the i3, i5= 3.5 ghz and i7 = 3.8 ghz

 

[Loog]

at @Loog , he stated 100% at 3.2 ghz so that would be the i3, i5= 3.5 ghz and i7 = 3.8 ghz

Thanks, I did assumed that @IngerMan, I was asking for the avoidance of doubt of other higher frequency CPU throttling to 3.2 as the machine hasn't been stated (or i've missed it)

 

[srkirt]

I just realized that the Pentium is not completely flat ... it seems curved !!! if so, the first sheet should be in the shape of the cpu ... it's an option ...