2020 Air Heatsink Modification Thread

[jonatanheyman]

To anyone thinking about putting a copper shim on top of the heatsink: What would be the point?

The shim between the CPU and the heatsink works because it closes the gap between the CPU and the sink. A thermal pad on top of the sink extends the size of the heatsink which can help in dissipating heat, possible connects it to the back plate which can also help, and possible also improves the way the air flows.

However, I can't see what one might think that a shim on top of the sink would do.

 

[twintin]

@twintin If you click your name and go to settings, the option is in there.

It appears the default is to have GPU hardware acceleration enabled.

 

[octoviaa]

To anyone thinking about putting a copper shim on top of the heatsink: What would be the point?

The shim between the CPU and the heatsink works because it closes the gap between the CPU and the sink. A thermal pad on top of the sink extends the size of the heatsink which can help in dissipating heat, possible connects it to the back plate which can also help, and possible also improves the way the air flows.

However, I can't see what one might think that a shim on top of the sink would do.

Same idea as with the thermal pad basically 'extend' the size of the heatsink, and copper-shim have much better thermal character capacity than the thermal-pad.

 

[AdamNC]

Well I won’t have to worry about the heat issue. Ended up buying a 13 in MacBook Pro i5 1T 16 gigs ram. Getting it set up now.

 

[BLiT_Racing]

I went to install a shim, got lazy, and just dropped a thermal pad on top of the heatsink with the factory thermal paste.

I'm honestly a little shocked that it worked as well as it did. My initial Cinebench R20 score was 987 on the day my i5/16/256 Air arrived. After the thermal pad, I was getting around 1100 points consistently, with a max of 1105.

Idle temps are a hair over 40*C, with my standard workflow (6-24 Chrome tabs, Tidal, Slack) idles in the low 50*C range. Heat-soak is an issue according to Geekbench, but for everyday use this seems like an extremely viable warranty-friendly option. I can slap together a how-to video for the extreme-newbs, if anyone would be interested.

Video of Cinebench R20 temps:

 

[rba1989]

I went to install a shim, got lazy, and just dropped a thermal pad on top of the heatsink with the factory thermal paste.

I'm honestly a little shocked that it worked as well as it did. My initial Cinebench R20 score was 987 on the day my i5/16/256 Air arrived. After the thermal pad, I was getting around 1100 points consistently, with a max of 1105.

Great result!

I can slap together a how-to video for the extreme-newbs, if anyone would be interested.

That would be very helpful!

 

[violentlychill]

I went to install a shim, got lazy, and just dropped a thermal pad on top of the heatsink with the factory thermal paste.

I'm honestly a little shocked that it worked as well as it did. My initial Cinebench R20 score was 987 on the day my i5/16/256 Air arrived. After the thermal pad, I was getting around 1100 points consistently, with a max of 1105.

View attachment 919151
View attachment 919153

Idle temps are a hair over 40*C, with my standard workflow (6-24 Chrome tabs, Tidal, Slack) idles in the low 50*C range. Heat-soak is an issue according to Geekbench, but for everyday use this seems like an extremely viable warranty-friendly option. I can slap together a how-to video for the extreme-newbs, if anyone would be interested.

Video of Cinebench R20 temps:

I would second a video even as a somewhat experienced user. I'm curious to see what coverage on the CPU/GPU you have with the pad. Also, apologies if you already mentioned, but what thickness was the pad?

Great results though. I'd much prefer to not have to cut out a copper shim and just go with this. When/if I get my 2020 MBA I'll probably try this route with fresh paste first.

 

[demize]

I went ahead and did the shim/re-paste/pad mods on my i5/8/512 this morning and so far I couldn't be happier. The entire process took about 40 minutes, most of which was spent trimming/filing the second shim.

Here is what I used:

Honbay15mm x 15mm .3mm copper pads x2 (one of which was cut down to roughly 15mm x ~7mm
Thermal Grizzly Kryonaut Thermal Paste
Thermal Grizzly Minus 8 Thermal Pad 30mm x 30mm x 2.0mm

Here are the results:

Pre mod:

Cinebench R20: 923
Geekbench 5:Single 1055 Multi 2709

Through R20 the temp was essentially maxed at 100 with the fans spinning up quite loudly at the end of the test.

Post mod:

Cinebench R20: 1120 +21.3%
Geekbench 5: Single 1177 +11.5% Multi 3371 +24.4%

What's even more remarkable is during the R20 tests temps never got above 93 degrees and seemed to hover around 90. Through both Cinebench and Geekbench the fans never once spun up which is probably the most satisfying part of the mod. I didn't buy the MBA for "pro tasks" but having the fans spin up almost immediately on video calls was tremendously annoying.

I'll continue to go about my normal tasks and report back anything interesting.

Lastly a big thank you to everyone on this forum for helping make the 2020MBA that much better... now if someone could just figure out how to make the webcam not so terrible this would be the perfect ultrabook!

UPDATE: I decided to hook my MBA up to my "desktop" setup: Dell S2716DG, Logitech G610, Logitech G203, Logitech C920, Audio-gd NFB-11.32 and Blue Yeti. These devices are all fed into a USB switch, which combined with the HDMI out on my monitor allow me to have both my desktop and either PC work laptop or my MBA connected simultaneously. Pre-mod connecting my MBA to this setup via the Apple USB-C Digital A/V multiport adapter resulted in the fans being on constantly, even if I just had a couple tabs in Safari open. Opening FaceTime or Zoom immediately caused the fans to max. and the processor to stay at 100 degrees.

I'm happy to say that post mod, I connected my MBA to this setup and never once heard the fans. I had Mail, Outlook, Excel, Spotify, Acrobat DC, Safari and Messages open, temps hovered around 78 degrees. While on a test FaceTime call the temp increased to about 88, and on a test Zoom call the temperature increased to 95 but the fans never once came on and eventually the temp settled at about 90.

Again, it's really hard to explain how much better this machine is post mod (for me).

 

[kinchee87]

What would be interesting is to see if the temperature around the Air has increased noticeably after the mod under some standardised max processor load. For example, here are my i7 temperatures after my wind channel + shim + thermal pad mods (15 min into running "All Frequency" stress test by Intel Power Gadget):

 

[vyruzreaper]

I have just ordered a 2020 MBA i5 526 8gig. All this talk about the heat and all has me nervous. I don’t do VC’s or any real hard use work, mostly weather tracking emails, Quicken, Number crunching. Should I be concerned? And I had a idea I may try and wanted to see if you think it would be worth it. I noticed the pattern on the heat sink. I was wondering if getting heat pad and cutting strips shown in yellow so air could flow through a little would help? Any thoughts?
View attachment 919082

No, except Quicken might make the fans blow at max speeds. Random apps have random outcomes with the fan.
[automerge]1590614877[/automerge]

I went to install a shim, got lazy, and just dropped a thermal pad on top of the heatsink with the factory thermal paste.

I'm honestly a little shocked that it worked as well as it did. My initial Cinebench R20 score was 987 on the day my i5/16/256 Air arrived. After the thermal pad, I was getting around 1100 points consistently, with a max of 1105.

View attachment 919151
View attachment 919153

Idle temps are a hair over 40*C, with my standard workflow (6-24 Chrome tabs, Tidal, Slack) idles in the low 50*C range. Heat-soak is an issue according to Geekbench, but for everyday use this seems like an extremely viable warranty-friendly option. I can slap together a how-to video for the extreme-newbs, if anyone would be interested.

Video of Cinebench R20 temps:

This is what I've been saying!!!! Thermal Pad might be the perfect way to mod for those that are not keen on screwing with logic boards etc.

Great results and thanks for posting!

 

[srkirt]

 

[BLiT_Racing]

That would be very helpful!

I would second a video even as a somewhat experienced user. I'm curious to see what coverage on the CPU/GPU you have with the pad. Also, apologies if you already mentioned, but what thickness was the pad?

The pad was 1.5mm thick. See the link below for the exact product I used.

ARCTIC TP-2 (APT2560): Economic Thermal Pad, 50 x 50 x 1.5 mm (1 Piece) - Thermal pad, Excellent heat conduction, Low hardness, Ideal gap filler, Easy installation, thermal pads gpu, thermal pad cpu -

ARCTIC, a leading manufacturer of low-noise PC coolers and components, initiated and continues to shape the trend towards quiet cooling systems. In addition to an extensive range of CPU coolers and fans, ARCTIC produces high-quality monitor arms, wall mounts, thermal paste as well as thermal pads...

www.amazon.com

I'll edit a video together then! I've got GoPro footage of the entire ordeal.

 

[RiaKoobcam]

I'll edit a video together then! I've got GoPro footage of the entire ordeal.

Given it's a GoPro, I expect an extreme soundtrack complete with guitar solos and an EDM section where you drop the bass (maybe on the reveal of the heat sink?!) 😂

 

[daystar]

What would be interesting is to see if the temperature around the Air has increased noticeably after the mod under some standardised max processor load. For example, here are my i7 temperatures after my wind channel + shim + thermal pad mods (15 min into running "All Frequency" stress test by Intel Power Gadget):

View attachment 919223

Here is mine after same 15min all frequency test (i7 shim+repaste+heatpad)

 

[ctjack]

Here is mine after same 15min all frequency test (i7 shim+repaste+heatpad)

Wow, your results are incredible. Could you be so kind to share with us the exact steps which you've made to achieve this? Also which parts you used?
I've read your previous posts and it seems like you used thermal grizzly paste and pads.
So was it CPU+grizzly paste on it+copper shim+grizzly paste+original heatsink+grizzly paste+heatpad?

 

[excelsior.ink]

Here is mine after same 15min all frequency test (i7 shim+repaste+heatpad)

View attachment 919311

It can keep ~3GHz in sustained load. Nice job!

 

[daystar]

Wow, your results are incredible. Could you be so kind to share with us the exact steps which you've made to achieve this? Also which parts you used?
I've read your previous posts and it seems like you used thermal grizzly paste and pads.
So was it CPU+grizzly paste on it+copper shim+grizzly paste+original heatsink+grizzly paste+heatpad?

Sure! It was CPU-Thermal Grizzly Hydronaut-Copper shim that I made 22x16x0.3mm lapped with 1000 and 1500 sandpaper-Hydronaut-Heatsink-on top of the heatsink Thermal Grizzly Minus Pad 8 30x40mm.

Just ran trough Cinebench R20 with factory fan settings and got 1240 points. Temps never went over 82c and fans kept at minimum speed (2700).
So overall excellent combination with the Thermal Grizzly products

 

[Ghostrider72]

No, except Quicken might make the fans blow at max speeds. Random apps have random outcomes with the fan.
[automerge]1590614877[/automerge]



This is what I've been saying!!!! Thermal Pad might be the perfect way to mod for those that are not keen on screwing with logic boards etc.

Great results and thanks for posting!

I don't really understand what he did, basically he did put just a thermal pad over the heatsink? Did I understand well?
[automerge]1590656262[/automerge]

Sure! It was CPU-Thermal Grizzly Hydronaut-Copper shim that I made 22x16x0.3mm lapped with 1000 and 1500 sandpaper-Hydronaut-Heatsink-on top of the heatsink Thermal Grizzly Minus Pad 8 30x40mm.

Just ran trough Cinebench R20 with factory fan settings and got 1240 points. Temps never went over 82c and fans kept at minimum speed (2700).
So overall excellent combination with the Thermal Grizzly products

And what about if just putting a lot of thermal paste over the cpu, then the heatsink and finally the thermal pad? Did anyone try this? I have been reading lots of messages but cannot figure out why you need to put a piece of copper sandwiched between the cpu and the heatsink while probably a lot more thermal paste could do. Can you provide more info please?

 

[Nütztjanix]

ImAnd what about if just putting a lot of thermal paste over the cpu, then the heatsink and finally the thermal pad? Did anyone try this? I have been reading lots of messages but cannot figure out why you need to put a piece of copper sandwiched between the cpu and the heatsink while probably a lot more thermal paste could do. Can you provide more info please?

Because that is exactly what Apple did in the factory, and the reason for the bad thermal performance in the first place. The gap between CPU and heatsink is just to big. You need to fill it with something that conducts heat better than thermal paste, and doesn’t tend to change shape or consistency over time.

 

[Ghostrider72]

Because that is exactly what Apple did in the factory, and the reason for the bad thermal performance in the first place. The gap between CPU and heatsink is just to big. You need to fill it with something that conducts heat better than thermal paste, and doesn’t tend to change shape or consistency over time.

Thanks a lot, now it's clear. I thought that just pushing more paste would alleviate the supposed issue but at this stage I see being these the premises it requires a deeper intervention. Thanks again.

 

[daystar]

I don't really understand what he did, basically he did put just a thermal pad over the heatsink? Did I understand well?
[automerge]1590656262[/automerge]

And what about if just putting a lot of thermal paste over the cpu, then the heatsink and finally the thermal pad? Did anyone try this? I have been reading lots of messages but cannot figure out why you need to put a piece of copper sandwiched between the cpu and the heatsink while probably a lot more thermal paste could do. Can you provide more info please?

Because thermal paste works best when applied as a thin layer, it's job is to fill the micro holes/non flat surfaces between heatsink and the die and in this case the space between the die and the heatsink is quite big so that's why you need a copper shim between the layers. Putting a lot of thermal paste is exactly what apple did ( it was quite thick)and we are trying to avoid that..

 

[Ghostrider72]

Because thermal paste works best when applied as a thin layer, it's job is to fill the micro holes/non flat surfaces between heatsink and the die and in this case the space between the die and the heatsink is quite big so that's why you need a copper shim between the layers. Putting a lot of thermal paste is exactly what apple did ( it was quite thick)and we are trying to avoid that..

That's a great explanation, thank you very much, I've got the picture now.

 

[RiaKoobcam]

Sure! It was CPU-Thermal Grizzly Hydronaut-Copper shim that I made 22x16x0.3mm lapped with 1000 and 1500 sandpaper-Hydronaut-Heatsink-on top of the heatsink Thermal Grizzly Minus Pad 8 30x40mm.

Just ran trough Cinebench R20 with factory fan settings and got 1240 points. Temps never went over 82c and fans kept at minimum speed (2700).
So overall excellent combination with the Thermal Grizzly products

That is extremely good work, especially with the i7. Great job!

 

[IngerMan]

I read one poster had better heat and fan results after updating to 10.15.5.

I hope Apple did not dumb things down. Can a poster that did the mod confirm they are getting same results from .4 version.

 

[Saul Giordani]

The results obtained after putting the thermal pad ON TOP of the heatsink make me think that the bigger problems of the cooling system are:

• air flow, as already said
• but especially no good contact between heatsink and bottom cover of the laptop