2020 Air Heatsink Modification Thread

[Viper-FFM]

I use my Air mostly on the couch on my lap.

Reasons for the easy peasy pad-solution:
- Warranty remains as it is plug and play
- Less or zero danger of damaging anything on the chip-level
- cheap
- Performance gain plus lower overall temps

Reasons against:
- Slightly warmer case but honestly it isn't really noticeable


Please note that the surface area of your Air is huuuuge. It would be more than surprising when the back lid would reach a temp delta of even 4C. I have a thermal scanner in my office but honestly, as I am working on cooling of rugged servers on a daily basis there's really no reason to test this...

ATM I have 6 tabs open in safari plus Mail plus Wifi over Bluetooth/iPhone. Temp 42 degrees celsius while the fan is completely down at zero. With this mod, the Air is what I was hoping for since 2013. It is super energy efficient while having the potential for medium to high workloads when needed. But for your daily office work its just a slim, cool, performant device with no Touch Bar and a decent keyboard.
[automerge]1590861945[/automerge]
I re-ran Geekbench just now as I did it only 2-3 minutes after the Cinebench score which wasn't enough time for the processor to go back to idle temps.

Therefore
OS10.15.4 Stock
Cinebench R20 891pts
Geekbench5 Singlecore 1081 Multicore 2592


OS10.15.5 Stock
Cinebench R20 979pts
Geekbench5 Singlecore 1118 Multicore 2778


OS10.15.5 with 2mm PAD only between sink and chassis
Cinebench R20 1100pts
Geekbench5 Singlecore 1191 Multicore 3421

I guess that's pretty huge.

 

[Loog]

keep in mind my score is without any fan activity. With fan at full blast i got 1150 /3630 and 1100 in cinebench and with Catalina 10.15.4. The benefit is, that the case remains cool.

@nill1234 Looking at @Viper-FFM original results before upgrading to 10.15.4 there was around 100 points difference without modification which does seem to indicate the latest update may also have some tweaks included, some voodoo has happened with my battery going from 98% to 101% peak capacity. It would be interesting to compare your results when you take the upgrade.

Where my mind is, we seem to have established there are differing tolerances between the heatsink and the CPU, the shim addresses this, better paste such as Grizzley alone also has some great benefits.

If the heat pad increases the thermal performance, is this working by increasing the thermal area i.e. the back of the case or is this also applying light pressure on the case/heatsink which is causing better thermal contact between the CPU and heatsink? It may be a little bit of both here.

Both methods are seeing really great results, keep up the good work!

 

[Saul Giordani]

Just run the weekly cleaning from Clean My Mac, last week the temperature went up to 75º, now after the thermal pad mod didn't exceed the 62º. I'd really love to do add the shim pad but I don't want to lose the warranty.

 

[srkirt]

[CITA = "Saul Giordani, publicación: 28517344, miembro: 1211657"]
Simplemente ejecute la limpieza semanal de Clean My Mac, la semana pasada la temperatura subió a 75º, ahora después de que el mod de almohadilla térmica no excedió los 62º. Realmente me encantaría agregar la cuña, pero no quiero perder la garantía.
[/CITAR]
I still have not paid the first installment and it is already scratched and inside it looks old ... hahaha if something breaks to pay ...

 

[kinchee87]

@nill1234 Looking at @Viper-FFM original results before upgrading to 10.15.4 there was around 100 points difference without modification which does seem to indicate the latest update may also have some tweaks included, some voodoo has happened with my battery going from 98% to 101% peak capacity. It would be interesting to compare your results when you take the upgrade.

Where my mind is, we seem to have established there are differing tolerance s between the heatsink and the CPU, the shim addresses this, better paste such as Grizzley alone also has some great benefits.

If the heat pad increases the thermal performance, is this working by increasing the thermal area i.e. the back of the case or is this also applying light pressure on the case/heatsink which is causing better thermal contact between the CPU and heatsink? It may be a little bit of both here.

Both methods are seeing really great results, keep up the good work!

To test this theory, people could replace their thermal pad with corrugated cardboard and see the effect. I have no idea if it would lead to any improvement over the stock configuration 🤷🏻‍♂️

 

[hmhmmhmmm?]

At last i got all the parts needed, took forever! 😕

2020 i5/16/256

Added shim 0.4mm
Gelid GC Extreme

before shim mod - fan auto

C20
multi avg 1012
single avg 367

Geekbench
single 1182
multi 3106

After shim mod
fan went to ~3600during geekbench test.

C20
Multi 1035
single 371

Not that much of a difference?
Idle is around 40, pretty sure that was the same as before mod.

Are these results ok? only made a difference in geekbench multi.
Also my temp is spiking coremax?

 

[violentlychill]

At last i got all the parts needed, took forever! 😕

2020 i5/16/256

Added shim 0.4mm
Gelid GC Extreme

before shim mod - fan auto

C20
multi avg 1012
single avg 367

Geekbench
single 1182
multi 3106

After shim mod
fan went to ~3600during geekbench test.

View attachment 920127

C20
Multi 1035
single 371

Not that much of a difference?
Idle is around 40, pretty sure that was the same as before mod.

Are these results ok? only made a difference in geekbench multi.
Also my temp is spiking coremax?

Maybe your setup was fine before? Idling at 40 seems to be what people get after the mod. I really am starting to think there are QC and consistency issues with the thermal solution and yours was fine out of the box while others are reporting much worse idle temps.

 

[hmhmmhmmm?]

Maybe your setup was fine before? Idling at 40 seems to be what people get after the mod. I really am starting to think there are QC and consistency issues with the thermal solution and yours was fine out of the box while others are reporting much worse idle temps.

Could be a QC issue, but still weird that the heatsink isn’t flush with the cpu? At least my multi geekbench improved. Also noticed my coremax temp spikes compared to the pkg temp. But after i run a C20 test it follows the same temp, and at some point it starts spiking again when idle?

 

[IngerMan]

Could be a QC issue, but still weird that the heatsink isn’t flush with the cpu? At least my multi geekbench improved. Also noticed my coremax temp spikes compared to the pkg temp. But after i run a C20 test it follows the same temp, and at some point it starts spiking again when idle?

This video talks about the spikes with the Gen 10 at 5:23 . It’s designed that way

 

[srkirt]

I have achieved my goal !!! I have outperformed a 4 core i7 !!! It has been filling all the heat sink (not Pad) and on top the fin radiator. I'm still thinking that it could create more pressure at the fins entrance and expand the air at the exit, which is called the Venturi effect. If you could stop the entry speed by compressing the air, it would come out faster and hotter.
At the moment I have the fastest 2019 in the entire forum simply by using trial / error and a little ingenuity. It is incredible that making the bermack does not exceed 80º and that it is at 35 / 38º with the Chrome open at rest.

 

[hmhmmhmmm?]

This video talks about the spikes with the Gen 10 at 5:23 . It’s designed that way

Thank you for the video. But it doesn’t answer my question. They are talking about cpu freq. My question is about temperature. Take a look at my screenshot #1581 and you can see how it’s spiking when idle. I’m not sure if that is normal?

After i run a test both PKG and CoreMax temp. is the same. After a while coretemp starts spiking like in the screenshot.

 

[srkirt]

Could be a QC issue, but still weird that the heatsink isn’t flush with the cpu? At least my multi geekbench improved. Also noticed my coremax temp spikes compared to the pkg temp. But after i run a C20 test it follows the same temp, and at some point it starts spiking again when idle?

Logical ... when more heat dissipates and in this case it flows towards the laptop lid if it does not have air flow it goes back ... it is elementary physics, the ideal thing (Mr. Tim) is that the CPU was upside down so that without ventilation the heat would come out in front of the keyboard ... And this Cia has the most prodigious minds on the planet ??? Hahahaha They go for money, I had Apple shares and I made a lot of money but when I saw how the company works I was amazed ... investors and managers (including the son of Al Gore) are only interested in making quarterly profits and they have have to get there anyway. In this case it is clear that making the system cheaper so as to at least have less expenses at less income. The proof is clear ... I have an MBP Early 2016 and inside it is another universe !! nothing to do with now.

 

[IngerMan]

Thank you for the video. But it doesn’t answer my question. They are talking about cpu freq. My question is about temperature. Take a look at my screenshot #1581 and you can see how it’s spiking when idle. I’m not sure if that is normal?

After i run a test both PKG and CoreMax temp. is the same. After a while coretemp starts spiking like in the screenshot.

Kinda like this?

 

[srkirt]

4K 58º/ 63º en Vimeo.
We have leftover processor to see up to 8K (and more you those of 2020). It's just that YouTube has an incompatible codec system and on Vimeo if it is, you can see that your machine is left over with the Srkirt Kit ... hahahaha

 

[hmhmmhmmm?]

Kinda like this?


View attachment 920167

Not quite. My idle PKG is ~40 while Coremax is spiking to 60-70 all the time as in my screenshot #1581
After i run C20 both PKG and CoreMax are idle at ~40 ?

 

[srkirt]

[QUOTE = "hmhmmhmmm ?, publicación: 28518074, miembro: 1214031"]
No exactamente. Mi PKG inactivo es ~ 40 mientras que Coremax está aumentando a 60-70 todo el tiempo como en mi captura de pantalla # 1581
Después de ejecutar C20, tanto PKG como CoreMax están inactivos a ~ 40?
[/CITAR]
Why don't you expel the heat !!! with the modification without ventilation it is a microwave oven !!!

 

[IngerMan]

Not quite. My idle PKG is ~40 while Coremax is spiking to 60-70 al the time as in my screenshot #1581
After i run C20 both PKG and CoreMax are idle at ~40 ?

Can you run R20 and just take a picture of Intel power gadget only about half way thru

 

[hmhmmhmmm?]

Can you run R20 and just take a picture of Intel power gadget only about half way thru

Here is a screenshot halfway through

[QUOTE = "hmhmmhmmm ?, publicación: 28518074, miembro: 1214031"]
No exactamente. Mi PKG inactivo es ~ 40 mientras que Coremax está aumentando a 60-70 todo el tiempo como en mi captura de pantalla # 1581
Después de ejecutar C20, tanto PKG como CoreMax están inactivos a ~ 40?
[/CITAR]
Why don't you expel the heat !!! with the modification without ventilation it is a microwave oven !!!

Not sure what you mean? Should i do something else when adding the shim?

 

[IngerMan]

That looks normal. And your score was good. Now let it idle to 40c for a few minutes doing nothing but moving your trackpad Mouse to keep the screen on and snap another picture.

 

[hmhmmhmmm?]

Can someone confirm the same behaviour in temperature when idle with stable PKG and spiking CoreMax?

[automerge]1590889566[/automerge]

That looks normal. And your score was good. Now let it idle to 40c for a few minutes doing nothing but moving your trackpad Mouse to keep the screen on and snap another picture.

See #1595 🙂

 

[excelsior.ink]

Logical ... when more heat dissipates and in this case it flows towards the laptop lid if it does not have air flow it goes back ... it is elementary physics, the ideal thing (Mr. Tim) is that the CPU was upside down so that without ventilation the heat would come out in front of the keyboard ... And this Cia has the most prodigious minds on the planet ??? Hahahaha They go for money, I had Apple shares and I made a lot of money but when I saw how the company works I was amazed ... investors and managers (including the son of Al Gore) are only interested in making quarterly profits and they have have to get there anyway. In this case it is clear that making the system cheaper so as to at least have less expenses at less income. The proof is clear ... I have an MBP Early 2016 and inside it is another universe !! nothing to do with now.

I don't think cheap is Apple's motivation here. My 2020 MBA has a list price in USA of 1499 USD (i5/16/512). In my country with VAT included I paid amost 1800 EUR. These are not cheap laptops. They pay a premium to Intel to get the best binned processors and in the case of the 2020 MBA they get processors no one else get. The display is premium and they pay a premium to LG. Everything in this laptop in premium. And the quality of the components is usually confirmed by the fact that Macbooks usually last for years. How much money can they save by using cheap paste? The main problem (at least with the 2020 MBA) is that the heatsink has a recessed design which people on this forum fixed it with a 0.3-0.5 mm copper thermal pad (shim). Why would they design it like that if it is inneficient? To me this looks on purpose. Likely to stop the MBA enter the MBP teritory.

 

[nill1234]

At last i got all the parts needed, took forever! 😕

2020 i5/16/256

Added shim 0.4mm
Gelid GC Extreme

before shim mod - fan auto

C20
multi avg 1012
single avg 367

Geekbench
single 1182
multi 3106

After shim mod
fan went to ~3600during geekbench test.

View attachment 920127

C20
Multi 1035
single 371

Not that much of a difference?
Idle is around 40, pretty sure that was the same as before mod.

Are these results ok? only made a difference in geekbench multi.
Also my temp is spiking coremax?

It looks good you got more performance without the high temperatures. You didn't even reach 100 degrees.

 

[dcicer1]

Have you ever seen those temp?

 

[srkirt]

Air compression test with glass sheet ...
[automerge]1590933313[/automerge]
[QUOTE = "dcicer1, publicación: 28518986, miembro: 1216119"]
¿Alguna vez has visto esos temp? ags
[/CITAR]
Yo 26º.

 

[IngerMan]

Can someone confirm the same behaviour in temperature when idle with stable PKG and spiking CoreMax?

View attachment 920196
[automerge]1590889566[/automerge]

See #1595 🙂

Its something running in the background. I can simulate it by opening my turbo boost switcher. See to the left I had it open, not even on just open. Then I enabled it in the middle then I closed it in the end. But the key here it is normal temps. If your so concerned undo your thermal paste mod and put it back to stock.