2020 Air Heatsink Modification Thread

[adrianstuartt]

this is a heatsink modification thread not 'fixing apple's failed engineering'. you only see what you want to see perhaps you should also read Erehy Dobon and few other comments regarding the thermal system then you know if its 'creating a problem when there isn't one' or not. We all can agree that butterfly keyboard and macbook pro flexgate issue is the real failing design or oversight from apple engineers which eventually lead to a huge issue, but it is different story here.

all these hiccups that have been mentioned over and over again could possibly be optimisation problem for now, and ive said it wont be a drastic change even after optimisation but improvement to what it should be over time and how funny y'all complicate yourself and twist my words where i find it hilarious and also speechless at the same time. This happens is all due to over-expectation over an ultra thin entry level laptop and overreact on how a 3 month old baby behave.

by one of the forumer Erehy Dobon:
Every design choice is a compromise: power, weight, bulk, cost, noise, convenience, battery life, etc.

The point of a MacBook Air isn't to harness the maximum capability of the CPU 100% 24x7. Maybe they could add a heat pipe and a CPU fan but that would force an increase in the case thickness. Maybe you need a bigger battery to handle the power needs. All of sudden you end up with a MacBook Pro.

this says it all, Macbook air is Macbook air and it was only built to target at casual user for casual use in the very first place, its a bonus where it still capable of performing some power intense task, just same like any other windows laptop at the same category. if you want power and consistent performance, seriously i believe an extra few hundred bucks is not really a matter for yall, go for pro series already and most of your problems will be solved instantly, simple as that. ?‍♂️

LTT proved essentially that if the standoffs for the cpu heatsink were shorter to account for the milled out heatsink, the heatsink could have made sufficient contact with the cpu to run much cooler and quieter. Alternatively, apple could have just... not... milled out the heatsink. How is this not a failing of the design or manufacturing? The only 'optimization' you could possibly be talking about is further crippling the cpu

 

[jackiez]

that level of comprehension and blind stubbornness, whatever ?‍♂️

 

[adrianstuartt]

that level of comprehension and blind stubbornness, whatever ?‍♂️

Yeah, no matter how much we try we cannot seem to bring you up to our level. Your argument is basically that 'apple didn't fail in hardware, they only failed in software!' LOL

 

[srkirt]

Jackiez you are an egocentric of noses ... but I don't know what you do in this thread ... let people do with their things what comes out of their eggs. You write very well but you are not God and you are not right ... I started this !!! You understand ??? and we all saw the result !!! lower temp and increase turbo boost ... What you can not get is to spend more energy than the one designed for the laptop and if your 10W son there you will not go ... but if you can cool it and that fudge of a Huge thick of pasta is a deliberate apple maneuver. We have all seen it !!!

I repeat, if you do not like this thread or what people think, LEAVE.

 

[jackiez]

Yeah, no matter how much we try we cannot seem to bring you up to our level. Your argument is basically that 'apple didn't fail in hardware, they only failed in software!' LOL

well for references, make sure your level is beyond them then ?‍♂️

Every design choice is a compromise: power, weight, bulk, cost, noise, convenience, battery life, etc.

The point of a MacBook Air isn't to harness the maximum capability of the CPU 100% 24x7. Maybe they could add a heat pipe and a CPU fan but that would force an increase in the case thickness. Maybe you need a bigger battery to handle the power needs. All of sudden you end up with a MacBook Pro.

Guys, just because some people don't understand how something works – and seem to think it's a "deal-breaker" – doesn't mean it is inferior. The proof is in the pudding: does the machine get hot? Are lots of MacBook Air owners complaining about excessive heat or throttling or fan noise?

And it's really amusing to see people debating airflow in a three-dimensional space based upon two-dimensional photos.

The thermal design, like every other component, is chosen to achieve specific goals. What this heated discussion boils down to is a disagreement with Apple's goals, not Apple's competency at meeting those goals.

You want to run your Honda Civic as if it was a pickup truck? Don't blame Honda if it doesn't meet your desires.

Yes it is. The most amusing thing is that some people are commenting on this design like they are experts in thermodynamics, but they don't even understand how it works in the first place.

So let's be honest. How many people commenting here actually took the time to look at the design before they commented? And how many actually understand it, even after all this explanation?

By the looks of it, not many.

If you don't understand how something works, how are you to make a value judgement on whether it is good design or not? I work with heatsinks every day, so I know a little bit about what I'm talking about.

Now some may make a valid observation that the machine gets hot – if in fact it does – but unless I missed something, none of those commenting have the machine in their hands, and none have conducted thermodynamic tests on them or any other machine that I can see.

They are relying on a few YouTube videos that are taxing the machine far beyond what the design was initially intended for. MacBook Air's are not designed as gaming machines, and nor are they designed as data-intensive number crunchers. If the cooling system is throttling the CPU because it is being taxed to its core for 10-15 minutes at a time – something it is not designed for – then it is not a design fault, it is a user error.

You wouldn't rev your passenger car to 9000rpm compared to a race engine, so why complain about a consumer model CPU being throttled when asked to perform tasks beyond its intended design?

I can't help thinking he and everybody else who declares the computer unusable just really expected Apple to sell a refreshed Macbook Pro for $999. I would love it if they did, and I am no business person, but that would seem to be like a really stupid move for a company designed around making money.

[automerge]1592933541[/automerge]

 

[christophosphorus]

@jackiez, with due respect, this isn't as difficult as you're making it sound. There is no argument about the fact that Apple screwed up on the thermal design of the MBA.

Is the Air an entry-level Mac that isn't meant to run at 100% 24/7? Sure.

Are there power users in here that would likely benefit from a Pro? Of course.

But did Apple drop the ball (either unintentionally, or intentionally) in designing the thermals? Absolutely. We've opened up the guts of the laptops we're typing on this very second and have seen how deficient the system is (especially the extra space between the CPU and heatsink that, in the name of physics, does not make sense).

You can quote others all you want, but it is downright laughable that any-level MacBook Air suffers when running a couple of programs at the same time, especially during today's day and age of WFH and videoconferencing.

None of us here expect our MBAs to edit multiple 4K videos simultaneously while gaming. But I think we all know that even the most basic tasks that a $200 Chromebook can handle are an uphill battle for the MBA.

That's my $.02.

 

[SeeG]

Hi all,

It's been a while ... Hope you're all fine ... Didn't have time to follow the thread but I wanted to show you this video :

Maybe you have already seen it.

By the way, as I said, I returned the MBA for a 10th gen I5 MBP that I'm using for 2 weeks now : No comparison possible, MBP is a delight even if I love the MBA but that cooling thing f....d everything !!

 

[LarsMac]

Hey, thanks a lot to all of you for those great resources!
I'm wondering if the shim or heat pad mod does improve the results of a gpu demanding benchmark like Unigine Heaven. Youtuber MaxTech said in his comparison video that the MBA i5 reaches 9.0 fps vs. 10.7 on base Macbook Pro. Would be interesting if you were able to close that gap..

 

[IngerMan]

I will say even though the thermal Pad only is huge gains in multi core applications about 13% for me, the new thermal paste and copper shim made much more difference in heat. It only gained an added 3-4% in multicore compared to the pad only but the heat reduced is the biggest bonus for new paste and shim, like running GB5 Hovered around 75-80C With Shim and Pad. With the thermal Pad only, there is a portion of the test I always got very close to 100C. It never broke 80C on Cinebench R20 with Shim and Pad but with Pad only, it hit 100C. Thats a 20C difference when using intensive apps. WOW

 

[jackiez]

that level of comprehension and blind stubbornness, whatever ?‍♂️

?‍♂️

 

[adrianstuartt]

pot, meet kettle

 

[IngerMan]

Hi all,

It's been a while ... Hope you're all fine ... Didn't have time to follow the thread but I wanted to show you this video :

Maybe you have already seen it.

By the way, as I said, I returned the MBA for a 10th gen I5 MBP that I'm using for 2 weeks now : No comparison possible, MBP is a delight even if I love the MBA but that cooling thing f....d everything !!

Take a good look at the MBP threads. Plenty of fan and heat threads for your beloved 13”

ever hear “people who live in glass houses don’t Throw stones”.

 

[kazune_karin]

Take a good look at the MBP threads. Plenty of fan and heat threads for your beloved 13”

ever hear “people who live in glass houses don’t Throw stones”. Go back.

Just a tiny comment.
MacBooks eventually get hotter during stress tests, no matter Air or Pro.
Performance wise, Apple limits the performance by TDP wall (max time for turbo boost).

Guys in this thread made full use of the TDP wall by preventing thermal throttling before hitting the TDP wall.
For not prolonged usage, MacBook Air and Pro 10th gen can have very comparable performance, eapecially after mod.

 

[fcracer]

I will say even though the thermal Pad only is huge gains in multi core applications about 13% for me, the new thermal paste and copper shim made much more difference in heat. It only gained an added 3-4% in multicore compared to the pad only but the heat reduced is the biggest bonus for new paste and shim, like running GB5 Hovered around 75-80C. With the thermal Pad only, there is a portion of the test I always got very close to 100C. It never broke 80C on Cinebench R20 but with Pad only, it hit 100C. Thats a 20C difference when using intensive apps. WOW

View attachment 926813 View attachment 926814

With the shim and paste making such a big difference in temperature at the CPU, it means that the heat has gone somewhere. That extra 20C has gone into either the air or the thermal pad.

For my personal use, I'm trying to get the thermals to be just below the point where performance degrades under worst case sustained use. Anything beyond that is (IMO) introducing too many tradeoffs.

For example, I had originally tried covering the entire heatsink with a 1.5MM Arctic pad. The performance of the laptop was ~3950 on GB5 MC from a cold start. I then reduced it to about 3/4 of the surface area and the performance stayed the same at ~3950.

I still found the heat transfer to the bottom case to be too much and reduced the pad's surface area by 1/2, so the actual pad is now very small and barely touches the four screws holding down the heat sink. Performance is now ~3850 so not much given up but a much cooler bottom cover.

The tradeoff with the smaller pad is that the fans do spin up more now, but I'm ok with that since that's what they are there for. Fans should spin up under heavy and sustained usage like video encoding or raw image exporting. With the small pad, Zoom and other VC tech does not spin up the fan and it stays well below the temp threshold of 100C.

I'll soon be getting the new thermal paste and shims. I'll be happy to get back the 100 I lost from the smaller pad and with a bit cooler temps for good measure. If the temps drop dramatically, then I may consider reducing the surface area of the pad further and let the heat sink/air/fan do the work of heat dissipation.

Just my 2 cents.

 

[Loog]

I will say even though the thermal Pad only is huge gains in multi core applications about 13% for me, the new thermal paste and copper shim made much more difference in heat. It only gained an added 3-4% in multicore compared to the pad only but the heat reduced is the biggest bonus for new paste and shim, like running GB5 Hovered around 75-80C. With the thermal Pad only, there is a portion of the test I always got very close to 100C. It never broke 80C on Cinebench R20 but with Pad only, it hit 100C. Thats a 20C difference when using intensive apps. WOW

View attachment 926813 View attachment 926814

Glad to see you took the final step @IngerMan and 'shimmed up' well done sir. You've got some great results there and some great write ups on the process!

 

[nill1234]

I will say even though the thermal Pad only is huge gains in multi core applications about 13% for me, the new thermal paste and copper shim made much more difference in heat. It only gained an added 3-4% in multicore compared to the pad only but the heat reduced is the biggest bonus for new paste and shim, like running GB5 Hovered around 75-80C. With the thermal Pad only, there is a portion of the test I always got very close to 100C. It never broke 80C on Cinebench R20 but with Pad only, it hit 100C. Thats a 20C difference when using intensive apps. WOW

View attachment 926813 View attachment 926814

Thats the combination of heatpad and shim. Shim only improves your score but wont lower the temperature much. If you add a pad it has a larger dissipation area over the housing. Overall nice results.

 

[silvia_18]

Plucked up the courage to do the shim mod!

Geekbench
Stock: 3437
1.5mm Pad: 3819
1.5mm Pad & Shim: 4166

Cinebench
Stock: 1148
1.5mm Pad & Shim: 1206

 

[excelsior.ink]

I just want to report, I installed macOS Big Sur Beta on my MBA and Geekbench 5 score dropped from ~3600 to ~3200 for multicore, i5. Don't know why and I will test more, maybe it's other factors than the OS version. Big Sur is beta anyway.

Also, if you want to check Fan Speed and CPU Temperature without installing additional software you can use in terminal something like:

sudo powermetrics -s smc -i 50

where 50 is a value representing miliseconds. The output will be something like:

**** SMC sensors ****


CPU Thermal level: 25
IO Thermal level: 0
Fan: 2693.4 rpm
CPU die temperature: 58.37 C
CPU Plimit: 0.00
GPU Plimit (Int): 0.00
Number of prochots: 0

Maybe you will find this info useful.

 

[usernametakenforever]

I just want to report, I installed macOS Big Sur Beta on my MBA and Geekbench 5 score dropped from ~3600 to ~3200 for multicore, i5. Don't know why and I will test more, maybe it's other factors than the OS version. Big Sur is beta anyway.

Also, if you want to check Fan Speed and CPU Temperature without installing additional software you can use in terminal something like:

sudo powermetrics -s smc -i 50

where 50 is a value representing miliseconds. The output will be something like:

**** SMC sensors ****


CPU Thermal level: 25
IO Thermal level: 0
Fan: 2693.4 rpm
CPU die temperature: 58.37 C
CPU Plimit: 0.00
GPU Plimit (Int): 0.00
Number of prochots: 0

Maybe you will find this info useful.

I think Fanny uses this SMC library in their app. I vaguely remember seeing it in source code.

 

[Mopar]

LTT proved essentially that if the standoffs for the cpu heatsink were shorter to account for the milled out heatsink, the heatsink could have made sufficient contact with the cpu to run much cooler and quieter. Alternatively, apple could have just... not... milled out the heatsink. How is this not a failing of the design or manufacturing? The only 'optimization' you could possibly be talking about is further crippling the cpu

I agree with you to a point, but I think I also need to defend @jackiez here. I believe what he's saying is mostly true. Have a look at the video below and you can see that the CPU is being power throttled – not heat throttled – at about 13W.

From about 13:22 you can see that no matter how much they cool the CPU, performance is static. So it's not a heat issue.

Removing heat does increase performance – up to a point. And then the CPU is power throttled by the OS.

So where I tend to agree with @jackiez is that the cooling system design is not the failure many here are tying to make it out to be. Certainly we now know a heat pipe is not an easy fix – despite all the shrill posts in the early part of this thread claiming otherwise.

Indeed, if there's any design in failure, it is the gap between heatsink and CPU/GPU that is filled with nasty thermal paste that appears to be mostly to blame, and the only reason for it to be like this is to facilitate poor tolerances on the assembly line.

But it is possible those tolerances will get better and eventually the design will work the way it was . . . er, designed for.

We can all agree that shimming the heatsink has realised a significant performance gain in most of the machines here by properly cooling the CPU. And this is in a chassis with no heat pipe. So the airflow design inside the case is not that bad. Some of you guys are seeing even bigger performance gains with heatpads ducting to the case. But – each to his own – it's a bit counter-productive to do that if it means the machine gets too warm on your lap. Fine if it's sitting on a desk, but a few degrees on your lap makes a difference. Again, I am referencing the video above where this is stated. There need to be compromises in terms of comfort and safety.

The flip side is that not every i5/i7 2020 MBA is affected to the same degree. I sit here typing this with seven programs open, almost 40 tabs open in Firefox and a few in Chrome, Skype and Mail running etc, and my CPU is 39.4C and the fan isn't even on. I almost never hear the fan on this machine: 2020 i5/16/256. In fact, for nearly all day-to-day use this machine runs cool and quiet.

So while it's amazing to see what this thread has become – truly legendary – can we at least tone down the rhetoric? Almost everything we buy can be improved, but I don't believe the 2020 MBA is a failure. Because I am almost certain there will be many, many owners out there who do not even know this thread exists . . . yet have absolutely no problems with their machines and are perfectly happy with them.

 

[adrianstuartt]

Indeed, if there's any design in failure, it is the gap between heatsink and CPU/GPU that is filled with nasty thermal paste that appears to be mostly to blame, and the only reason for it to be like this is to facilitate poor tolerances on the assembly line.

This is precisely what we are disagreeing with him on. Everything else we already know and accept. We are well aware we cant turn these into MBPs.

All i wanted was for my $900 laptop to play 1080p60 YT in safari without the fan spinning up to 8k rpms. i dont feel like that is an unreasonable expectation in 2020. the fact that it can now hit TDP wall rather than a thermal wall for most tasks was the bonus for me. On that note, now i can play 1440p60 in Chrome without hearing the fan.

 

[Mopar]

Well, at the risk of repeating myself, it seems to me to be a failure of execution and not design.

Also, looking at the way the 2020 heatsinks have been step-milled, I wonder if the original heatsinks delivered to Apple were not milled that way, and were in fact meant to sit closer to the CPU/GPU cluster with minimal thermal paste – which is effectively what you guys are doing by shimming (replacing copper that was taken out during the milling process). Maybe Apple made the decision late in the day to step-mill the heatsinks for extra clearance because when they tested them on the production line they discovered there wasn't enough tolerance. Prototypes are usually built to finer tolerances than production models.

I'm only guessing, and maybe I'm wrong, but isn't the step in the heatsinks roughly the same thickness as the shims you guys are using? Something to think about.

 

[nill1234]

No the cpu seems a bit higher so the have to mill. Think the problem are tolerances during the manufacturing process. To overcome those problems they are using more thermal paste and a deeper milled out square. Nothing special about it, i dont think its on purpose. Its still a mba with a 10 W cpu, dont know what some people are asking here. It wont achieve thw performance of a 25 or 35 W CPU.

 

[adrianstuartt]

No the cpu seems a bit higher so the have to mill. Think the problem are tolerances during the manufacturing process. To overcome those problems they are using more thermal paste and a deeper milled out square. Nothing special about it, i dont think its on purpose. Its still a mba with a 10 W cpu, dont know what some people are asking here. It wont achieve thw performance of a 25 or 35 W CPU.

It's quite simple. If apple can't design a $900 MBA where 98% of the units wont immediately reach low earth orbit watching 1080p60, using Safari for hardware accelerated h264, taking into consideration manufacturing process tolerances... then it shouldn't have ever even made it into production. It shouldn't be a 50/50 chance of getting an adequate product. I personally believe that they CAN design great products. It is with optimism in their ability to produce great products that i am suggesting this is a design flaw.

 

[agaskew]

Design or execution flaw, couldn't care less. The ball was dropped and the result is some of these run below par. Not the first time this has happened, and it won't be the last.