Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

christophosphorus

macrumors newbie
Original poster
Jun 16, 2020
24
64
Based on the heatsink modification megathread, and on the advice and guidance from users like @DanSilov, @Loog, @Robotronic, @kinchee87, and others I'm forgetting, I performed both the heatsink and the heat pad mods on my 2020 i7 MBA. The results have been absolutely amazing.

Even though all of this information is above, I'm putting together this thread to (hopefully) give other newbies a one-stop writeup. All credit goes to the awesome users from the megathread who are mentioned above!

Why you should consider this: The MBA runs hot; VERY hot. Thermal throttling is a huge issue that's covered in the megathread, and my biggest problem is that videoconferencing or 4K video viewing sends my MBA's temperature through the roof, constantly hitting 100*C and causing the laptop to throttle (thereby reducing performance and causing choppiness). I had resorted to manual control of my fans AND had disabled Turbo Boost as a temporary solution, but the temperatures were still insanely high.

What exactly are we doing? First, We're going to ADD copper shims between the CPU and the heatsink because there's a lot of room between the two; that doesn't allow for the most optimal heat transfer, and so we're helping by slipping in a couple of copper shims between the two and creating a copper sandwich. Second, we're going to use a much better thermal paste than what Apple used. Third, we're going to install a "heat pad" on top of the heatsink, which will further aid with heat dissipation and cooling.

Does it make a difference? Yes, and a huge difference at that. Two points you should consider: (1) my multicore performance increased almost 25% from the mods alone, and (2) my temperatures are much lower and stable, and my fans haven't turned on once since doing this mod three hours ago.

Is it easy to do? Relatively, yes. The megathread has a TON of great advice from DanSilov, Robotronic, Kinchee, and a lot of others. If you have time, I suggest you read through as much of it as you can.

_____________________________________________

Step 0:

Baseline tests were conducted using GeekBench, and temperatures and CPU parameters were logged using Intel Power Gadget. Download both of these.

A good way to test your GeekBench score is to restart your laptop, close any open programs or addons that you generally don't need other than Intel Power Gadget, and let the laptop "rest" for a few minutes so that most of the charts in IPG are stable and not peaking. Then, open GeekBench and click "Run CPU Benchmark" under the CPU test tab.

Once the test is complete, take a screenshot of your score and save the score's URL (you can share this with us later, and you'll compare this with your scores AFTER you perform the mods).

My personal comparison of scores and temperatures is toward the end of this guide.

Step 1:

I purchased these parts from Amazon
based on others' suggestions in the megathread:
IMG_3981.jpg

Step 2:

We need to take apart the MacBook's rear cover
by removing the screws with the "Pentalobe 1.2" screw head. NOTE: The four screws at the top of the rear cover are longer than the rest. Keep this in mind when you're reinstalling the screws when completed.

IMG_3986.jpg


Step 3 (Optional):

2020 MBAs apparently work best with 0.3mm shims. The Amazon product I linked above provides three (3) copper shims that are 15x15x0.3mm. You'll need two of these shims. We'll be sanding down the shims down with 400, 600, 800, 1000, 1200, and 1500 grit sandpaper in that order. To be honest, I'm not sure if this was at all necessary. It definitely removed some impurities from the shims, but I think you'll get 99% of the same performance improvements if you skip this step. I'd also caution you to avoid sanding for a prolonged period, as you don't want to remove too much material.

The unsanded/original shim is on the left; the sanded shims are on the right.

IMG_3984.jpg

Step 4:

You'll need to remove the heatsink cover located near the center of the top half of the laptop (I'm pointing at it in the picture below). Note that there are four rather shallow screws that require a Torx T4 screwdriver head to open. These are NOT the same Pentalobe screws that you removed from the MacBook's rear cover.

IMG_3989.jpg

Step 5:

As you can see, Apple used a ton of gunky/sandy thermal paste to fill the void left by the space between the heatsink and the CPU. First, we need to remove the gunk from both the CPU/chassis and the heatsink that we just removed. The easiest way to do this is by using paper towels and/or q-tips. Once most of the gunk is removed, you can use ArtiClean to remove any of the remaining residue.

IMG_3992.jpg

IMG_3993.jpg

IMG_3994.jpg

Once fully clean, your CPU and heatsink should look like this:

IMG_3995.jpg

IMG_3996.jpg

Step 6:

Now, we're going to apply our new thermal paste directly onto the CPU/GPU on the MacBook itself. Put a small, rice-grain sized clump directly in the middle of each shiny piece shown below, and then spread it around as evenly as possible using a q-tip or an applicator (provided with the MX-4).

NOTE: You need VERY little paste; less is more when it comes to thermal paste.

NOTE: You don't need to cover each and every crevice of this area perfectly.The paste will spread once pressure is applied by reinstalling the heatsink in the next few steps.

Here's what my coverage looked like after applying the MX-4:

IMG_4001.jpg


(continued below)
 
Last edited:
As an Amazon Associate, MacRumors earns a commission from qualifying purchases made through links in this post.
Step 7:

Remember the copper shims we sanded in Step 3? We'll need them now. We'll be using ONE full shim from Step 3, but we'll have to cut the second shim to size since two shims cannot fit in the recessed portion of the heatsink. Take a good pair of scissors and cut one of your shims exactly in half. If the shim ends up getting a bit deformed or isn't completely flat anymore, take the end of pen or pencil and rub the shim while it's laying on your desk. This will help straighten out the shim.

Now, we'll work with one full shim, and one half shim:

IMG_4014.jpg


Step 8:

Per @DanSilov's suggestion, we are now going to paste these copper shims directly to the heatsink. This is a lot easier than installing the shims directly onto the CPU that we already covered with thermal paste in Step 6 above.

So, you want to follow the same routine for applying paste as we did in Step 6: take a small amount and spread it onto one side of both copper shims (the full shim, and the half-size shim we cut, pictured above).

It should look something like this:

IMG_3999.jpg

Step 9:

Once you cover the shims with thermal paste, go ahead and stick these shims directly onto the heatsink that we removed in Step 4. You'll notice that the heatsink has a recessed area where our shims can almost perfectly fit into place. THAT's where you want to stick our shims, paste side down.

I've highlighted the area where our shims will rest (paste DOWN) in red below:

IMG_4015.jpg

Once you install the shims, you'll have a heatsink that looks like this:

IMG_4004.jpg


Step 10:

Now, carefully place the heatsink back into its place over the CPU. Take care to do this rather slowly, as you do NOT want to wobble the heatsink too much once it sits in place. Note the orientation of the heatsink in Step 4 above if you need a reminder.

Note: As other users suggested in the megathread, tighten the screws diagonally and turn by turn so that all of our gooey thermal paste is evenly spread and squeezed down.


You're now done with the heatsink mod! If you're also proceeding with the heat pad mod, read below.


Step 11:

The heat pad mod is relatively simple. We need to cut the Wathai Blue pad (in Step 0) to fit over the top of our heatsink. This can be done with a pair of scissors, as the pad cuts easily. I cut one rather large piece to cover most of the heatsink, and then cut a smaller strip to cover the right side of the heatsink.

I then peeled the red-colored side of the adhesive from the thermal pads and pasted them directly on top of the heatsink. See the pictures below:

Note: I chose to cover the entire heatsink with the thermal pad. Some users left areas uncovered, but I have been experienced no issue with covering the entire thing.

IMG_4007.jpg

Lastly, I removed the plastic from the top of this blue heat pad. If you're having trouble grabbing the plastic covering, use the tweezers that came with our Xoom screwdriver set.


Step 12:

You're done! Reinstall the MacBook's cover using the Pentalobe 1.2 screw head, and make sure to place the LONGER screws toward the top of the MacBook.

IMG_4011.jpg
[automerge]1592730261[/automerge]
GeekBench Scores

These mods were extremely fun, but the performance gains and temperature reductions were even better. Here were my GeekBench scores before doing the mods:

Screen Shot 2020-06-20 at 11.08.20 PM.png

And here they are after:

Screen Shot 2020-06-20 at 11.17.50 PM.png

Along the lines of other users' results:

Single core score: 2.8% increase over stock
Multi core score: 26.7% increase over stock!


Temperatures

GeekBench scores are one thing, but what I really cared about was taming the heat/throttle issues with my MBA. What I've noticed is absolutely amazing: temperatures are running MUCH cooler under load, and are rarely hitting 100*C (where the unmodded MBA will gladly live at 100*C all day during 4K videos, benchmarking tests, videoconferencing (VC), etc.

Here's a side-by-side comparison of Intel Power Gadget's stats and temperatures during GeekBench. The charts on the left side of the screen are pre-mod; the charts on the right are after the heatsink and heat pad mods.

Pre and Post Mod CPU and Temps.png

Take a look at how much lower and steadier the temperatures are. Also, fun fact: the fan NEVER turned on during the GeekBench test after performing these mods. That's absolutely insane, and very impressive.

____________________________________________

Overall, I'm 100% pleased with these two mods, and could not have asked for a better introduction to MacForums and the amazing people here. All of the credit and kudos belongs to the people in the original heatsink modification thread; I merely wanted to put together a separate, linkable thread that will be a little easier for newbies to follow and reference if they decide to perform this mod.

And if you're reading this... DO IT! It has transformed my 2020 MacBook Air from a throttled, handicapped laptop into an absolute beast. This is what we all deserved from the factory.


Cheers!
 
Last edited:
For the i3/Base Models, i'd probably recommend NOT doing the thermalpad mod due to the design of the heatsink. It's a more traditional passive heatsink so i don't think the thermal pad would really make sufficient contact between the heatsink and the chassis cover.

My Geekbench score after installing the copper shim was ~1100/2400. I want to say before the mod, they were ~1000/2100. CBr20 went from ~620 to ~720. The real improvement was simply not hearing the fan anymore.

Somebody started an informal google sheet if people would like to add their before and after scores, etc.

If you would like to participate in keeping better track of Mods and Results, here is an Edit URL (Against my better judgement) to a Google Sheet. If you have any questions please let me know. Please don't do us dirty and put false results in.
 
Last edited:
  • Like
Reactions: ruslan120
My Geekbench score after installing the copper shim was ~1100/2400. I want to say before the mod, they were ~1000/2100. CBr20 went from ~620 to ~720. The real improvement was simply not hearing the fan anymore.

Somebody started an informal google sheet if people would like to add their before and after scores, etc.

That's about a 14% improvement in multicore. Nicely done!

I've added a couple of my GeekBench runs to that spreadsheet. Thanks for linking it.
 
Thanks for the outstanding guide, I have just a question: all the thermal pads I have seen used for this mod are 1,5 mm thick, is the 2 mm a better choice?
Thanks
 
Thanks for the outstanding guide, I have just a question: all the thermal pads I have seen used for this mod are 1,5 mm thick, is the 2 mm a better choice?
Thanks

That's a very good question. I went with the 2mm pad because (1) someone in the megathread had suggested it, and (2) it was the only one in stock via Amazon Prime.

I will say that the laptop does get quite warm to the touch after 15-20 minutes of videoconferencing/4k monitor/7-10 apps running, but this doesn't bother me as I'm using the MBA on my desk.

I do not know if a 1.5mm pad would make any difference as to how warm the MBA's bottom cover gets, as I believe that both sizes will still maintain direct contact with the bottom cover. I think the 2mm just gets squished down a bit more once you reinstall the cover, but hope to hear from others about it.
 
  • Like
Reactions: Ghostrider72
I've got some Artic Silver ArcticClean to be able to do this mod. The instructions say to apply the Thermal Material Remover directly onto the paste but to me it seems crazy to apply a liquid on top of the CPU but is this okay?

Appreciate the help!
 
I've got some Artic Silver ArcticClean to be able to do this mod. The instructions say to apply the Thermal Material Remover directly onto the paste but to me it seems crazy to apply a liquid on top of the CPU but is this okay?

Appreciate the help!

I applied it onto a cotton ball and cleaned the CPU (instead of applying the liquid directly onto the CPU itself). Hope that helps.
 
somewhat interesting update for the i3 model; so initially, i had only added the coppershim, which bumped both my geekbench and cinebenchR20 scores. my thermalpad just arrived and i installed it. the geekbench score didn't really change, it was actually slightly worse but within the margin of error. my cinebench r20 score however improved quite a bit. and bonus, the fan never became audible at all.

cbr20 @ stock: ~620
cbr20 w/ shim: 727
cbr20 w/ shim + pad: 785

[edit/] I should have posted this in the megathread. sorry, mods. i'd have deleted the post if i could :p
 

Attachments

  • Screen Shot 2020-06-23 at 6.02.57 PM.png
    Screen Shot 2020-06-23 at 6.02.57 PM.png
    211.6 KB · Views: 440
Last edited:
  • Like
Reactions: silvia_18
I am seriously considering this mod, but would there be any issues with warranty, specifically, would apple remove the heatsink when replacing the machine or fixing a problem? if so, is there a way to get the same or similar thermal paste that apple uses so if they removed the heatsink, it would look the same?
 
Thanks for posting, parts on order. On the Intel power gadget, which actual attachment are you using? Is there extra steps after installing? I tried this before on a 2020 MBP and got stuck on os booting screen. Thanks again.
 
I just did this and it made a pretty significant difference on my i5 Air. In stock form, idle temp with just Safari opened would be at mid-high 40's C. Now with the full copper shims and heat pad, it's at low-mid 30's C.

I didn't do a Geekbench test since I'm a pretty light user. Only tested with the Intel power gadget's test. With just the copper shims, on the all threads frequency test, it would still reach 95C-98C, although it climbed up there much more slowly than stock. With the heat pad, it now maxes out at 85C. I didn't peel off the foam pad heat shield from the factory. I do wonder if it'll further improve it if I did. This is good enough so I'll just leave it as is.

BTW I just some Arctic Silver that I already have rather than the MX4. Been using Arctic Silver for years with good results I'll stick with what I know.
 
I just did this and it made a pretty significant difference on my i5 Air. In stock form, idle temp with just Safari opened would be at mid-high 40's C. Now with the full copper shims and heat pad, it's at low-mid 30's C.

I didn't do a Geekbench test since I'm a pretty light user. Only tested with the Intel power gadget's test. With just the copper shims, on the all threads frequency test, it would still reach 95C-98C, although it climbed up there much more slowly than stock. With the heat pad, it now maxes out at 85C. I didn't peel off the foam pad heat shield from the factory. I do wonder if it'll further improve it if I did. This is good enough so I'll just leave it as is.

BTW I just some Arctic Silver that I already have rather than the MX4. Been using Arctic Silver for years with good results I'll stick with what I know.

not peeling off the plastic liner might not change the thermal conductivity of the pad much and the results will be similar. BUT the plastic will melt to the heat sink leaving an impossible mess for warranty return.
 
I also didnt remove the plastic liner which is attached to the case and had zero problems with melting. I wanted some insulation to prevent that the case gets too hot.
 
not peeling off the plastic liner might not change the thermal conductivity of the pad much and the results will be similar. BUT the plastic will melt to the heat sink leaving an impossible mess for warranty return.
I peeled a bit of the corner to test how easily it can be removed. It seemed like it’ll leave a mess so I left it alone.

I also didnt remove the plastic liner which is attached to the case and had zero problems with melting. I wanted some insulation to prevent that the case gets too hot.
That was my thought too, to prevent too much heat getting drawn to the bottom and onto my lap.
 
i removed mine; the part where i initially lifted it left adhesive, but once you start it you can then lift off the majority of the adhesive while pulling up the pad. Then i went back and scraped the adhesive where i started it to clean it off completely. with the i3, even with the cpu hitting +90c, the case gets very hot, but not enough to burn. I wouldn't keep it on my lap while running cinebench, though. for the most part everything stays cold. right now, my cpu temp is 36c and the entire case is cold to the touch.
 
I used the Arctic 1.5mm After a good week I took it off to try a Gelid 1.0mm, There was no mess left on the heatsink from the bare thermal pad that sat on there for a week. The Gelid pad seemed more fragile and I only left it on for a day and went back to cutting a new Arctic Pad. Cleaning with isopropyl alcohol on a q-tip or coffee filter will clean the sink like new. Just make sure you don't have excess liquid that will run down on something important.
 
I’d like to contribute another potential optimization which is bumping up the fans to around 20% (or before they’re audible) using TG Pro. It should result in more thermal headroom.

Guide: Stop Burning Your Lap | MacBook
 
Received my 2020 refurbished macbook air i5,16gb, 250GB. Ran GB5 a few times and the results appear to be much lower than many of the premods listed in threat. Cant tell if this normal range or not - what do you think?
1) 1014/1933
2) 1014/1933
3) 1050/2019
4) 1041/1992
 
Received my 2020 refurbished macbook air i5,16gb, 250GB. Ran GB5 a few times and the results appear to be much lower than many of the premods listed in threat. Cant tell if this normal range or not - what do you think?
1) 1014/1933
2) 1014/1933
3) 1050/2019
4) 1041/1992

Yeah, those are lower than my premod i3. Reboot and make sure all programs are closed. Would double check you got an i5 and not an i3.
 
Years ago, I got a heat-sink stand from "id east end", a Japanese outfit, I believe, for a MB. It's quite elegant, made of a solid hunk of brushed aluminum, has several fins on the bottom and is vented in back. The MB lies perfectly flat on it, the bottom in full contact with the stand. I've used it with a number of MB's since then, including my new 2020 MBA, and it reduces the temp considerably. It seems to be no longer available for sale elsewhere, but maybe still in Japan. Here's a site I found it on today, but I don't read Japanese: <https://zigsow.jp/item/152099>. I highly recommend it.
id east end.jpeg
id east end 2.jpg
 
Just be aware that if you sink heat into the chassis with a heat pad (removing the insulator) you are likely to raise the bottom case temp hot spot to above the point at which it will physically burn you if you use the device on your lap.

That's not just speculation here, LTT tested it, and the bottom case hot-spot will get above 46C which will burn you (i.e., there's a reason that insulator is there).

I'm considering replacing the TIM and shimming, but sinking heat into the case is, imho a bad idea (performance be damned).
 
Last edited:
Just be aware that if you sink heat into the chassis with a heat pad (removing the insulator) you are likely to raise the bottom case temp hot spot to above the point at which it will physically burn you if you use the device on your lap.

That's not just speculation here, LTT tested it, and the bottom case hot-spot will get above 46C which will burn you (i.e., there's a reason that insulator is there).

I'm considering replacing the TIM and shimming, but sinking heat into the case is, imho a bad idea (performance be damned).
You're probably right, so warning noted. (I've never removed any insulator or somesuch from anywhere, and don't even know where it'd be or how to do it, so I don't know what that's all about.) But I always use my MB's on a desk (so they're not, strictly speaking, "laptops" :rolleyes:), so I never noticed that danger. I may be mistaken, but I have the feeling that keeping the temp down is healthier for the innards of the machine, and probably reduces thermal throttling, regardless the temp of the bottom case. In any case, my id stand sometimes gets warm when I'm doing something intensive (and the fan very occasionally turns on) but never hot.
 
Last edited:
Just be aware that if you sink heat into the chassis with a heat pad (removing the insulator) you are likely to raise the bottom case temp hot spot to above the point at which it will physically burn you if you use the device on your lap.

That's not just speculation here, LTT tested it, and the bottom case hot-spot will get above 46C which will burn you (i.e., there's a reason that insulator is there).

I'm considering replacing the TIM and shimming, but sinking heat into the case is, imho a bad idea (performance be damned).
I left the factory heat insulator on the bottom lid with the heat pad. I did a quick before and after test with the heat pad with just the Intel power gadget test. With just the 0.3mm shims and Arctic silver 5, it would still climb to 90-95C on the all threads test. After I put on the heat pad, it dropped to 80-85C, with the factory heat shield in place. Geekbench 5 score is 1160 single and 3768 multi, with temps in the low-mid 80's during the test and fan at 2700rpm. This is the i5 model. The bottom is slightly warmer than before but not uncomfortably so.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.