Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
M1 Air with Thermal Pad Mod
- Thread starter downshiftdre
- Start date
- Sort by reaction score
Have you read through the thread? The information is there.
Perhaps, that 3% drop is a combination of thermal mod heat and prolonged 100% charged state? If I may ask, how long have you been using your MacBook in this state?
Sorry to be too lazy to search for myself, but: does anybody know where I might find a translation table for these temp readings to a somewhat more understandable name (something like CPU or GPU or Battery)?
My eyes inevitably begin to flutter after the first dozen or so ANE MTR Temp.
My eyes inevitably begin to flutter after the first dozen or so ANE MTR Temp.
The pACC MTR Temp is for the performance cores (one for each core). The eACC MTR temp is efficiency cores (one for each core). GPU MTR Temp is the GPU (apparently just one for all GPU cores). ANE MTR Temp is the Apple Neural Engine. ISP MTR Temp is the Image Signal Processor. NAND CH0 temp is the NAND controller for the SSD. Gas gauge battery might be the battery temp. The rest probably aren’t very important to most uses.
Weird. Looking again at my test code, I'm only seeing 2 sensors for the efficiency cores and 7 sensors for the performance cores. So either my interpretation of what the labels mean is wrong (likely) or something changed with macOS 12.0.1. I'm guessing my code has bugs but what seemed obvious before no longer makes any sense.
This is on an M1 MacBook Air.
While I am afraid to do the mod and I use my MBA 16GB in standard factory configuration this is exactly what I've been experiencing too. Apple on their website states that "safe" temps for battery is around 35C max so seeing 45C on the battery does worry me.
You are absolutely right about rosetta games taxing the CPU. I played ROTR the other day and CPU was at 97C constant while "battery gas gauge" went up to 46C (without any mod, all factory.) I got worried and stopped playing.
However some games like Hollow Knight can be played with much ease as they run much cooler (45C max on the CPU!)
I knew this would happen before purchasing the MBA but I didn't have money to go for MBP and I thought I could resist my urge to play games. Please let me know if you have any suggestion for me.
hi workerbee! bzzzz bzzzzzz bzzzzzzzzzzzzzzzzzzz (sorry couldn't resist)
I had the same questions as of you while using the "hot" app. https://github.com/exelban/stats/issues/210 this helped me a bit.
Slightly delayed reply, apologies. After several overseas trips I am now at a cycle count of 31 and maximum capacity of 94%. That is a MacBook Air bought 11th December 2020. MacOS 12.1 keeps the charge state near 80% as it is usually plugged in. I won't complain at a 6% loss of maximum capacity after 14 month's use, mod or no mod.
iStatistica Sensors actually got an Update not too long ago where they specifically changed the temperature sensors setting, it's now *much* simpler to read and understand.
Attachments
As someone who did the thermal mod early on, I had a question about Apple’s attitude regarding warranty work and trade-ins. Has anyone here had any problems getting (hardware) warranty support? If you traded your MBA in for a higher tier product how did Apple price your MBA or just refuse it?
Just wanted to share my recent experience - after reading all there is I decided to get one of the best pads, the Gelid Ultimate with 15w/mk rating. That's better than the classic Thermalright and doesn't cost that much more. There'r a nice comparison here:
I also decided to go for 2mm height and just cover the central raised portion of the heatsink - completely skipping the second thicker square - seemed to me with the 2mm pressure and high thermal transfer rating it wouldn't bring more benefits if any. Also no scraping of the back insides etc, just one 90x50x2mm pad, cut, paste, close, done.
Everything worked out great I think, I also left some 2mm space between the pad and the central battery divider (most people just shove it in touching all sides), but it might not make any difference, the battery gets probably 2-3 degrees hotter in a long run, but who cares, it's just for a peak performance.
Also from all the thermal camera comparisons it's clear that while the central area over the M1 does get hotter quickly, the rest of the laptop back is gradually cooler than leaving it in a "toaster oven" default scenario - so I guess it sort of evens out
As for performance, I get Cinebench R23 default 10min 7770+ multicore, the first in loop got me 7818 the most so far. Don't think there's much to improve with even more cooling - Created Tech made a interesting water cooling comparison and got less in Cinebench, personally I think it's about the pad quality and "speed" at which it takes away the heat from the chip (but the difference doesn't matter in real life of course):
Everything worked out great I think, I also left some 2mm space between the pad and the central battery divider (most people just shove it in touching all sides), but it might not make any difference, the battery gets probably 2-3 degrees hotter in a long run, but who cares, it's just for a peak performance.
Also from all the thermal camera comparisons it's clear that while the central area over the M1 does get hotter quickly, the rest of the laptop back is gradually cooler than leaving it in a "toaster oven" default scenario - so I guess it sort of evens out
As for performance, I get Cinebench R23 default 10min 7770+ multicore, the first in loop got me 7818 the most so far. Don't think there's much to improve with even more cooling - Created Tech made a interesting water cooling comparison and got less in Cinebench, personally I think it's about the pad quality and "speed" at which it takes away the heat from the chip (but the difference doesn't matter in real life of course):
That’s the same thermal pad I used on my original 2020 Intel MacBook Air with good results before I modded my M1Just wanted to share my recent experience - after reading all there is I decided to get one of the best pads, the Gelid Ultimate with 15w/mk rating. That's better than the classic Thermalright and doesn't cost that much more. There'r a nice comparison here:I also decided to go for 2mm height and just cover the central raised portion of the heatsink - completely skipping the second thicker square - seemed to me with the 2mm pressure and high thermal transfer rating it wouldn't bring more benefits if any. Also no scraping of the back insides etc, just one 90x50x2mm pad, cut, paste, close, done.
Everything worked out great I think, I also left some 2mm space between the pad and the central battery divider (most people just shove it in touching all sides), but it might not make any difference, the battery gets probably 2-3 degrees hotter in a long run, but who cares, it's just for a peak performance.
Also from all the thermal camera comparisons it's clear that while the central area over the M1 does get hotter quickly, the rest of the laptop back is gradually cooler than leaving it in a "toaster oven" default scenario - so I guess it sort of evens out
As for performance, I get Cinebench R23 default 10min 7770+ multicore, the first in loop got me 7818 the most so far. Don't think there's much to improve with even more cooling - Created Tech made a interesting water cooling comparison and got less in Cinebench, personally I think it's about the pad quality and "speed" at which it takes away the heat from the chip (but the difference doesn't matter in real life of course):
Would it make sense for the best internal thermal temperatures (incl. battery) to cover only half of the heasink with thermal pads? So some of the heat stays inside the chassis while the other half goes to the backcover?
My plan is to put some doubled 1,5mm thick 1cm2 on the lower heatsink and some 1,5mm thick 1cm2 on the taller one.
My plan is to put some doubled 1,5mm thick 1cm2 on the lower heatsink and some 1,5mm thick 1cm2 on the taller one.
That may well work well. If you do an image search for the M1 Air thermal sponge you can see the sponge is about the area of a large postage stamp. Try that first and take some readings. There's plenty of benchmark info in previous thread posts so you could compare those with what you're getting.
I was actually thinking that, that's why I skipped the lower part completely. And maybe I could have used less on the upper part, especially with the effective pad I've used, it gets really hot when doing longer exports
So I'd just experiment there.
It is impossible to answer that question without a scientific, randomised, double-blind study which I doubt will ever happen. However, you can take heart and opinion from the vast evidence on this thread that some sort of thermal connection between Soc and case makes a dramatic improvement and a sustained dramatic improvement if the MBA is lifted off the desk surface at an angle to allow the bottom case to give off its heat.
Half and half? I doubt it. Any thermal transmission to the case. aided by a free flow of air, will help. Batteries have been referred to many times. The effect of the mod on the lifespan of the battery is an unknown as yet. Perhaps in a few years time there may be a measurable percentage difference. Perhaps not. However there is an immediate and measurable difference in using the bottom case to disperse heat under sustained load as documented in this thread.
Remember that the thermal conductivity is measured per unit thickness. Don't use 1.5mm or 3mm where 1mm will suffice.
I love my MBA and I love that after the mod it can outperform the MBP released at the same time, on a laptop stand, when needed, silently. It can drive a huge display and remains the 'perfect laptop'. Not even Apple can better it yet, at any price.
Misinformation the Air does not consistently out perform the 13" MBP. The mod can enhance the Air's performance, however in hotter climates it will not be as effective as colder climates and it may result in higher battery temperatures which is not desirable. Passively cooled and all...
Stock base model 13" M1 MBP, ambient temp circa 25/27CC
If you need guarantied sustained performance opt for the M1 13" MBP or the new 14" MBP. If on a budget an Air and the mod will help, just be mindful the Air is passively cooled so don't expect miracles in hotter climates. Only passively notebook/tablet I've observed to be really effective is the Acer Switch 5 with it's liquid cooling loop and that's just the i3 model, i5 & i7 both throttled past the i3's threshold.
Q-6
Thanks for your reply. I should have said my goal: My Air is mostly in light-medium use, almost never on maximum load more than a few seconds. Just want to stay the device in a most healthy condition. Might be excessive, however this mod cost almost nothing, is fun and is free of danger (if you check the temperatures). If i remember correctly you have posted your temperatures a few months back showing that for example the battery is indeed cooler with the mod if the air is not heavily in use. Thats what i am aiming for.
I have ordered the cheapest one from aliexpress i could find:
RGeek 6,0 W/mK sliced 100x100 1,5mm - 2,5 USD
Left part: 2x1,5mm (stacked) = 3mm, right part: 1,5mm
People in this thread used 2mm for the right part, youtube people used 1,5mm. Went for 1,5 so i can use it for both parts.
Very cheap thermal pad + stacked might be much worse than the 30 USD pads some people bought, but i think its more than sufficient for my needs.
Last edited:
With just small bursts of maximum load you might not need the mod at all, but yeah, it's the nature of us "nerds" not to leave any possible performance on the table and we like to tinker with things
Anyway, with your 100x100 you seem to have more than enough to make more than one version, might be informative to try from a smaller strip just on the higher middle part with some space around, run CB, and add more. I only did one 15W/mK on the middle part, it's highly effective (maybe too much) but suspect I could have used even smaller strip.
Not related to this hack method but I need to tell you about different mod I did about better cooling MB AIR M1 7 GPU.
I cleaned the CPU thermal paste and replaced it with Thermal Grizzly Carbonaut - carbon fiber cloth with alleged 64W / (mK). Only thing what is unsure about this stupid idea (not great reviews about this product) is some type of silicon pad at aluminium heatsink. I kept it there. There is height difference in that hetsink but I think CPU portion has good contact with it.
Now I want to do some performace tests with temp readings. If you can help, send me tips in DM. I am mac beginner.
Cinebench 6716 pts
I cleaned the CPU thermal paste and replaced it with Thermal Grizzly Carbonaut - carbon fiber cloth with alleged 64W / (mK). Only thing what is unsure about this stupid idea (not great reviews about this product) is some type of silicon pad at aluminium heatsink. I kept it there. There is height difference in that hetsink but I think CPU portion has good contact with it.
Now I want to do some performace tests with temp readings. If you can help, send me tips in DM. I am mac beginner.
Cinebench 6716 pts
Last edited:
Changing/upgrading the paste might make things better only after the heat is actually having a way to get out I think. That's what the thermal mod does, without it (and maybe even after) I wouldn't expect paste tinkering to change much.
You need to have a thermal pad in contact with the base plate for the mod to work to dissipate the heat. Cinebench 6716 pts is less than what others have seen who have applied the mod. Unless your ambient temperature is high as that greatly impact. M1 SOC can push close to 8K with adequate cooling, but for that you need a fan or live in Antarctica.
Q-6