crwdns2933423:0crwdne2933423:0

Repair guides and disassembly information for the 16-inch MacBook Pro released in October of 2021, featuring Apple-designed M1 Pro and M1 Max SoCs. Model A2485.

Power management chip getting overheated need solution

Hi everyone

I have m1 max 16 inch. I was running beta dev version of mac os tahoe from last couple of months but this overheating of this bottom chip started occuring from last 1 month. I kinda figured because apart for these 4 the circled one is extremely hot to touch that its impossible to touch with finger even for 1 second.

If i don't connect battery cable i think then it doesn’t get that.

I fixed last month in august after reinstalling Tahoe developer beta and removing some apps now recently they released stable and it started happening again.

If someone has any ideas would really be helpful its very painful to start migrating all apps and data its huge and time consuming to fix this for long term.

Block Image
Block Image
Block Image

Update (09/20/25)

Sorry i think its wifi module. Marking on chip and gpt says: That chip is a Murata Wi-Fi/Bluetooth module, specifically Murata 339S00761 / BCM4377 family (the markings like K1CM227 CB6692 TENA12230 are Murata’s lot codes).

crwdns2934081:0crwdne2934081:0 crwdns2934083:0crwdne2934083:0 crwdns2934093:0crwdne2934093:0

crwdns2934109:0crwdne2934109:0

crwdns2889612:0crwdne2889612:0 0
crwdns2934285:0crwdne2934285:0

crwdns2933313:01crwdne2933313:0

crwdns2934051:0crwdne2934051:0

The chip you have marked is one of four identical chips which your system has these are the Flash (NAND) chips which is your data storage (often referred to as a SSD). In an M Series system Apple splits the parts that make up a typical SSD which has a controller and the Flash chips. The controllers is baked in the M1 chip so here we only have the Raw Flash chips.

So let's dig into the cause of the heating a bit.

SSD's and other Flash based storage have an inherent weakness which is a limited amount of write cycles.

Let's look at a model so you can see this... you have a piece of paper which has four squares drawn and you fill in each square with a letter in alphabetical order A, B, C & D ... Now we need to get rid of some of this data as we need to plug in newer data! Erase the letters B & D now putting E & F in their place and we repeat that a few more times letting A & C alone. The paper is getting worn where we are constantly rewriting! If we tried to get to the last letters of the alphabet I doubt the boxes would be readable clearly! Here is where things get interesting! Before the paper is too badly worn the controller does what we call wear leveling! So now the not modified boxes are now modified but we just can't erase the data we need to move it so this time we make the B box A and then erase the original A box and likewise we do the same with the C box writing it in the D box and erasing the original C! Now the empty boxes can then be filled in continuing our Alphabet run! At some point this happens again on other more static cells.

So a very full storage will need to do this dance as there is not enough empty boxes as you just updated a sizable and more static bit of code the system needed to move things about as well. So it's not uncommon for the Flash chips getting warm.

Now the killer! As the system is used some of the storage cells wear out as we encountered in our model the paper gets worn and dirty so the box is less readable. Your usage over time could have worn quite a few cells! This is one of the hazards of a lean system with not enough storage given the usage the root cause of the heat!

You may need to unload your system of most of your created data to an external drive so the system can have more space to do what it needs with its wear leveling as the more free space it has it can move things once Vs multiple times which is what I think you are seeing is the avalanche effect of a lean system trying to wear level!

So quickly copy off and then delete as much as you can of your music and vids you've download and anything else you can like old work to a second drive. Hopefully you are backing up as well. On a 1TB system I try to get better than 256GB free before I upgrade see if you can get to that here. And don't let the system run with less than 128GB of free space.

crwdns2934105:0crwdne2934105:0

crwdns2889612:0crwdne2889612:0 1

crwdns2944067:010crwdne2944067:0:

Thank you that makes a lot of sense actually i did clean up my space from 900GB to 450GB in august when i cleaned up and did reinstall tahoe but again it went upto 950GB 2-3 days back.

I can clean up to keep it under 800GB

So those 4 silver chips aren’t power or wifi module as told to me by gpt or serial number online?

I initially thought to remove the one silver chips that was getting way too hot to get few more months buffer because i dont want to buy m4 max rather wait for m5 but keep this running until next 2-3 months.

crwdns2934271:0crwdnd2934271:0crwdne2934271:0

@danrowns - Please don't remove any chips! As once you do the phone won't work any more. Not sure how that got this so confused they are Flash chips! Look though the Teardowns you'll see they are the same genera p,ace on the older Intel models which this system used as its start. The 12 & 13" have only two chips.

The way these chips work is a RAID type of way! Here the given write of data is split across the chips so using the example as a frame work... each block is on a different chip so A is on the first B on the second and onwards to if you remove one chip you loose the integrity of the file.

So you now have a plan! But 800GB is still on the high side. What do you have on here that takes up so much space?

crwdns2934271:0crwdnd2934271:0crwdne2934271:0

You mean mac right?

Its my primary development machine so lots of stuff. I can move around more to bring to 600 something just these days AI models etc takes lot of space.

Your suggestions and knowledge was very helpful its amazing to see how things go on hardware side for data transfer. Very grateful for it.

crwdns2934271:0crwdnd2934271:0crwdne2934271:0

@danrowns - Yes, freudian slip was checking out the new iPhones!

Local storage for AI will be a big issue as it progresses. Are you mostly desk bound? Maybe you'll need a desktop level system if you are getting deep into AI or any serious coding. You can compile code on it while using your MacBook system. I like the Mac Mini Pro! While the RAM is soldered the flash is a module so you can boost it to 4TB via third party options.

crwdns2934271:0crwdnd2934271:0crwdne2934271:0

@danj yes i do train llm as well so dataset and combination of testing and running on cloud cluster just adds up. when i bought this in 2021 it was more than enough now 1TB feels less. Training sets, simple ai models just takes up easily lots of space.

I am thinking to get nvidia dgx spark to offload local task and wait 2-3 more months and get m5 max for coding/ai. I am hoping they will release by jan/feb if lucky may be november.

Last monthly only i tried gpt oss ai model which is still around 20-25GB i believe.

Then there are other image generation ones stable diffusion xl :) which are fun to use.

crwdns2934271:0crwdnd2934271:0crwdne2934271:0

crwdns2934275:05crwdne2934275:0

crwdns2934285:0crwdne2934285:0

crwdns2934229:0crwdne2934229:0

Dannrown crwdns2934231:0crwdne2934231:0
crwdns2936625:0crwdne2936625:0:

crwdns2936751:024crwdne2936751:0 0

crwdns2936753:07crwdne2936753:0 5

crwdns2936753:030crwdne2936753:0 94

crwdns2942667:0crwdne2942667:0 94