You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When playing music/youtube/whatever with the screen off as I go to sleep, the power LED lights up the whole room. Options to dim it were added here, but the ability to disable it entirely would be appreciated. This has also been mentioned in the discussion thread.
Putting electrical tape over a button gets a bit messier than putting it over the charge lights, and I haven't got up the courage to start cutting on the board just yet :)
Exposing it via /sys/class/leds could also be nice (because we could set up keybinds), but the possibility of that is deeper than my firmware knowledge goes.
Reacted by CodeAsm, haarp, James, Dries, Guillaume, Arian Baishya, stylefish, Kaukov and Briella Bugs
EDIT: Not sure if there is any way to restore the default behavior afterwards, and I can't seem to be able to lower the brightness, it is either full or nothing, putting values between 0 and 100 do nothing, if value is >0 then the light looks like it's at full brightness.
EDIT 2: Add Kernel version: Linux 6.11.3 #1-NixOS SMP PREEMPT_DYNAMIC Thu Oct 10 10:04:18 UTC 2024 x86_64 GNU/Linux
Reacted by Mariusz Stępień, Jonathan Plasse, Hunter Horsman, Arian Baishya, Matt Noakes and Briella Bugs
See FrameworkComputer/framework-system#37
We won't add the option to turn off the power button completely, as the LED is used to tell you the power state of the system.
But we are going to add an ultra low power mode and let you set the power button to any percentage you like, except 0%.
Reacted by Nicolaos Skimas, Guillaume, Samy Lahfa, anijatsu and Harry Bond
Reacted by Sheikh, Michael Mior, Emil Fihlman, Jesse Koivukoski, Dawid Potocki, Vishal Sharma, yum-food, Charmunk, Evgeny, zoey-on-github and 14 more
I still feel like there is value in having the option to set it to 0%, the LED has never worked on my FW13 AMD350; and a replacement input cover is on the way (the cover is warped apparently according to support), but I actually had no idea it even was supposed to light up until a friend of mine asked out I disabled it (since his was illuminated).
Once input cover, I am concerned that since I am used to having everything nice and dark when I use my laptop in the dark, I am likely to find the LED annoying.
I think the side LEDs could be used under the same argument of having a visual representation of if the laptop is asleep or off, but honestly it seems pretty irrelevant if you are a poweruser who could manage to use this tooling to set it to 0% anyway. If you are about to interact with the computer and its asleep or off, just pressing a key to wake will tell you if its off really fast when it wakes, and in either scenario pressing the power button would wake/boot the device. So is it really a "it has to be on", since like I said, mine has never worked and it has never really caused me a problem.
TL;DR: What is really the harm in letting powerusers disable it?
Reacted by Robert Ulrich Hofmann, Isaac Parker, percy, Michael Braun, Pat Myron, matt beary, Nathan, Nick Noble, Sean Lang, Jaakko Sirén and 7 more
On any recent Linux system you can manage the LED through the standard APIs.
$ cd /sys/class/leds/chromeos:white:power/ # specific name depends on the device
$ echo 0 > brightness # turn off
$ cp max_brightness > brightness # turn on, unfortunately EC does not support non-maximum brightness
$ echo chromeos-auto > trigger # switch back to automatic management through embedded controller
Reacted by Arian Baishya, Matt Noakes, Michael Nedokushev, Ivan Habunek, no-usernames-left, Charmunk and Briella Bugs
On any recent Linux system you can manage the LED through the standard APIs.
$ cd /sys/class/leds/chromeos:white:power/ # specific name depends on the device
$ echo 0 > brightness # turn off
$ cp max_brightness > brightness # turn on, unfortunately EC does not support non-maximum brightness
$ echo chromeos-auto > trigger # switch back to automatic management through embedded controller
Is it possible to do anything approaching this on Windows? I see that @grouzen has a pretty incredible tool built in Rust which seems to take advantage of these same interfaces, but again it's targeted at a Linux kernel and I see no hints of Windows access to similar APIs without setting up an entire dev environment. Am I correct, or am I missing something wonderful to save a ton of effort for a simple change (mess with my power button brightness on a purely Windows Framework 16)?
Activity
hauntingEcho commentedon Mar 7, 2022
Exposing it via
/sys/class/ledscould also be nice (because we could set up keybinds), but the possibility of that is deeper than my firmware knowledge goes.add 'Off' setting for power LED
add 'Off' setting for power LED
rex: Enable KB8010 retimer driver
james2432 commentedon Sep 2, 2023
I wish this was merged :(
hauntingEcho commentedon Oct 21, 2024
Looks like
/sys/class/ledssupport may be available with kernel 6.11: https://old.reddit.com/r/framework/comments/1g8gbm2/what/lt0fxia/AkechiShiro commentedon Oct 22, 2024
@hauntingEcho can confirm this works on kernel 6.11, I'm on Framework 13th AMD.
I can shutdown the power LED with
Full brightness
Here are also the other leds :
EDIT: Not sure if there is any way to restore the default behavior afterwards, and I can't seem to be able to lower the brightness, it is either full or nothing, putting values between 0 and 100 do nothing, if value is >0 then the light looks like it's at full brightness.
EDIT 2: Add Kernel version: Linux 6.11.3
#1-NixOS SMP PREEMPT_DYNAMIC Thu Oct 10 10:04:18 UTC 2024 x86_64 GNU/LinuxJohnAZoidberg commentedon Mar 27, 2025
See FrameworkComputer/framework-system#37
We won't add the option to turn off the power button completely, as the LED is used to tell you the power state of the system.
But we are going to add an ultra low power mode and let you set the power button to any percentage you like, except 0%.
kariudo commentedon May 11, 2025
I still feel like there is value in having the option to set it to 0%, the LED has never worked on my FW13 AMD350; and a replacement input cover is on the way (the cover is warped apparently according to support), but I actually had no idea it even was supposed to light up until a friend of mine asked out I disabled it (since his was illuminated).
Once input cover, I am concerned that since I am used to having everything nice and dark when I use my laptop in the dark, I am likely to find the LED annoying.
I think the side LEDs could be used under the same argument of having a visual representation of if the laptop is asleep or off, but honestly it seems pretty irrelevant if you are a poweruser who could manage to use this tooling to set it to 0% anyway. If you are about to interact with the computer and its asleep or off, just pressing a key to wake will tell you if its off really fast when it wakes, and in either scenario pressing the power button would wake/boot the device. So is it really a "it has to be on", since like I said, mine has never worked and it has never really caused me a problem.
TL;DR: What is really the harm in letting powerusers disable it?
t-8ch commentedon Jun 28, 2025
On any recent Linux system you can manage the LED through the standard APIs.
queesamor commentedon Dec 26, 2025
Is it possible to do anything approaching this on Windows? I see that @grouzen has a pretty incredible tool built in Rust which seems to take advantage of these same interfaces, but again it's targeted at a Linux kernel and I see no hints of Windows access to similar APIs without setting up an entire dev environment. Am I correct, or am I missing something wonderful to save a ton of effort for a simple change (mess with my power button brightness on a purely Windows Framework 16)?
kiram9 commentedon Dec 30, 2025
@queesamor have you tried framework-tool? https://github.com/FrameworkComputer/framework-system/releases
JohnAZoidberg commentedon Dec 30, 2025
FrameworkComputer/framework-system#37