I’ve been experiencing an occasional issue where the CPU fan doesn’t spin up until I reboot. When I run the sensors command, the cros_ec-isa-0000 section lists cpu@4c as FAULT and the fan1 speed as 0 RPM.
Rebooting always fixes the issue. I’m not sure how frequently this happens because the fan is usually pretty quiet. I haven’t contacted support because this feels like a Firmware bug.
Distro: Arch Linux w/ CachyOS Kernel (although I remember seeing this before on stock Arch Linux).
Kernel: 6.11 (but not new to this release).
Laptop: Framework AMD 13.
That’s strange. That is something I have never experienced in my Framework AMD 13. Perhaps BIOS update or looking for kernel update. Could also be a hardware problem but mostly those problems have software solutions, as a first step.
I am on stock Arch Linux (but LTS kernel) and I’ve experienced this multiple times. I’ll be in a video conference for work and notice that my temperature reading in my status bar shows 99℃ and my fan isn’t spinning at all. Just like you, I’d reboot and then suddenly my fans would roar on at their highest speed because the laptop was literally hot enough to boil water.
Distro: Arch Linux
Kernel: 6.6 LTS (Can’t go more specific because I don’t remember which point release I was on the last time it happened. It was within the past month.)
Laptop: Framework AMD 13.
Yeah it’s happened to me once too this week, thought I was going crazy. A reboot fixed it - I should have tried using ec-tool to see if manually setting the fan controller to 100% then auto would fix it but next time.
If you run a load do you ever see a RPM reported? One person had erratic fan behavior and one of the fan connector pins was bent.
You could also check ectool to see what the temperature/fan settings are. Maybe something is out of wack, a sensor is reporting something weird or a sensor stops reporting.
Happened to me again today, this time I used ectool and saw the same behavior described in the first post and my status bar listed my temperature as 180℃, and it would have burned me if I touched the bottom of my laptop for more than a split second. Dear framework, please fix your buggy firmware, I don’t think my laptop appreciates being baked from the inside out.
ectool pwmgetfanrpm:
Fan 0 RPM: 0
ectool temps all:
--sensor name -------- temperature -------- ratio (fan_off and fan_max) --
local_f75303@4d 307 K (= 34 C) 0% (313 K and 343 K)
cpu_f75303@4d 305 K (= 32 C) 0% (319 K and 327 K)
ddr_f75303@4d 306 K (= 33 C) N/A (fan_off=401 K, fan_max=401 K)
Sensor 3 error
That “Sensor 3 error” is referring to “cpu@4c” which I can confirm by running the same command after a reboot.
Distro: Debian sid
Kernel: 6.11.6-amd64
Laptop: Framework 13 AMD Ryzen 7 7840U (new mainboard upgraded from an 11th gen Intel)
Bios: 03.05
Near the end of a video call just now I noticed my laptop was extremely hot to the touch. sensors reported cpu@4c: FAULT, and ectool temps all reported Sensor 3 error.
I was able to turn on the fan with ectool fanduty 100. The fan immediately turned off upon ectool autofanctrl. Even after the temperature is back down, sensors and ectool temps all still report the fault status. I don’t see anything obvious in dmesg.
Suspending then immediately awakening seems to have reset it for me, after doing that sensors and ectool temps all report normal-seeming values.
ectool temps all
--sensor name -------- temperature -------- ratio (fan_off and fan_max) --
local_f75303@4d 307 K (= 34 C) 0% (313 K and 343 K)
cpu_f75303@4d 309 K (= 36 C) 0% (319 K and 327 K)
ddr_f75303@4d 307 K (= 34 C) N/A (fan_off=401 K, fan_max=401 K)
cpu@4c 325 K (= 52 C) 0% (376 K and 378 K)
If the EC is detecting a problem reading a cpu@4c sensor, then one would need to understand which bits of hardware and software are responsible for that sensor value. We don’t have detailed enough schematics to answer this.
If the EC is responsible for the sensor and the fans then maybe it is a bug on the EC side or a near faulty sensor.
This is a little perplexing because rebooting the laptop does not necassarily reboot the EC so it is odd how rebooting fixed this.
What reboot procedure did people use? FYI, the EC is only reset if one powers off, removes the psu and wait 40 seconds before powering on again.
This does maybe highlight a failure mode that is not covered correctly by the EC. I would suggest that if one cannot read the temp sensor, fans should more safely be set at max, not off.
Same issue here, but it happened on my Windows partition. Noticed because I was compiling a project and curiously I had no fan noise. Checked hardware info and saw my CPU pinned at 100C.