Hardy does not control the CPU fan properly.

[Mike Hix]

Description: Ubuntu 8.04
Release: 8.04
Clean (no previous configuration or data), full install, 64bit

ThinkPad X61 7675-CTO
BIOS Version: 2.07 (latest as of May 30, 2008)

What I expect to happen:
- The CPU fan should run between 0 and 6000 RPM based on processor temperature, as experienced with previous versions of Ubuntu, and with Windows XP.

What Happens instead:
- When the system's temperature is above spec, the fan spins no faster than ~3200 RPM.
- When the system's temperature is normal or cooler, the fan spins at ~3200 RPM, and very rarely operates at a slower speed.

The fan's maximum speed is 6000 RPM. Sourced from Lenovo's HMM Documentation & verified with lmsensors output as described:

modprobe ibm_acpi experimental=1
echo 0x2F 0x40 > /proc/acpi/ibm/ecdump

This disengages the PWM control. Via lmsensors, one can see that the fan ramps up to about 6000 RPM and stays there. At this fan speed, I can put a heavy load (`yes | sha512sum` is my test) on the system and see that the CPU temperature stabilizes below spec - the heat sink works.

I suspect thinkpad_acpi, but cannot prove it.

ProblemType: Bug
Architecture: amd64
Date: Wed Apr 30 11:19:07 2008
DistroRelease: Ubuntu 8.04
Package: linux-ubuntu-modules-2.6.24-16-generic 2.6.24-16.23
PackageArchitecture: amd64
ProcEnviron:
 PATH=/home/username/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games
 LANG=en_US.UTF-8
 SHELL=/bin/bash
SourcePackage: linux-ubuntu-modules-2.6.24
Uname: Linux 2.6.24-16-generic x86_64

[lophiomys (lophiomys)]

My subjective perception is, that Hardy causes more fan activity than Gutsy.
I have experienced this on 2x Thinkpad T42p 2373-KYG with
Kubuntu 8.04 Hardy Heron LTS installed.
Sorry, I momentarily dont have the time to investigate in depth.

uname -a
Linux T42p 2.6.24-16-generic #1 SMP Thu Apr 10 13:23:42 UTC 2008 i686 GNU/Linux

[Mike Hix (mike-musl)]

It's a different version of the fan hardware (X60 vs. T42p) and a different arch (i686 vs x86_64). Kind of apples / oranges, no?

[lophiomys (lophiomys)]

You might be right.
For my T42p the fan activity has reduced after the recent OS updates for Hardy.
The average idle temp has gone down from 42C to 37C after the OS updates
and the fan activity has reduced to the levels experienced with Gutsy.

[Mike Hix (mike-musl)]

I'm more concerned about the fact that the system will overheat instead of using the full capability of the fan.

[lophiomys (lophiomys)]

My last post was not quite correct.
The T42p system is cool but the fan is blowing cold air at 3600rpm (approx. level 4)
I just didn't hear the fan at the beginning.
Seems to be a similar problem, despite the different hardware.

root@T42p:/# aticonfig --lsp
    core/mem [flags]
-----------------
* 1: 111/111 MHz [low voltage]
  2: 317/203 MHz [default state]
root@T42p:/# cat /proc/acpi/thermal_zone/THM0/temperature
temperature: 36 C
root@T42p:/# cat /proc/acpi/ibm/fan
status: enabled
speed: 3658
level: auto

[Mike Hix (mike-musl)]

I might (be | have been) a bit paranoid.

Based on information from Intel (listed below) I've found that for my particular processor (T7300 2Ghz) that the specs might be a bit more forgiving than I originally thought.

A heat sink compatible with my processor has to be able to dissipate 35W of heat - Intel calls this number "Thermal Design Power". The "Thermal Specification" for the processor is 100 C. Thus, when my CPU is crankin' out 35W of heat, the temperature measured at the joint between the CPU and the heat sink should be no more than 100C.

From Intel.com:

Thermal Design Power:

Thermal Design Power: (Also referred to as Thermal Guideline) The maximum amount of heat which a thermal solution must be able to dissipate from the processor so that the processor will operate under normal operating conditions.

Thermal Specification:

Thermal Specification: The thermal specification shown is the maximum case temperature at the maximum Thermal Design Power (TDP) value for that processor. It is measured at the geometric center on the topside of the processor integrated heat spreader. For processors without integrated heat spreaders such as mobile processors, the thermal specification is referred to as the junction temperature (Tj). The maximum junction temperature is defined by an activation of the processor Intel® Thermal Monitor. The Intel Thermal Monitor’s automatic mode is used to indicate that the maximum TJ has been reached.

[Mike Hix (mike-musl)]

I've tested my system by starting two copies of 'yes | sha1sum' and 'watch acpi -V'.

After 30 minutes, both cores' temperature never broke 86C. The system was stable and responsive during the test.

Anyone else have temperature data from an X61 (tablet or otherwise)? Ubuntu, Windows, or otherwise?

[oscar (garcia-unbc)]

I thought that my ThinkPad X60s was getting warm. I installed the sensors applet, and it showed that the temperature was creeping into the 80's, with the fan steady somewhere around 3300 rpm (or was it 3600?). Seems rather high/dangerous. I guess it is back to Windows until this is fixed, if ever :-(

[Mike Hix (mike-musl)]

Oscar - Does the CPU fan behave differently for you when you're running Windows?

[Gerson "fserve" Barreiros (fserve)]

on my Acer 5315, ubuntu 64 doest not start fan, so i got very HOT and shutdown

i'm using 32bits ubuntu and it's ok now.

there is a way to manual control fan?

[jajo (aliced-tektix)]

i have a thinkpad x61s with the same problem as the bug report starter. fan speed is after booting constant at ~3800. if i stress the laptop it gets really hot!
is there a temporary solution to get the automatic fan control working? when will it be fixed?

[Tor Håkon Haugen (torh)]

I've just done the test Mike Hix wrote about on both a ThinkPad x31 and a ThinkPad x61 (yes, I love the Thinkpads).

On my X61 the fan will be steady somewhere around 3200/3300 rpm. That is even if it's idle or if I'm doing the sha1sum test.

The X31, on the other hand, it's a lot more interesting since the fan is OFF when idle. When I ran the sha1sum test it didn't start before the temperature reached 68 degrees celcius, and then "only" at 3400 rpm. This was however enough to keep the temperature steady at 70 degrees.
When I stopped the test the fan turned itself off when passing 53 degrees celcius. Silence is golden.

So my question is, why does the fan in the X61 keep spinning at ~3300 even when the machine is staying idle?
I must say I prefer the way the X31 handles this problem.

Hope this input is of any value.
Btw, I'm running 32bit Ubuntu on both machines. X61 is running on Hardy and the X31 is using Gutsy. I can upgrade the X31 and do the same test again if necessary. Heck, I can even reinstall it. :-)

[Andrew (andrewkvalheim)]

> Anyone else have temperature data from an X61 (tablet or otherwise)? Ubuntu, Windows, or otherwise?

Here's data from an X60. From what I can tell, everything here looks the way it should.

[Mike Hix (mike-musl)]

It seems that the CPU is intended to operate in the ranges we're observing. It simply seems really hot, not to mention about 90% of spec - not a lot of room for changes in the ambient temperature...

[Mike Hix (mike-musl)]

I guess I just didn't understand the specs.

[[e-mail] (johsuars)]

I have a T42, running Hardy Heron, and my fan doesn't go > ~1900 rpm. When trying to rip a DVD using handbrake (for person backup of course), the computer shuts down stating that the CPU temp has gone too high. I am unable to increase the fan speed using the command to change the setting in /proc/acpi/ibm/fan. It changes the setting, just not the fan speed... I can stop the fan and start it, it just doesn't seem to go fast enough.

Linux chem-smith-nb2 2.6.24-19-generic #1 SMP Wed Jun 18 14:43:41 UTC 2008 i686 GNU/Linux

[kelvie (kelvie)]

~ $ cat /proc/acpi/ibm/thermal
temperatures: 41 38 38 35 23 -128 23 -128 35 34 -128 -128 -128 -128 -128 -128
~ $ cat /proc/acpi/ibm/fan
status: enabled
speed: 3574
level: auto

I get something similar.. all of my temperatures seem OK, and the computer is mostly idle (load was at 0.01 0.06 0.07 when I did this).

The fan doesn't turn off after it starts.

~ $ uname -a
Linux mykons 2.6.27-7-generic #1 SMP Fri Oct 10 03:55:24 UTC 2008 i686 GNU/Linux

on an X61 tablet -- this is on Intrepid Beta.