CPU fans race when loading it87 module for lm-sensors. Why?

gcostanz

I'm trying to use lm-sensors. I've compiled I2c modules and lm-sensors and all went well. Sensors detect reports:

weltraumfahrer · Posted: Sun May 04, 2003 11:09 am Post subject:

Hi,
> I'm using a shuttle with an 845GE intel chipset.
me too, with the same problem:-(
This is a known problem of the module (found somewhere in the www).

For me, i've replaced the original fan with a papst fan, who runs always at full speed (1500 RPM).

Frank

gcostanz · Posted: Sun May 04, 2003 12:58 pm Post subject:

Hey thanks. This is a big help and time saver. I don't thinkt that I'll spend any time on this right now. I'll keep on the look-out for an updated module.

Cheers!

BlackB1rd · Posted: Tue Nov 04, 2003 7:43 pm Post subject:

Just curious if there is already a solution for this one. I am experiencing the same increase of fan speed after modprobing this module (it87).

[edit] Took some time to search the lm_sensors site and found this bug report: http://www2.lm-sensors.nu/~lm78/readticket.cgi?ticket=1204
Editing it87.c might be a solution, though haven't tested it out yet
[/edit]

lazarusrat · Posted: Wed Nov 12, 2003 4:53 pm Post subject:

BlackB1rd,
I did the edit suggested in the link you gave, and it did fix the problem for me (Shuttle SB51G).
_________________
obpiper: pipe menu generator for openbox
obtheme: pipe menu to switch openbox themes

rob.sharp

For the sake of anyone searching on this problem, the "patch" also fixed the problem on my Shuttle SN 41G2.

Bear in mind that doing this relies on the BIOS to control the temperature, so make sure you have the appropriate settings.

torklingberg · Tux's lil' helper Joined: 30 May 2004 Posts: 86

I had the same problem with an Asus K8V motherboard, with the CPU fan controlled by the board's Q-fan system. The fan went to full speed wheneither i loaded the w83627hf or w83781d modules.

I solved it by adding the init=0 option when loading w83627hf, like

serendipity · n00b Joined: 05 Jun 2004 Posts: 69

I have not had this problem on my Shuttle SB75G2. I have, however, noticed that the fan control through pwm settings does not work as expected, so this type of problem is not surprising. I found that there is no correlation between fan rpm and pwm value over the full fan range : full pwm range.

On my FB75 mainboard. the fan is at 0 rpm at a pwm of 4, at one notch less than full at 30 (over the continuous range 4 - 30), and max rpm is set at a pwm value of 101. Odd, but I write a script to test it, and this behaviour is totally consistent.

I use a hacked version of fancontrol from the lm_sensors userspace tools, and I find that it does an excellent job of keeping the CPU and chipset at the right temperature while keeping my noise level as low as possible.

Oh, and it does a lot better job than the BIOS.