Some newer sensor chips support "smart fan control" which varies the speed of the fan automatically based on the temperature reading. This happens without software intervention. I wouldn't be surprised if a laptop had that type of chip. We do NOT support that operation in ANY of our drivers. So don't expect it to happen even when you do get the right driver. Unless the BIOS configures a chip for smart control and our driver doesn't override it, you will not get fan speed variation using our drivers, unless you write an application your self to control things via /proc, or just set the speed manually by writing a speed to /proc. Once you have positively identified your sensor chip we may be able to help you further. mds phil at netroedge.com wrote: > > If the chips are power cycling, then it's probably going to be > difficult to fix that somehow, unless the Bios can let you override > that. A more realistic solution is to use a 'wake up' script which > resets your limits for you each time the computer is woken up (taken > off suspend). > > Phil > > On Fri, Jan 04, 2002 at 08:38:20AM +0100, Marcus Eger wrote: > > I know, it's a 440BX board, but which sensoring chips...!? > > I'll have a look the next time I open the case. > > BTW, the bios does not seem to work well on its own. On win98 > > (and also linux without sensors) the cpu fan does not stop once ever > > started. However, it starts correctly when the temperature seems to be high > > enough. > > > > Is there a possibilty to override the bios? (reset alarms?!) Or to control > > the fan directly -- I ask because the noise generated by the fan is terrible. > > > > To Phil: > > Yes, I get the effect you describe. Is there a possibility to suppress > > this power cycling of the chips? I don't have a clue. > > > > Marcus > > > > > > On Saturday 29 December 2001 00:26, Mark D. Studebaker wrote: > > > You may not really have a LM80 or MAX1617, > > > those chips are often misdetected. > > > I'd tell you to look at your board but you said > > > it's a laptop so that's harder... > > > mds > > > > > > phil at netroedge.com wrote: > > > > Boy, that is odd. It's like the Bios or something else is trying to > > > > set the limits as well. Or perhaps the chip is power-cycled during > > > > the event, which causes it to reset to defaults? Or, another idea is > > > > that your distibution has a daemon or something running which might be > > > > setting values? I'm at a loss. :'( > > > > > > > > At the end you note that a sensors -s only partially fixes the > > > > situation. What if you remove the module, re-insmod it, and then do a > > > > 'sensors -s'. Does it return to the state that you show after the > > > > first 'sensors -s'? If so, then it's likely that the chip is > > > > power-cycled during a suspend which loses it's config/limits settings. > > > > Removing the driver and reinstalling it causes it to be reinitialized. > > > > > > > > Phil > > > >
