Salut Philippe, Merci de r?pondre ? la liste plut?t qu'? moi, d'autres personnes auront peut etre d'autres id?es. > The thermal sensors is good because the CPU fan is well regulate by > the CPU temperature. If you mean that you have a self-regulating CPU fan, then your analysis is not correct, because such fans regulate themselves using an internal thermal sensor. This is not the same sensor as the ones lm_sensors is reporting values for. > w83l785ts-i2c-1-2e > Adapter: SMBus nForce2 adapter at 5500 > temp: +61?C (high = +110?C) *This* is the CPU temperature, as obtained from a thermal diode built inside the CPU. 61 degrees C is a quite reasonable value, so I'd expect it to be correct. > M/B Temp: +48?C (high = +45?C, hyst = +40?C) And this should be the temperature in the CPU socket, most likely from a thermistor. The value is expected to be lower, and to change much slower, that those returned by the built-in thermal diode. This seems to be your case so there's nothing to worry about. Now, if the value really *never* changes, there might be a problem. You might try increasing the high and hyst limits for this input (to, say, 55 and 50 degrees C, respectively). I know that some sensor chips stop monitoring channels in alarm conditions until the alarm is read. I also have to admit that 48 degrees C is a bit high for a socket temperature. Remember that the asb100 driver was written without the benefit of a datasheet. Blame Asus for that. As a result, we did our best but when things don't seem to work properly, there's almost nothing we can do. Maybe Mark Hoffman will have more ideas, as he wrote the driver. > Power Temp: > -0?C (high = +45?C, hyst = +40?C) > CPU Temp (AMD): > +25?C (high = +60?C, hyst = +50?C) These two are obviously not wired so you can add "ignore" statements for them in your sensors.conf file. -- Jean Delvare
