You mentioned earlier that you've changed the .conf file a lot; have you tried detection with the default .conf file? I'm sure the results will still be wrong, but it might be worthwhile seeing what they are anyway.

Don't give up--chances are excellent it can be made to work, but you may need the help of the trouble-shooters at lm_sensors. Just be prepared to furnish a whole lot of data and follow their instructions. And to be fair to them, read past trouble reports that look similar to yours. I spent about two or three hours, probably, looking at their FAQ and old trouble reports before deciding it was time to file one of my own. Then within two or three days with their help my setup was OK.

It's working fine.

You may need to adjust the multiplier values.


The best thing to do, I've found is to bring up Linux and let your system warm up.

Then drop reboot and go to the cmos page that also displays these values.

Note what the temperatures and fan speeds are.

Then go back into Linux and look at what lm_sensors reports.

Using these two values you can adjust lm_sensors to more accurately reflect what is actually being reported...

HOWEVER, that said...

Your VCORE1 is a bit too high. This indicates that your power supply may not be able to put out enough current for the amount of draw that there is on it.

VCORE2 looks fine though, and the ALARM is erroneous. Probably the trip point for it in lm_sensors needs to be adjusted.

3.3v looks fine as it's pretty close to spec. Again Alarm needs to be adjusted.

5V is a little low, something may be amiss here...
-5 and -12 are way off too.

You may want to invest in a beefier power supply if these readings are correct.

no they are not correct but the thing is why I have to manually fix multipliers till it shows something? it's +7% on CPU for 300Mhz FSB and chip is detected correctly I guess and loaded (modules) I made the suggestion thing to modprobe and edited not that much .conf

It's hard to read what you wrote... or rather I find it a bit difficult to comprehend...

But if I am getting what you are saying...

Sensor chips are fairly generic. They have no actual way of determining temperature. Rather they return a value which software, bios or other applications interpret as temperature values.

The same chip may be used in many different motherboards, all of which use differing thermal probes.

So it is effectively not possible to write ONE kernel module which correctly reads all values.

That said, often the values returned to the chips are handled by the bios. In this case the result returned to the OS is more accurate.

I would love to see the lm_sensors modules offer the possibility of reading bios results.

However I understand the rather VAST problem this creates, as different manufacturers choose to do this differently.

The way it is currently done offers the greatest possibility of finding a compatible driver.

Even motherboard monitor, etc. for XP require vendor & motherboard specific configuration files... much like Linux.

The problem for Linux users, is that there are few (yes they are available on the lm_sensors site!) already adjusted configuration files available for the myriad of systems and motherboards which are out there.