[lm-sensors] ATK0110

Mark Altern mark.altern at gmail.com
Fri Oct 2 21:23:54 CEST 2009


On Fri, Oct 2, 2009 at 12:44 PM, Luca Tettamanti <kronos.it at gmail.com>wrote:

> (do not strip the CC)
>
> On Fri, Oct 2, 2009 at 7:16 PM, Mark Altern <mark.altern at gmail.com> wrote:
> > No,  I am NOT using any fancontrol or pwmconfig, the coretemp reported
> > around 37-40, which is consistant as in windows,  so the
> > CPU diode temp should definitely be lower than that. The w83667hg
> reported
> > the same diode CPU temp (36C) as I observed by several softwares under
> > windows 7, where I am sure it is correct. So w83667 indeed reported a lot
> of
> > other wrong values, like the voltages, but I believe it reports the
> correct
> > CPU diode temp.
> >
> > Thus,  atk0110 is working, but it report wrong value and push the fan
> > speeding up, which I confirmed by opening the case and touch the CPU heat
> > sink, and it is stone cold, but the fan is running like crazy, more than
> > 1000rpm, where it should only run at about 500rpm, as it does in Windows.
>
> I asked about fancontrol because atk0110 itself does not control the
> fan, but an external monitoring program may react to the higher
> temperature.
> Asus BIOS defines a "Q-FAN" mode, which - when set to auto - allows
> the system to adjust fan rotation based on CPU temperature. atk0110
> driver does not change the settings defined in the BIOS screen (I
> haven't decoded the interface yet!).
>
> > The funny thing is that right after computer startsup, the lm_sensors
> report
> > correct diode CPU temp, like around 36C, and fan is spinning at around
> > 600rpm, but after idling a minutes or two, that value suddenly jump up to
> > 47C and you hear the fan speeding up to 1100rmp, while the coretemps are
> > still the same as before, further more, I touched the "cold" CPU heat
> sink,
> > 47C is definitely NOT true.  So how does the atk0110 work?
>
> It just report the value calculated by the firmware, the driver itself
> does not perform any "adjustment" on the temperature.
>
> > Why this wrong reporting only happens after a minites of startup, instead
> of from the
> > beginning?
>
> Because the monitoring stuff embedded on the board decided so ;-)
> Really, the driver is just passing up the data. Please send me a copy
> of the DSDT.
>
> L
>
I will send you the DSDT after I come home from work. I have Q-fan in BIOS
enabled.

By the way, how does w83667 report/calculate the CPU temperature?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lm-sensors.org/pipermail/lm-sensors/attachments/20091002/df1db1f0/attachment.html>


More information about the lm-sensors mailing list