Hi Juerg,
> Hi Jeff,
>
> Thanks for the detailed feedback.
>
>
>> Experiences so far:
>>
>> (1) "fancontrol" script does work, but I would like it to run as a
>> "background" process or even as a daemon. I'm not a programmer so I
>> don't
>> know if this is even possible or how to do it.
>>
>> (2) Fan speed on "Fan1" does go up and down with system temperatures.
>
> I don't understand. Does fan1 change as a result of your running the
> fancontrol script? The SCH5127 has a built in fan controller so you
> shouldn't need to run any external scripts. The chip should
> automatically take care of adjusting the fan speed based on the temp.
> Can you please send the output of the following command:
>
> grep "" /sys/class/hwmon/hwmon*/*
>
> Thanks
> ...Juerg
>
No output from your command in my SSH session, but there is symbolically
linked 'content' in that file structure.
[root@anas-01 ~]# tree -af /sys/class/hwmon
/sys/class/hwmon
/sys/class/hwmon/hwmon0 -> ../../devices/platform/coretemp.0/hwmon/hwmon0
/sys/class/hwmon/hwmon1 -> ../../devices/platform/dme1737.2160/hwmon/hwmon1
These directories have all sorts of files in both of them.
[root@anas-01 ~]# ls /sys/devices/platform/coretemp.0
driver hwmon modalias name power subsystem temp1_crit
temp1_crit_alarm temp1_input temp1_label uevent
[root@anas-01 ~]# ls /sys/devices/platform/dme1737.2160
driver fan3_alarm in1_alarm in3_input in5_max pwm1
pwm2_auto_point1_pwm pwm3_enable temp2_alarm temp3_min
zone2_auto_point3_temp
fan1_alarm fan3_input in1_input in3_max in5_min
pwm1_auto_channels_zone pwm2_auto_point2_pwm pwm3_freq
temp2_fault uevent
fan1_input fan3_min in1_max in3_min in6_alarm
pwm1_auto_point1_pwm pwm2_enable pwm3_ramp_rate
temp2_input zone1_auto_channels_temp
fan1_min fan3_type in1_min in4_alarm in6_input
pwm1_auto_point2_pwm pwm2_freq subsystem
temp2_max zone1_auto_point1_temp
fan1_type hwmon in2_alarm in4_input in6_max pwm1_enable
pwm2_ramp_rate temp1_alarm temp2_min
zone1_auto_point2_temp
fan2_alarm in0_alarm in2_input in4_max in6_min pwm1_freq
pwm3 temp1_fault temp3_alarm
zone1_auto_point3_temp
fan2_input in0_input in2_max in4_min modalias pwm1_ramp_rate
pwm3_auto_channels_zone temp1_input temp3_fault
zone2_auto_channels_temp
fan2_min in0_max in2_min in5_alarm name pwm2
pwm3_auto_point1_pwm temp1_max temp3_input
zone2_auto_point1_temp
fan2_type in0_min in3_alarm in5_input power
pwm2_auto_channels_zone pwm3_auto_point2_pwm temp1_min
temp3_max zone2_auto_point2_temp
Hopefully the formatting doesn't look ugly. Only PWM1 (Fan1...4-pin) and
PWM2 (Fan2...4-pin) connect to motherboard hardware. PWM3 does not have a
motherboard connector.
I did not know that SCH5127 had an internal fan controller, but I
suspected something going on when I changed the system fan from a 3-pin
format to a 4-pin format by reconnecting the 'blue' PWM wire on the
header. With the PWM wire added back I saw my Fan1 speed drop to ~750 rpm;
full speed is ~1900 rpm. I run a tool called "systemgraph" (google
it...nice application and "open source" too) to display a webpage of
various values it collects via Perl scripts and deposits into RRD tables
for a nice "history" view.
In the 4-pin format and no other adjustments, sensors reported the Fan1
speed around 750 rpm and HDD temperatures are ~40 C. With the 3-pin
format, Fan1 speed is ~1900 rpm and HDD temps are ~34 C. So in 4-wire Fan
mode I tinkered with "pwmconfig" and "fancontrol" for a while to
understand how the "/etc/fancontrol" config file works. I finally settled
on this:
[root@anas-01 ~]# cat /etc/fancontrol
# Configuration file generated by pwmconfig, changes will be lost
INTERVAL=10
DEVPATH=hwmon1=devices/platform/dme1737.2160
DEVNAME=hwmon1=sch5127
FCTEMPS= hwmon1/device/pwm1=hwmon1/device/temp1_input
FCFANS= hwmon1/device/pwm1=hwmon1/device/fan1_input
MINTEMP= hwmon1/device/pwm1=30
MAXTEMP= hwmon1/device/pwm1=32
MINSTART= hwmon1/device/pwm1=255
MINSTOP= hwmon1/device/pwm1=254
MINPWM= hwmon1/device/pwm1=254
Then I edited "/etc/rc.local" to add this line:
/usr/sbin/fancontrol &
Fancontrol now runs in the background all the time unless I "kill it". I
don't have a 4-pin fan with a "loose" PWM wire inside the chassis. I can
tape and secure the PWM wire, but then I have to disconnect stuff and open
the case when I wish to experiment. I can restore full 4-pin operation by
removing the line added to "/etc/rc.local" and then echoing the desired
value to "pwm1_enable" or rebooting the system. Some might call my method
the "lazy sysadmin" approach to system thermal control. Works for me!
This setup does a nice job of keeping the 4-pin fan running at it's
maximum speed of ~1900 rpm and HDD temperatures stay below 40 Celsius even
under load.
_______________________________________________
lm-sensors mailing list
lm-sensors@xxxxxxxxxxxxxx
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
