Ticket #1763 (closed task)
Opened 10 years ago
Fans report zero RPM. Increasing divisor doesn't help. (Examined/Solved - Linux 2.6.7)
|Reported by:||contact||Owned by:||somebody|
The system is an ASUS NCCH-DL dual Nocona board
Chipset: Intel 82875P, Intel 6300ESB ICH
The sensor chip is w83627hf
I get the following fan speed readings:
/sys/bus/i2c/drivers/w83627hf/1-0290 # cat fan1_input
/sys/bus/i2c/drivers/w83627hf/1-0290 # cat fan2_input
/sys/bus/i2c/drivers/w83627hf/1-0290 # cat fan3_input
Temperature readings seems to work:
/sys/bus/i2c/drivers/w83627hf/1-0290 # cat temp1_input
/sys/bus/i2c/drivers/w83627hf/1-0290 # cat temp2_input
/sys/bus/i2c/drivers/w83627hf/1-0290 # cat temp3_input
The BIOS hardware monitor does show valid readings for the fan speed.
Here is the script I used to initialize the drivers:
modprobe w83627hf init=0
I also tried without the init=0 argument and I recieved the same result. I
tried increasing the fan divisors, but that didn't make any difference. Any
suggestions on how to proceed? Let me know if I can provide further information.
--- It looks like the NCCH-DL has 6 fan headers:
PC Health Monitoring support the following:
- 6 x Fan sensor & control: CPU fan*2, System fan*2, Chassis fan*2
- Voltage sensor: Vcore0, Vcore1, +1.5V(vNB), +2.5V(vDDR), +3V, +5V, +12V,
- 3 x Thermal sensor: Environment x 1, CPU x 2
The Winbond W83627HF does support only 3 fan inputs.
So I think that either there is a second hardware monitoring chip (or Super-I/O
with fan monitoring capabilities) on the board, or the fan inputs are multiplexed.
In the first case, the output from sensors-detect may reveal something. If it
doesn't name the chip, it may at least find something and fail to recognize it
(but we humans may succeed). Please send the output of sensors-detect to the
lm-sensors mailing-list (reference your ticket number). Make sure you use a
recent version of the script (2.8.7 or CVS).
In the second, there's almost nothing we can do, since the way the multiplexer
is controled is left to the motherboard manufacturer. Your best chance is to
contact Asus and ask them. But I hope there is a second chip instead, because
multiplexing hardware monitoring inputs is a silly thing to do. (How the hell
are we supposed to monitor a signal if we cannot read it continuously?) It was
seen already though.
After investigations, it happens that the board actually has another hardware
monitoring chip, a Winbond W83792D at I2C addresses 0x2f/0x48/0x49. This is
where the fan speeds must be reported.
We do not support this chip now. No datasheet is available from Winbond at the