'sensors' never installed

'sensors' never installed

[Brian Papantonio]

I installed i2c-2.7.0 and lm_sensors-2.7.0, and
"values" are reported in
/proc/sys/dev/sensors/as99127f-i2c-0-2d/ "properly",
but they are way off:

]# cat < temp1 temp1 temp2 temp3
60.0 50.0 33.0
60.0 50.0 27.0
60.0 50.0 112.2

those are in deg C, right? temp1 seems to be ok
assuming it is the mobo. I dunno what the hell temp2/3
are supposed to be.

I try editing /etc/sensors.conf, but no changes to
compute temp2/temp2 have any effect. i try to do
"sensors -s" but it says "command not found"--it never
got installed (I cant find it anywhere on my system).
Something tells me that getting 'sensors' installed
will be a bitch, so is there a way to "manually"
update the compute line in libsensors somewhere? it
must be stored somewhere, because I restart the
computer, and the values for temp1/2/3 never change,
so the sensors.conf file must not be being reloaded
(it is set to startup, which works).

Thanks,
Brian


==========
Asus A7V-266-E
Linux 2.4.20-8+RH_9.0

[root@x]# lsmod
Module                  Size  Used by    Not tainted
via686a                 9696   0  (unused)
w83781d                22608   1 
i2c-proc                9040   1  [via686a w83781d]
i2c-isa                 1832   0  (unused)
i2c-viapro              4912   0  (unused)
i2c-core               19012   0  [via686a w83781d
i2c-proc i2c-isa i2c-viapro]




__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com

'sensors' never installed

[Brian Papantonio]

because...

[root@localhost as99127f-i2c-0-2d]# cat < temp1 temp2
temp3
60.0 50.0 22.0
60.0 50.0 112.2

it "eats" the output for the first value! don't ask me
why, I am a linux noob. So I just wrote 'temp1' twice
to be used as a dummy value. it means nothing.

Anyway...

I agree that temp1 is the motherboard temp in C. It
agrees nicely with what the hardware monitor in my
BIOS setup program displays. It also agree with the
fact that temp1 computation doesn't exist in the
sensors.conf file which has proven to be useless
anyway.

For the processor, the BIOS program reports 58C/136F
with case and HeatSink/Fan set to low. The minimum
i've seen in the BIOS is 46C/114.5F with side door
removed and proc/case fans set to highest. temp3 MIGHT
have been close as F, EXCEPT for the fact that I just
noticed it NEVER moves. I don't have any idea where it
comes from. Temp2 is active, and SEEMS to be the
processor, however, the original output of "26.0" or
"22.0" is obviously miscalculated.

I think using that information, at the very least I
should be able to interpolate the final values, but
depending on the computation used to obtain "22/26",
it may not be very accurate since when dealing with
temperature, the value/starting point of 0 you would
use when comparing two sets of values is meaningless,
especially since I don't know if "22/26" is C, F, or
neither.

I am going to use this value in a perl script I
downloaded that displays the temp and fan stats on an
LCD module mounted to the front of the computer (fun
picture: http://12.203.238.243/lcd.jpg ). Since it is
a perl script, I can edit the appropriate lines to to
a "post-convert" on the temp2 value, as the program
does not output the value of temp2 directly. However,
the temp2 value must be the result of some calculation
that is completely unknown to me. Do you think you
might be able to figure out what the raw value should
be, so we can take it from there? Or is "22.0" the raw
value? (See 'useless info' section for specifications
on my computer, which should give you an idea of what
the value might be)

For your enjoyment <g> I have included a quasi-random
sample of outputs from temp1/temp2/temp3 (quasi
because whenever there was not change yet I didn't
copy it again, and these are not even close to even
intervals):

WARNING: yes this is overkill!! feel free to skim/skip
it. The executive summary is that when I pumped up the
processor's HS/F, temp2 went down, followed by temp1
immediately going UP, at which time I opened the case
door and turned up the case fans to high, and then
both temperatures dropped from there, so you can just
skim the following:

############## BEGIN USELESS INFO ###############

NOTE: in the previous e-mail I incorrectly stated my
mobo as A7V-266-E, when it is actually A7V-133-C. My
Windoze computer is the A7V-266-E. Same sensor chip
for both, so it doesn't change anything.

Duron 1.1ghz, Asus A7V-133-C, Thermaltake Volcano 9,
Lian-Li PC-69 aluminum.
Server: contains one 7200rpm HD, one AGP vid card
(only use console, no GUI), one NIC, and an LCD that
runs off the serial port.
Last but not least, a 275w PC Power & Cooling PSU.
That is it! Very empty indeed!

THE POINT IS THIS CASE RUNS COOOOOOOOOOL.

..except for the fact that my other 1.4Ghz Athlon
with 1280mb PC2100, Geforce4 128mb, WD Special, a
couple more HDs and a mess of other stuff is stitting
right next to it, runs HOT, at IDLE: 64C/147F proc and
27C/80F mobo with the case open. The room temperature
is around 79F.
see http://12.203.238.243/computers.jpg for a fun
picture :)


==At idle for days, case fans on low, proc HS/F at
low (~1700rpm)=
[root@localhost as99127f-i2c-0-2d]# cat < temp1 temp1
temp2 temp3
60.0 127.0 33.0
60.0 50.0 27.5
60.0 50.0 112.2

==Proc fan turned to high (~5800rpm)==

]# cat ...
60.0 127.0 34.0
60.0 50.0 27.2
60.0 50.0 112.2
]# cat ...
60.0 127.0 34.0
60.0 50.0 27.0
60.0 50.0 112.2
]# cat ...
60.0 127.0 34.0
60.0 50.0 26.5
60.0 50.0 112.2
]# cat ...
60.0 127.0 35.0
60.0 50.0 26.2
60.0 50.0 112.2
]# cat ...
60.0 127.0 35.0
60.0 50.0 26.0
60.0 50.0 112.2
]# cat ...
60.0 127.0 35.0
60.0 50.0 25.7
60.0 50.0 112.2
]# cat ...
60.0 127.0 35.0
60.0 50.0 25.5
60.0 50.0 112.2
]# cat ...
60.0 127.0 36.0
60.0 50.0 24.5
60.0 50.0 112.2

==Turned case fans to high, removed side door.=
]# cat ...
60.0 127.0 36.0
60.0 50.0 24.0
60.0 50.0 112.2
]# cat ...
60.0 127.0 35.0
60.0 50.0 24.0
60.0 50.0 112.2
]# cat ...
60.0 127.0 35.0
60.0 50.0 23.5
60.0 50.0 112.2
]# cat ...
60.0 127.0 35.0
60.0 50.0 23.2
60.0 50.0 112.2
]# cat ...
60.0 127.0 34.0
60.0 50.0 23.0
60.0 50.0 112.2
]# cat ...
60.0 127.0 34.0
60.0 50.0 22.5
60.0 50.0 112.2
]# cat ...
60.0 127.0 33.0
60.0 50.0 22.5
60.0 50.0 112.2
]# cat ...
60.0 127.0 33.0
60.0 50.0 22.0
60.0 50.0 112.2
[root@localhost as99127f-i2c-0-2d]# cat < temp1 temp1
temp2 temp3
60.0 127.0 33.0
60.0 50.0 22.0
60.0 50.0 112.2
==Steady temp=
############## END USELESS INFO #################

Thank you for coming this far.

My method of "installing" lm_sensors-2.7.0 was what
was in the QUICKSTART file in:

http://www.lm-sensors.nu/archive/lm_sensors-2.7.0.tar.gz

"Quickstart for i2c option 1 (modules) and lm_sensors
option 1 (modules).
This is the recommended EASIEST METHOD."

No doubt I screwed it up.

Hopefully, you will be able to tell me some possible
formulas to try with that "raw" temp2 value so it is
represented in a degree format I can use.

Also, since sensors.conf is useless, I suppose the
answer to this question is "You can't" also: At fan
speed set to lowest, fan2 (the only fan showing
anything) is 0. At max it shows:

[root@localhost as99127f-i2c-0-2d]# cat < fan2
3000 5487

Anything below ~2600rpm shows 0 (when it does show a
value, it is exactly what the mobo would report in the
BIOS's hardware monitor). If this cannot be fixed, no
biggie. The processor temperature is my main concern,
as I generally do NOT have any computer monitor hooked
up to the video card (I may even remove that), so I
have to rely on this program to report the temperature
to me.


Again, thank you for reading this entire e-mail (or
most of it). Any more info you request will be
availible to you. I look forward to hearing from you.

Thanks,
Brian




--- Jean Delvare <khali@linux-fr.org> wrote:
> 
> > I installed i2c-2.7.0 and lm_sensors-2.7.0, and
> > "values" are reported in
> > /proc/sys/dev/sensors/as99127f-i2c-0-2d/
> "properly",
> > but they are way off:
> > 
> > ]# cat < temp1 temp1 temp2 temp3
> > 60.0 50.0 33.0
> > 60.0 50.0 27.0
> > 60.0 50.0 112.2
> 
> What a strange command. What were you trying to do?
> (I mean, "cat temp1
> temp2 temp3" does just the same, so what's the big
> idea?)
> 
> > those are in deg C, right? temp1 seems to be ok
> > assuming it is the mobo. I dunno what the hell
> temp2/3
> > are supposed to be.
> 
> They do not have to be in deg C. They can happen to
> be, and actually I
> believe temp1 is, in your case. But temp2 and temp3
> are probably not.
> What you get in /proc/sys/dev/sensors are raw
> readings. You have to use
> the formulae in /etc/sensors.conf (or, more
> commonly, let libsensors do)
> in order to get them in deg C.
> 
> > I try editing /etc/sensors.conf, but no changes to
> > compute temp2/temp2 have any effect. i try to do
> > "sensors -s" but it says "command not found"--it
> never
> > got installed (I cant find it anywhere on my
> system).
> 
> Logical. The /etc/sensors.conf has no effect on
> /proc/sys/dev/sensors
> values. You have to use an application that itself
> uses libsensors to
> see the effects of any change done on the file.
> 
> > Something tells me that getting 'sensors'
> installed
> > will be a bitch, so is there a way to "manually"
> > update the compute line in libsensors somewhere?
> 
> You simply cant.
> 
> > it must be stored somewhere, because I restart the
> > computer, and the values for temp1/2/3 never
> change,
> > so the sensors.conf file must not be being
> reloaded
> > (it is set to startup, which works).
> 
> Raw values don't even care about the config file.
> 
> You have to get the sensors file. You did not tell
> us how you installed
> lm_sensors, but I guess it comes from a package
> designed for your
> distribution. You should look for other packages,
> there *must* be one
> that provide sensors (because the drivers aren't
> useful to anything if
> you can't issue a "sensors -s" at startup).
> 
> -- 
> Jean Delvare
> http://www.ensicaen.ismra.fr/~delvare/




__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com

'sensors' never installed

[Mark D. Studebaker]

I'm not even going to attempt to answer any of the rest of it but for
an explanation of fan monitors and why anything below 2600 is 0 see doc/fan-divisors.

Brian Papantonio wrote:

> Also, since sensors.conf is useless, I suppose the
> answer to this question is "You can't" also: At fan
> speed set to lowest, fan2 (the only fan showing
> anything) is 0. At max it shows:
> 
> [root@localhost as99127f-i2c-0-2d]# cat < fan2
> 3000 5487
> 
> Anything below ~2600rpm shows 0 (when it does show a
> value, it is exactly what the mobo would report in the
> BIOS's hardware monitor). If this cannot be fixed, no
> biggie. The processor temperature is my main concern,
> as I generally do NOT have any computer monitor hooked
> up to the video card (I may even remove that), so I
> have to rely on this program to report the temperature
> to me.
> 

'sensors' never installed

[Jean Delvare]

> because...
> 
> [root@localhost as99127f-i2c-0-2d]# cat < temp1 temp2
> temp3
> 60.0 50.0 22.0
> 60.0 50.0 112.2
> 
> it "eats" the output for the first value! don't ask me
> why, I am a linux noob. So I just wrote 'temp1' twice
> to be used as a dummy value. it means nothing.

This is because you don't *need* to use '<'. This is probably not the
place for a complete lesson about shell basics, but if you use "cat
temp1 temp2 temp3" (*without* the '<') you'll get what you want.

> I agree that temp1 is the motherboard temp in C. It
> agrees nicely with what the hardware monitor in my
> BIOS setup program displays. It also agree with the
> fact that temp1 computation doesn't exist in the
> sensors.conf file which has proven to be useless
> anyway.

At least one that is working :)

> For the processor, the BIOS program reports 58C/136F
> with case and HeatSink/Fan set to low. The minimum
> i've seen in the BIOS is 46C/114.5F with side door
> removed and proc/case fans set to highest. temp3 MIGHT
> have been close as F, EXCEPT for the fact that I just
> noticed it NEVER moves. I don't have any idea where it
> comes from. Temp2 is active, and SEEMS to be the
> processor, however, the original output of "26.0" or
> "22.0" is obviously miscalculated.

Temp3 may simply be not linked to anything. Most motherboards only have
two temperature sensors, one for the CPU and one for the motherboard.

> I think using that information, at the very least I
> should be able to interpolate the final values, but
> depending on the computation used to obtain "22/26",
> it may not be very accurate since when dealing with
> temperature, the value/starting point of 0 you would
> use when comparing two sets of values is meaningless,
> especially since I don't know if "22/26" is C, F, or
> neither.

This is a raw value, that is actually in "neither" unit, as you say. See
below for more details.

> NOTE: in the previous e-mail I incorrectly stated my
> mobo as A7V-266-E, when it is actually A7V-133-C. My
> Windoze computer is the A7V-266-E. Same sensor chip
> for both, so it doesn't change anything.

It changes much if you consider that I own an A7V133C here, which means
I *must* have a working config for it :)

(skipped, the last part which is pointless I think)

OK, Let's make things clear.

1* You *need* the "sensors" program to have your sensors working
correctly.

2* Compiling and installing the "sensors" program is easy, if not
trivial. The program is included in the source package you say you
downloaded and installed. It is installed by default in /usr/local/bin
when you run "make install".

So, you please carefully take another walk through the installation
process and see if anything fails there. Check in /usr/local/bin that
the program is installed. Check that /usr/local/bin is in your PATH
environment var ('echo "$PATH"' will let you know). If anything fails,
we can't guess what it is unless you give us a details log of what you
did and how the installation process behave.

For what it's worth, the correct temp2 formula for your motherboard is:
    compute temp2 @*2.0, @/2.0
Which means that the 22-26 values you are seeing are actually in a "2
deg C" unit. Your CPU must be living around 48 deg C (which makes
sense).

-- 
Jean Delvare
http://www.ensicaen.ismra.fr/~delvare/