Re: fan noise control

From: Jean-David Beyer (jdbeyer_at_exit109.com)
Date: 06/29/04 

Date: Tue, 29 Jun 2004 08:51:55 -0400

Richard Kimber wrote:

>> After I switched my BIOS to 1.31 beta, I saw my reported CPU temps
>> drop. Not sure if that means anything or not- I also modified the air
>> flow in the case at the same time. Current sensor output:
>>
>
>
>> CPU1 Temp: +45 C (high = +80 C, hyst = +60 C) sensor =
>> PII/Celeron diode CPU2 Temp: +46.5 C (high = +80 C, hyst = +60
>> C) sensor = PII/Celeron diode M/B Temp: +29.0 C (high = +80 C,
>> hyst = +60 C) sensor = thermistor
>
>
> I have CPU0: 48C and CPU1 49C with fan3 (an Acoustifan with its own
> thermistor) at 39C
>
> I suspect my readings are higher than yours because I have installed
> fairly thick acoustic felt inside the case, which can reduce cooling.
>
> One oddity is that CPU0 fan is running at 2700, yet CPU1 is reported as
> 0. I guess I have no control over this, though I'd prefer they were
> both running at 2700 or thereabouts. Maybe this is a sensors thing
> (I've noticed that sensors reports fan3 as zero, even though the
> readings in the bios say circa 1600.
>
>
>> CPU temps go up by <10C when compiling under full load.
>
> I haven't tried this yet.
>
I am in doubt about trusting the lm_sensors package. I have a dual XEON
server motherboard; i.e., the processors are right at the top right of the
motherboard right near the front panel of the tower. Processor 0 is the
top one and immediately below it is Processor 2. They both have the Intel
Processor Wind Tunnel over the 1 pound heat sinks. At the intake side of
the PWT is a 60 mm 1.5" thick Intel-supplied thermistor controlled fan.

My readings, are:

Tue Jun 29 08:30:00 EDT 2004
w83627hf-isa-0290
VCore 1: +1.44 V (min = +1.34 V, max = +1.47 V)
VCore 2: +3.31 V (min = +3.13 V, max = +3.45 V)
+3.3V: +3.24 V (min = +3.20 V, max = +3.45 V)
+5V: +4.97 V (min = +4.84 V, max = +5.24 V)
+12V: +11.97 V (min = +11.48 V, max = +12.58 V)
-12V: -11.83 V (min = -13.11 V, max = -11.41 V)
V5SB: +5.43 V (min = +4.84 V, max = +5.24 V)
VBat: +3.23 V (min = +2.40 V, max = +3.60 V)
CPU0 fan: 5818 RPM (min = 3000 RPM, div = 2)
CPU1 fan: 4041 RPM (min = 3000 RPM, div = 2)
fan3: 0 RPM (min = 0 RPM, div = 2)
System: +42C (limit = +60C) sensor = thermistor
CPU0: +53.5C (limit = +60C, hysteresis = +50C) sensor = thermistor
CPU1: +51.5C (limit = +60C, hysteresis = +50C) sensor = thermistor
vid: +1.400 V

These processors run quite a bit cooler if the processors are idle, but I
am running 4 instances (hyperthreaded processors) of setiathome in the
background, so the processors are seldom idle. I wish they were not quite
so hot, though. Now it is quite cool in this computer room this morning
(78F). When I first got the lm_sensors working, the CPU0 fan was running
about 7200rpm and by examining the geometry of the inside of the box, I
inferred a hotspot near the CPU0 intake fan. So I punched a hole in the
front panel and stuck a little 40mm fan there to blow filtered cool air
right into the CPU0 fan. The cooler air seems to have cooled the
thermistor and slowed that fan down. So I am not sure if I have done the
right thing or made things worse. Is fast hot air better than slow cool
air? Of course, those thermistors are in the worst possible place. They
should use the diode in the CPU chip itself, or at least, sense the hot
air coming out of the heat sink.

But can I believe any of this? The MB has 6 tachometer fan connectors and
I have tachometer fans plugged into all of them. The BIOS gives me
readings for all of them. But fan3 indicates it is not running even though
it is. The only fan that is not running is the overheat fan, and that one
is plugged into the overheat connector that is not connected to tachometer
monitor.

And the processors contain a temperature sensing diode, not a thermistor.
Perhaps there is a thermistor near the processor socket (I cannot tell).
But the sensors package only provides for a diode in the case of a Pentium
II (IIRC), and when I tried that, it announced that my processors were
running at 200C.

Are these monitors accurate enough to be useful? 

-- 
   .~.  Jean-David Beyer           Registered Linux User 85642.
   /V\                             Registered Machine   241939.
  /( )\ Shrewsbury, New Jersey     http://counter.li.org
  ^^-^^ 08:40:00 up 1 day, 16:09, 3 users, load average: 4.21, 4.18, 4.15

Relevant Pages

  • Marshall AVT50
    ... 24V fan would seem to be variable to temperature sensed by thermistor close ... to but not touching the heatsink at the o/p hybrid and no trace of white ... Solder on a number of W/W ...
    (sci.electronics.repair)
  • Re: How fast should CPU processor cooling fans run?
    ... >but when I put in two statements as below, the readings changed ... BTW, for fan 3, it ... of RPM. ... >sensor chip near the back edge of the mother board. ...
    (comp.os.linux.misc)
  • Re: Thermistors! Ah!!
    ... Does anybody know of a replacement which would do ... Easy method - repalce the thermistor, in the meantime stick a fan at ... Leads me to think the switching cct is ...
    (rec.radio.amateur.homebrew)
  • Intel L440GX+ Server Board - Temperature Monitoring
    ... I am trying to get fan RPM, temperature & voltage readings from Intel ... have IPMI management capabilities. ...
    (freebsd-questions)
  • Re: Strange fan behavior [longish]
    ... My only guess is the air temp passing over the thermistor in the fan is ... and they've all been above 50C without Q-Fan. ...
    (comp.sys.ibm.pc.hardware.chips)
Loading