P4C800-E Deluxe - CPU Fan Speed Conrol?

[Eli]

What mechanism controls the CPU fan speed?

I recently built two nearly identical machines around the P4C800-E
Deluxe motherboard. Both with P4 3.0GHz Northwood processors and
using the stock Intel heatsink & fan with thermal paste instead of the
supplied pad. The only real difference between the two machines is
the case and PSU. The first machine has a larger, better ventilated
(and noisier) case, while the second is smaller, with just one 80mm
case fan.

Motherboard monitor on machine A show motherboard temperatures
typically about 32C and CPU temps 33C at idle. In the second machine
I have similar CPU temps, but the motherboard temp usually reads about
37C. The CPU fan in machine A is at about 2700RPM at idle and may go
up to 3200RPM under load. In machine B I have CPU fan speeds anywhere
from 4000RPM to 4600RPM. Needless to say, the CPU fan noise from the
second machine is louder and the fan drone is pretty annoying.

So, does the P4C800-E Deluxe use the motherboard temp sensor to
control the CPU fan speed? I would have thought it would only use CPU
temps to control this fan, which should give me similar fan speeds in
the two machines.

 

[Paul]

What mechanism controls the CPU fan speed?

The Intel retail fan, has an air temperature sensor. The design
trip points differ, depending on which kind of processor it is,
but basically, the fan speed is at a minimum, until the measured
air temperature passing through the CPU fan hits 30C. The fan
RPM rate increases until the air temperature hits about 38C, where
at that point the CPU fan ramps up to max RPMs. The Intel CPU fan
doesn't use the actual CPU temperature, but instead senses the
air temperature, and attempts to compensate for the loss in
cooling ability, of the hot case air, by speeding up.

The motherboard has a feature called Q-fan. It can be operated
at the same time as the Intel fan feature. Q-fan ramps up at
a measured CPU temperature of 50C and hits max RPMs at 60C.

When you combine the two cooling curves, the CPU will hit 50C,
before you see any real change. With Q-fan disabled, I think you
will see the behavior you've currently got. (4600RPM on the
motherboard with the 37C mobo/case air temp.)

I recently built two nearly identical machines around the P4C800-E
Deluxe motherboard. Both with P4 3.0GHz Northwood processors and
using the stock Intel heatsink & fan with thermal paste instead of the
supplied pad. The only real difference between the two machines is
the case and PSU. The first machine has a larger, better ventilated
(and noisier) case, while the second is smaller, with just one 80mm
case fan.

Motherboard monitor on machine A show motherboard temperatures
typically about 32C and CPU temps 33C at idle. In the second machine
I have similar CPU temps, but the motherboard temp usually reads about
37C. The CPU fan in machine A is at about 2700RPM at idle and may go
up to 3200RPM under load. In machine B I have CPU fan speeds anywhere
from 4000RPM to 4600RPM. Needless to say, the CPU fan noise from the
second machine is louder and the fan drone is pretty annoying.

So, does the P4C800-E Deluxe use the motherboard temp sensor to
control the CPU fan speed? I would have thought it would only use CPU
temps to control this fan, which should give me similar fan speeds in
the two machines.

The Intel fan is doing what it is supposed to do, on the second
machine. If you add a case fan to the second machine, that will
bring the case air temp down, and inside the computer, the Intel
fan will slow down in response. In any case, AMD recommends a delta
between case air and room air of no more than 7C. You should add
case fans to achieve that objective. Doing so helps the most
temperature sensitive part, which is the disk drive. The CPU can take
more heat, and you could enable Q-fan, to quiet the second machine.
If the processor temp is shooting past 60C, and the Intel fan
running at full speed cannot cool it, then at the very least you
need a better CPU cooler and/or more case fans.

Test with Prime95 and make sure there are no detected computing
errors, while doing your thermal testing (mersenne.org). Prime95
will give you a 100% load and ensure you have a worst case test
condition to work with.

HTH,
Paul

 

[Ken]

So, does the P4C800-E Deluxe use the motherboard temp sensor
to control the CPU fan speed?

Yes, in BIOS you can control the CPU fan speed.
11/16 lowest 15/16 highest

 

[Eli]

The Intel retail fan, has an air temperature sensor. The design
trip points differ, depending on which kind of processor it is,
but basically, the fan speed is at a minimum, until the measured
air temperature passing through the CPU fan hits 30C. The fan
RPM rate increases until the air temperature hits about 38C, where
at that point the CPU fan ramps up to max RPMs. The Intel CPU fan
doesn't use the actual CPU temperature, but instead senses the
air temperature, and attempts to compensate for the loss in
cooling ability, of the hot case air, by speeding up.

The motherboard has a feature called Q-fan. It can be operated
at the same time as the Intel fan feature. Q-fan ramps up at
a measured CPU temperature of 50C and hits max RPMs at 60C.

When you combine the two cooling curves, the CPU will hit 50C,
before you see any real change. With Q-fan disabled, I think you
will see the behavior you've currently got. (4600RPM on the
motherboard with the 37C mobo/case air temp.)

Thanks, Paul. That explains it. At the current CPU temperatures the
fan's speed must be purely controlled by its own air temperature
sensor.

 

[Formerprof]

This thread induced me to check the temperatures on my P4C800E-Deluxe,
Northwood 3.0 gHz P4. I was distressed to see that the CPU temperature was
about 60 c. I opened the case to see whether all the fans were running and
found that the slots on the Intel CPU fan were dust-clogged. I cleaned them
out using a tiny Rubbermaid vacuum cleaner and the CPU temperature dropped
back to 48 - 49 c. That's something I'll be doing every six months or so
from now on.

formerprof

 

[High Score]

I've been told to avoid using a vacuum cleaner. The air moving through
the nozzle can cause a static charge to build up. This charge could jump
from the nozzle to a vulnerable component ... Zap !!!!

You can get static controlled vacuums and also maybe could ground the
nozzle, but I find canned air to work well without the risk.

OTOH, maybe this is an urban legend. Anyone know of bad things happening
with a vacuum?

-G