Intel DP35DP motherboard, no fan speed control?

[John Doe]

Anybody have an Intel DP35DP motherboard with an Intel heatsink/fan?
Does the fan speed increase much beyond the minimal 1050 rpm? The
motherboard manual says the BIOS should have options for fan speed
control but I don't see them. I did a BIOS maintenance boot and
noticed the QST switch was off. Turned that on but the fan speed
controls did not up in the BIOS after restarting and Windows did not
detect any new hardware. So I have a new E6850 CPU with a fan that's
stuck at 1050 rpm. Ack.

 

[Paul]

Anybody have an Intel DP35DP motherboard with an Intel heatsink/fan?
Does the fan speed increase much beyond the minimal 1050 rpm? The
motherboard manual says the BIOS should have options for fan speed
control but I don't see them. I did a BIOS maintenance boot and
noticed the QST switch was off. Turned that on but the fan speed
controls did not up in the BIOS after restarting and Windows did not
detect any new hardware. So I have a new E6850 CPU with a fan that's
stuck at 1050 rpm. Ack.

If the solution is not in the BIOS, then one way to fix it, is to
power the CPU fan from a Molex power connector.

OK, this one looks pretty good (but out of stock). The two drive
connectors, allow daisy chaining, so no power supply connectors
are wasted. The middle connector (fan, male, 3 pin) goes to the CPU fan
(and the PWM pin is not connected and floats). The right hand connector,
with the single yellow wire, goes to the motherboard CPU fan connector - the
yellow wire carries the RPM signal. The RPM signal helps keep the BIOS
happy, as some BIOS freak if no RPM signal is detected.

http://www.svc.com/3pinto4pinad1.html
http://us.st11.yimg.com/us.st.yimg.com/I/svcompucycle_1999_107234983

I'd use a wire assembly like that, so that the CPU fan receives full
12V all the time, no PWM signal (so the fan is told to run 100%).
That is about the best you could do. You may have enough odds and
ends in your junk box, to make a cable like that.

Paul

 

[John Doe]

The RPM signal helps keep the BIOS happy, as some BIOS freak if no
RPM signal is detected.

I hadn't considered that, thanks.

The manual says this.
"... a fan with a 3-pin connector can be used. However, since the
fan with a 3-pin connector cannot use the onboard fan control, the
fan will always operate at full speed."

So, I severed the fan speed control wire (the fourth wire) and the fan
speed is now almost double at 1975.

I guess that's full speed, but if not, maybe the fan speed sensor lead
can be left attached to the CPU fan header (by simply severing all but
that fan speed sensor lead) and just the V+ and V- leads can come
directly from the power supply.

 

[Paul]

I hadn't considered that, thanks.

The manual says this.
"... a fan with a 3-pin connector can be used. However, since the
fan with a 3-pin connector cannot use the onboard fan control, the
fan will always operate at full speed."

So, I severed the fan speed control wire (the fourth wire) and the fan
speed is now almost double at 1975.

I guess that's full speed, but if not, maybe the fan speed sensor lead
can be left attached to the CPU fan header (by simply severing all but
that fan speed sensor lead) and just the V+ and V- leads can come
directly from the power supply.

On a four pin fan connector:

+12V, GND provide basic power.
RPM is a signal going from fan to computer - for monitoring and should not be cut.
PWM is the speed control, and as you've already noted, cutting PWM should
result in 100% speed request at the fan.

This is the spec for the four pin fan interface. Section 2.1.4, page 9,
mentions PWM has a pullup inside the fan, and when the wire is cut,
that is how it can run at 100%. (PWM is a logic level signal, and the
"width" of the duty cycle of the square wave on the signal, sets the
fan speed. When the pullup drives it, it is effectively a 25KHz
square wave with a 100% duty cycle.)

http://www.formfactors.org/developer/specs/4_Wire_PWM_Spec.pdf

If you still don't have enough cooling, then you may want to pick
up an aftermarket heatsink/fan.

Paul

 

[John Doe]

On Fri, 09 May 2008 06:08:53 GMT, John Doe
<[email protected]> put finger to keyboard

my voice to the microphone

I think this behaviour may be normal.

If it's normal, then they need to update the owner's manual.

And enable the QST without having to do a BIOS maintenance startup.

I think mine has something to do with the fact I bought an open box
product and maybe that the DP35DP is problematic (leading to greater
availability of open box products). I'm not whining though, unless
something else comes up, it's already been resolved. I really don't
mind doing a little electrical work to make the fan run normally.

Thanks.

http://www.intel.com/cd/channel/reseller/asmo-na/eng/299984.htm

"Intel has recently developed a new revolutionary motherboard
based fan speed control included with Intel 965 chipset (and
newer) based motherboards called Intel® QuietSystem Technology
(Intel® QST). This new technology uses a PID controller that can
measure the rate of change of the processor temperature, thus
predicting when the processor will reach its maximum temperature.
If implemented correctly by the motherboard manufacturer, the
control algorithm will operate the processor fan at minimum speed
under most operating conditions. Since Intel® QST can predict when
the processor will reach its maximum temperature,

Like when it's approaching its maximum temperature.

it will delay increasing the fan speed until just the right moment
in order to keep the processor from exceeding its maximum
temperature."

Like a few degrees below its maximum temperature.

 

[John Doe]

On Sat, 10 May 2008 00:02:51 GMT, John Doe
<[email protected]> put finger to keyboard

I said "voice to the microphone".

I think you really need to monitor the fan speed as the CPU's
temperature rises. All you've said so far is that the fan spins at
its minimum speed at whatever temperature. As suggested by TM, try
loading up your CPU with a torture test, eg CPUburn or Prime95.

I monitor lots of stuff. The temperature has exceeded what I
consider to be acceptable without any increase in fan speed. With
the fan speed at about 2000 rpm after the fan speed control wire was
severed, the temperature under that load is almost in the yellow
area in Intel's Desktop Utilities CPU Monitor. Supreme Commander
pushes both cores to about 100% (according to Performance Monitor)
while using a CoreMaximizer utility that evens out the core load.

Speaking of multiple core CPUs and stuff. That utility is
impressive, IMO. Some amateur/shareware/whatever programmer wrote
that little core equalizer utility that distributes Supreme
Commander program threads more evenly across multiple cores. I read
that Intel's Core 2 Duo spreads the load evenly among the cores. And
Chris Taylor and his Gas Powered Games says Supreme Commander is
designed for multiple core CPUs. But, using that utility the game
runs significantly better on my machine, equalizing core usage that
is otherwise unbalanced during play (like 100% - 40%). But hey, at
least the utility is free.

 

[John Doe]

So, I severed the fan speed control wire (the fourth wire) and the
fan speed is now almost double at 1975.

I guess that's full speed,

Umm, it's not. At the moment, the fan speed is at 2800 rpm. I hate
to think the fan speed is random, but who knows. Maybe Intel CPU
fans take a while to loosen up.