CPU fan question

  • Thread starter Thread starter Jethro
  • Start date Start date
Jethro said:
Why 3 wires on my CPU fan?

Jethro
Click to expand...

The three wires are +12V, GND, and RPM.

The first two power the fan, and the third allows the motherboard
to monitor the fan speed. To vary the speed of a fan with three
wires, you vary the +12V lead. Less voltage on +12V, less speed. The
RPM signal pulses twice, for every full rotation of the fan blade.
The motherboard measures the time between pulses, to figure out
the speed.

Intel has a four pin cooling fan. That one is +12V, GND, RPM, and PWM_control.
The PWM_control is used for setting the fan speed. In that case, the +12V
stays at 12V. The PWM_control is a signal at 25KHz (above human hearing).
The pulse width of the PWM_control signal, indicates the desired speed.
A 50% wide signal might indicate half speed. A 75% wide pulse in PWM
might mean three quarters speed. Something along those lines.

So that is why there are three and four pin connectors on computer
fans. The extra pins allow monitoring the speed (RPM) and setting
the speed (PWM_control).

Paul
 
Why 3 wires on my CPU fan?

Jethro
Click to expand...

It's more expensive to get an RPM reading from only two
wires, and with many of the newer fan controllers integrted
onto mainboards, harder to do. (so #3 is RPM on a typical
fan, it is actually up to the fan manufacturer what feature
they want to provide on the *3rd* lead, in industrial
equipment fans you may sometimes find the 3rd lead is rotor
lock or a speed control line instead, but either of these
latter two are quite uncommon in a PC).
 
Paul said:
The three wires are +12V, GND, and RPM.

The first two power the fan, and the third allows the motherboard
to monitor the fan speed. To vary the speed of a fan with three
wires, you vary the +12V lead. Less voltage on +12V, less speed.
Click to expand...

That's the old and simple way. Nowadays you modify the duty cycle so
the fan still get 12V but not 100% of the time. That's how Speedfan and
Motherboard Monitor work to vary the fan speeds.
Intel has a four pin cooling fan. That one is +12V, GND, RPM, and
PWM_control.
The PWM_control is used for setting the fan speed. In that case, the
+12V
stays at 12V. The PWM_control is a signal at 25KHz (above human
hearing).
The pulse width of the PWM_control signal, indicates the desired
speed.
A 50% wide signal might indicate half speed. A 75% wide pulse in PWM
might mean three quarters speed. Something along those lines.
Click to expand...

A lot of extra logic in the fan that isn't needed. PC fans have been
capable of being duty-cycle controlled for a decade, maybe longer. It
all depended on what the mobo maker provided for voltage control for the
fans.
 
the fan still get 12V but not 100% of the time. That's how Speedfan and
Motherboard Monitor work to vary the fan speeds.

capable of being duty-cycle controlled for a decade, maybe longer. It
all depended on what the mobo maker provided for voltage control for the
fans.
Click to expand...

Most DC motors can be controlled with PWM. In fact, I remember playing
with one from some electric drapes that was designed to defeat pwm
control, but after a bit of ajustment of the pwm control, PWM worked
just fine.

It also works very well for lights. And even solenoids.
 
[email protected] said:
.... snip ...

Most DC motors can be controlled with PWM. In fact, I remember
playing with one from some electric drapes that was designed to
defeat pwm control, but after a bit of ajustment of the pwm
control, PWM worked just fine.

It also works very well for lights. And even solenoids.
Click to expand...

Not for fluoroscents, including compacts. However it can control
PM DC motors down to a crawl. Very useful for model trains.
 
That's the old and simple way. Nowadays you modify the duty cycle so
the fan still get 12V but not 100% of the time. That's how Speedfan and
Motherboard Monitor work to vary the fan speeds.


A lot of extra logic in the fan that isn't needed. PC fans have been
capable of being duty-cycle controlled for a decade, maybe longer. It
all depended on what the mobo maker provided for voltage control for the
fans.
Click to expand...


Not necessarily, I don't recall which models (of fan) but
there have been people reporting problems burning out fans
controlled by PWM from bay controllers.

Also they don't entirely work as you described, they do not
deliver 12V to the fan at less than 100% of the time, that
is just the output BEFORE the smoothing capacitor that
ultimately results in a charge/discharge average a bit
lower.
 
kony said:
Not necessarily, I don't recall which models (of fan) but
there have been people reporting problems burning out fans
controlled by PWM from bay controllers.

Also they don't entirely work as you described, they do not
deliver 12V to the fan at less than 100% of the time, that
is just the output BEFORE the smoothing capacitor that
ultimately results in a charge/discharge average a bit lower.
Click to expand...

For motor control you don't filter the PWM output. It uses full or
zero voltage out. This allows the motor to run slowly. At the
same time (for a PM motor) the back EMF during the off period can
measure the speed. Two wire control can then be very accurate.
 
The three wires are +12V, GND, and RPM.

The first two power the fan, and the third allows the motherboard
to monitor the fan speed. To vary the speed of a fan with three
wires, you vary the +12V lead. Less voltage on +12V, less speed. The
RPM signal pulses twice, for every full rotation of the fan blade.
The motherboard measures the time between pulses, to figure out
the speed.

Intel has a four pin cooling fan. That one is +12V, GND, RPM, and PWM_control.
The PWM_control is used for setting the fan speed. In that case, the +12V
stays at 12V. The PWM_control is a signal at 25KHz (above human hearing).
The pulse width of the PWM_control signal, indicates the desired speed.
A 50% wide signal might indicate half speed. A 75% wide pulse in PWM
might mean three quarters speed. Something along those lines.

So that is why there are three and four pin connectors on computer
fans. The extra pins allow monitoring the speed (RPM) and setting
the speed (PWM_control).

Paul
Click to expand...


I have to ask then - wouldn't XP know when the fan is not running (for
whatever reason)? And if so, couldn't it (shouldn't it) shut the
system down or at least warn you in red? with beeps? I just ran into
an experience wherein my CPU fan burned out, and the CPU got hot
enough to crash the system. It all happened unbeknownst to me except
for the crash. The fan being inside the case was not visible until I
opened the case. Then I could see that the fan would not start up. Of
course, that was too late, No CPU damage, but there could have been.

Jethro
 
I have to ask then - wouldn't XP know when the fan is not running (for
whatever reason)? And if so, couldn't it (shouldn't it) shut the
system down or at least warn you in red? with beeps? I just ran into
an experience wherein my CPU fan burned out, and the CPU got hot
enough to crash the system. It all happened unbeknownst to me except
for the crash. The fan being inside the case was not visible until I
opened the case. Then I could see that the fan would not start up. Of
course, that was too late, No CPU damage, but there could have been.
Click to expand...


XP, by itself would not warn you of a fan problem...

There is sometimes a bios option to shutdown. throttle back or warn in the
case of high CPU
temperature.

Also there is software for such that works with XP to do the same...
It might even be on the cd that came with your motherboard.

(A utility such as Speed Fan may work with your board)
 
XP, by itself would not warn you of a fan problem...

There is sometimes a bios option to shutdown. throttle back or warn in the
case of high CPU
temperature.

Also there is software for such that works with XP to do the same...
It might even be on the cd that came with your motherboard.

(A utility such as Speed Fan may work with your board)
Click to expand...
Thanks

Jethro
 
kony wrote:

For motor control you don't filter the PWM output. It uses full or
zero voltage out. This allows the motor to run slowly. At the
same time (for a PM motor) the back EMF during the off period can
measure the speed. Two wire control can then be very accurate.
Click to expand...


You could, but it's typical for a motherboard to have
capacitance just before the fan header.
 
I have to ask then - wouldn't XP know when the fan is not running (for
whatever reason)? And if so, couldn't it (shouldn't it) shut the
system down or at least warn you in red? with beeps? I just ran into
an experience wherein my CPU fan burned out, and the CPU got hot
enough to crash the system. It all happened unbeknownst to me except
for the crash. The fan being inside the case was not visible until I
opened the case. Then I could see that the fan would not start up. Of
course, that was too late, No CPU damage, but there could have been.

Jethro
Click to expand...


I wouldn't want to trust windows for this, and hardware is
in theory supposed to be able to protect itself without
relying on one particular OS, rather a bios setting.

Windows doesn't "know", it is not a built-in feature as it
is something subject to variables MS could not resolve
except through knowing board details... and these being
boards that might be designed after windows' release and
any particular service pack.

If you want the system to shut down, use the manufacturer's
hardware monitor software or the bios setting. You don't
have to care if the fan isn't working if the CPU overtemp
shutdown still works. For a critical use system, it may be
prudent to use a high quality, lower RPM fan so the odds of
it failing are substantially lower, and to attempt to find
the temp threshold for instability (while running a Prime95
Torture Test, for example), then setting the shutdown
threshold temp in the bios to handle that.

Then again, other people would rather have an error but try
to complete, end, etc, their work instead of an abrupt
forced shutdown.
 
Back
Top