"Wes Newell" wrote in message news:v8DGi.1334$re2.1007@trnddc02...
>
> It doesn't matter if it has 2 or 3 wires. The third wire is just a
> speed
> indicator. What matters is if his MB has speed control support for
> the
> header he's plugged into. It can control the speed of a 2 wire fan
> just
> like it can a 3 wire fan. And it does do this with voltage changes.
> The
> only difference is that you obviously won't be able to monitor the
> speed
> of the 2 wire fan. Now if you don't believe that, just cut the third
> wire
> and hook it up to a speed controlled header and try it. I just
> finished
> doing it to be sure I wasn't going crazy after reading your very
> convincing BS.
The 3rd wire is for speed *sensing*, yes. How do you know if changing
the duty cycle on the fan has reduced speed unless you can actually
measure it? If you change the duty cycle from 100% to 50%, there may
be no noticeable difference in noise generated by the fan. Then fan
may have or may not have changed much in speed. The speed of rotation
of the blades is not linear to match the duty cycle. Often you will
need to lower the duty cycle below 20% before your ear can hear any
difference in the noise level even if the fan has reduced its speed
some. If you can't see what change there is in the RPM then you are
guessing by ear as to the effect of changing the duty cycle. Pianos
are tuned by ear by those with excellent pitch discrimination. You
shouldn't be adjusting computer components by ear.
The reason that the 3rd wire should be considered necessary when
changing the duty cycle of a fan is to ensure that it spins at some
minimal speed, if required. For example, you certainly wouldn't want
to be reducing the duty cycle on your CPU's fan below whatever the
BIOS deems as too low which results in powering down your computer.
On my host, I can get the duty cycle for the CPU fan down to 3%.
Below that and the fan spins too slow and too close to what the BIOS
thinks is too slow and would power off the computer. The case fan can
be lower than 3% but may not continue to spin once it gets down to
that duty cycle (Speedfan starts at 100% and then slowly decrements
the ratio to the configured minimum).
For a CPU fan, the speed sense line is required so the BIOS knows that
the fan is spinning and perhaps above some minimal threshold for RPM.
A case fan doesn't need a speed sense line as the BIOS usually doesn't
care if it is spinning or not. Since the OP is asking about a case
fan, it may have only 2 wires. With only 2 wires, the OP won't know
if the RPM is changing when altering its duty cycle. He could listen
by ear to see when it makes less or little noise but perhaps then the
fan is spinning too slow to keep down the internal temperatures. When
the fan is spinning above what his ear can easily detect for changes
in noise, or the change in noise is far more granular, he won't know
at what minimal speed he should run to keep at the minimal
temperatures but keep noise at a minimum. Try telling the cop giving
you a ticket that you were using the sound of your tires against the
pavement to regulate your speed.
The voltage is not changed that is sent to the fan. It is either zero
or 12V. It is a pulse. The less the duty cycle (the percentage of
time for the pulse within a fixed time interval) the less effective
power is delivered to the fan over that time interval, but it is still
a full 12V pulse. The voltage is not regulated downward in 1V or
other increments to some minimum, like 6V. If you want to regulate
the actual voltage delivered to the fan, you will need to insert a
rheostat in the power lead to the fan. I've seen some kits that sell
a rheostat that you can insert to lower the actual voltage delivered
to the fan, like down to 7V (I think the fans become unreliable at 6V
so they may not even begin to spin), like
http://www.vantecusa.com/p_nxp_201.html although much simpler kits are
available. However, most if not all motherboards have super I/O
chips, like the Winbond brand, where the duty cycle of a fan can be
modified and why *software* products like MBM and SpeedFan work. They
are software so they are obviously not inserting a rheostat in the
fan's power lead. The controllers on the motherboard don't contain an
adjustable rheostat to alter resistance in the power lead for the fan,
they don't control the voltage regulators on the motherboard (since
those voltages are delivered to far more components than just the
fans), and they don't waste the expense of a separate voltage
regulator for each fan header on the motherboard. There are some fans
that don't like to be ran at less than 100% duty cycle but those for
use in computers handle it just fine.
It depends on your motherboard's super I/O chip if it supports
modifying the duty cycle of the fans. I haven't ran across one in
over 7 years that didn't support duty cycling of the fans but then I
don't use every mobo made since then. I'm not going to dig into a
host right now but my understanding is that a MOSFET and capacitor
must provided for each fan header to provide the necessary
buffer/driver for controlling the fan, and that the monitor chips used
now have fan control on 3, or more, channels; however, without the
buffer transistor then control is not possible (the monitor chip only
provides control, not the power switching). MBM and Speedfan can
access the monitor chip (e.g., Winbond) but can't tell you if the
buffer circuit is included. You'll notice in MBM and Speedfan that
speed control is listed as a ratio, NOT as a voltage. You are
changing the duty cycle for the fan back in the monitor chip, not the
voltage delivered to the fan.
Personally I don't see much of a point in regulating the speed of the
fan if you can't actually see what IS the speed of the fan so the 3rd
wire for speed sense is required so you know what you are doing for
effecting changes. I suppose you could work blindly. It is also
possible to drive a car with your eyes closed.
Similar Threads
Linear Mode
