Prev: SATA on IS7-V2
Next: KV8 Pro 3rd Eye stops at post code 52
From: Folk on 14 Sep 2007 17:43
On Fri, 14 Sep 2007 13:56:33 -0500, "VanguardLH"
<VanguardLH(a)mail.invalid> wrote:

>"Folk" wrote in message
>news:209le31pveer5882fbtsk6i8a792nbqqoo(a)4ax.com...
>> Beryl wrote:
>>>
>>> Does that fan have a 3rd wire for speed sensing? I don't think so.
>>
>> I don't think that Abit's FAN EQ thingy works on fan speed for the
>> system fan. From what I can tell, it provides for different
>> voltages
>> based on a temperature variable. So, if the temp you're monitoring
>> (CPU or case temp) is, for example, 50C, then the fan could run at
>> 12V, but if it's 35C it could run at 6V.
>
>And where did you come up with that gem

From the manual. Here's a copy/paste from the relevant section of the
bios instructions:

SYS FanEQ Control Enabled
-Reference Temperature CPU Temperature
-Control Temperature High 50�C/104�F
-Control Temperature Low 35�C/86�F
-DC Fan Voltage High 12.0 V
-DC Fan Voltage Low 6.0 V

So the way I read that, the system would provide 6 volts to the fan at
35C and 12 volts when it reached 50C. Is this not correct?

> that any software is going to
>act as a voltage regulator, varistor, or rheostat? It's software, not
>hardware. It works by regulating the duty cycle in the fan
>controller: less duty (i.e., percentage of full-voltage pulse) means
>less effective power to the fan so it spins slower but the pulse is
>still at full voltage.
>
>>>> The cable I purchased can be seen here:
>>>> http://www.performance-pcs.com/catalog/index.php?main_page=product_info&products_id=608
>
>Irrelevant. You were asked how many wires the *fan* has. If it only
>has 2 wires, you will never be able to regulate its speed unless you
>insert additional *hardware* between it and its voltage source. While
>software can regulate the duty cycle for the voltage delivered to the
>fan, it still needs to know what is the RPM of the fan as feedback to
>know that it is making an effective change. So, back to the question,
>how many wires does the fan have? 2 or 3 wires?

If the system is modifying the voltage based on temperature, then why
does it need to know how fast the fan is spinning? But to answer the
question, it only has two wires...

But why the attitude? I'm just trying to solve a problem and I
thought I asked the question in a civil manner.
From: VanguardLH on 14 Sep 2007 22:24
"Wes Newell" wrote in message news:v8DGi.1334$re2.1007(a)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.

From: VanguardLH on 14 Sep 2007 23:21
As I discovered after reading the manual for the OP's motherboard, it
includes the uGuru chip. Because the OP paid for the extra hardware
then perhaps it does perform voltage-regulation on the case fans. The
manual shows PWM (pulse width modulation) for duty cycling the CPU fan
but voltage thresholds for the system and case fans. I haven't
bothered with the added expense of uGuru-equipped Abit mobos. So
perhaps someone has a technical description of how their uGuru chip
works because Abit isn't revealing anything at their web site that I
could find.

From: Folk on 14 Sep 2007 23:33
On Fri, 14 Sep 2007 21:24:47 -0500, "VanguardLH"
<VanguardLH(a)mail.invalid> wrote:

>"Wes Newell" wrote in message news:v8DGi.1334$re2.1007(a)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.

>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.

The case fan I'm trying to control is the stock fan in an Antec Solo
case. Antec calls it a "TriCool", but that's just market-speak for a
manual 3-speed fan control switch that is attached to the fan. There's
a noticeable difference in loudness between the three speeds, with the
lowest being whisper quiet and the highest being quite loud. It would
be very easy to tell if the fan was running at a lower speed.

You're insinuating that there's no way to tell if a fan is running
slower without a tachometer. That's why we have ears. Slower fan
speed = less noise otherwise what's the point in all the fancy fan
controllers found on high-end motherboards?

I agree that the system needs to know if a critical fan has stopped,
and that's why the CPU fan on this motherboard is controlled
differently than the other fans as shown:

CPU FanEQ Control Enabled
-Reference Temperature CPU Temperature
-Control Temperature High 66�C/150�F
-Control Temperature Low 35�C/95�F
-Fan PWM Duty Cycle High 100 %
-Fan PWM Duty Cycle Low 60 %

So the CPU fan uses PWM and measures the RPM but the other fan
controls are simply analog.

Look at the graphic on the lower portion of this page:
http://techreport.com/articles.x/9219/2

This article is over two years old, but even then Abit was providing
extensive fan control options in their bios. You've never heard of
this? The concept of lowering voltage to quiet a fan is news to you?

Anyway, in trying to find a solution (haven't found one yet) I've come
to the conclusion that suspecting the cable as the problem was dumb on
my part. If the fan is getting voltage at all (and it runs fine at
full speed on all five auxiliary fan controllers) then the cable must
be good. So, I guess that points to the mobo as the problem. I'll
ask around in some of the online forums to see if anyone else has run
into this.
From: Folk on 14 Sep 2007 23:47
On Fri, 14 Sep 2007 22:13:16 -0500, "VanguardLH"
<VanguardLH(a)mail.invalid> wrote:


>I was hurried at the time.

Still no excuse for your attitude. If you were in such a hurry maybe
you shouldn't have replied.

>Without the speed sense,
>however, I'm not sure how the uGuru logic is going to know if it
>reduced the RPM too low (but then there is no "too low" setting
>because it cannot detect the RPM which probably means uGuru sees it at
>zero RPM). Could be uGuru changes the voltage and simply *hopes* that
>the fan speed changed. Or maybe because it sees the fan is at zero
>RPM so it can't do any regulation at all (it's a dead fan to uGuru so
>it doesn't lower the normal voltage) - which is what I suspect because
>your complaint was that the fan always spins at full speed (so it
>appears you really do need a case fan with a speed sense line).

That could be true, but the verbiage from the my case manual leads me
to believe that it can be controlled:

***************************************************************************************
The TriCool fan:
The case includes one 120mm TriCool fan installed in the rear. This
fan has a three-speed switch that lets you choose between quiet,
performance, or maximum cooling. (See specifications below.) The fan
is installed so that the air is blowing out of the case. Connect a
large 4-pin connector from the power supply to the male 4-pin
connector on the fan.

Note: The minimum voltage to start the fan is 5V. We recommend setting
the fan speed to High if you choose to connect the fan to a fan
control device or to the Fan-Only connector found on some of Antec�s
power supplies. A fan-controlled-device regulates the fan speed by
varying the voltage to it. The voltage may start as low as 4.5V to 5V.
Connecting a TriCool fan set on Medium or Low to a fan-control device
may result in the fan not being able to start. The already lowered
voltage from the fan control device will be further reduced by the
TriCool circuitry below 5V.
*****************************************************************************************

All my testing has been with the fan control switch on high.

I'm still searching for answers. And it's certainly possible that I'm
missing a setting somewhere. I'll post around on some of the online
forums. If I find a solution, I'll post it here.
First  |  Prev  |  Next  |  Last
Pages: 1 2 3
Prev: SATA on IS7-V2
Next: KV8 Pro 3rd Eye stops at post code 52