Do you have a question? Post it now! No Registration Necessary. Now with pictures!
- Posted on
- Y Split Power Dist
- Bryan Hoover
August 24, 2005, 2:51 pm
rate this thread
[sorry if appears twice - didn't appear to have gotten out the first
I'm thinking about Y splitting a 4 pin molex, and adapting it to
connect to two 3 pin fan headers. This would feed power to an Aspire
fan speed controller that attaches to the case as a pci slot bracket.
Then, of course, I would Y split the power output molex, and adapt to
my two 3 pin fan plugs.
The board 3 pin fan headers are restricted to 1.2A. Would the
described set up equally distribute the amperage over the two headers?
Or is there danger one header could be "called upon" for more amperage
The fans are .32A each, so I don't suppose drawing off a single header
would be a problem, but still, I wonder about the nature of
distributing current relative to this set up - I mean, is it
"naturally" distributed (and equally so) according to flow through the
circuit? Or are there variables that could result in one header
bearing more than another? I may also, at some point, want to connect
more than just the two cpu fans, so the question may be more important
at that time.
In case wondering, I'd like to do this so can have ease, and
flexibility of software speed control through the board headers, and
still have control when reboot, or running DOS, or other system
without software control.
Re: Y Split Power Dist
On 24 Aug 2005 10:51:07 -0000, Bryan Hoover
There is no "equally" to it, the component at the end
dictates the amperage, the wires merely have to be
sufficient to sustain that current. Yes, it will be more
than sufficient for a couple of *normal* fans.
What board is "the board 3 pin fan headers" referring to?
Look at the fans, their ratings. Because of the fan
controller, the current requirement will only go down from
that. It's very rare for two fans and a controller to use
even 800mA. Your description isn't clear though, if you
plan on doing extensive reworking of the electrical system
you might use a multimeter to measure current.
There is not "distribution" per se, the weakest link needs
to sustain two fans, at .32A each max = 0.62A, plus the
current of the controller if it's an active rather than
passive type of controller.
From a reliability standpoint, it is generally good to
minimize the number of mechanical connections and splitters.
If you plan on several fans, buy a fan controller that has a
4-pin molex power input and several 3 pin outputs, one per
IMO, it is a VERY bad idea to use software for fan control.
The cooling system should be as reliable as possible, which
software is not, based on several variables including that
software, the OS, the stability of the system.
If you really "need" variable control, set it either
manually or with hardware sensors and no logic controller.
Re: Y Split Power Dist
Main board. Tyan s2462.
Sounds like you understand.
The "controller" is here:
It has two 4 pin connectors - a power in, and a power out,
essentially. If you want more than one thing going in or out, you've
got to split those connectors. For power in, I thought to use two 3
pin fan headers. This way, in addition to the controller power knob,
software can be used to control the power delivery through the board.
As you say however, I might as well use a single header - no advantage
using both as there would be no amperage distribution, as it were.
It's either that, or use two rheostats instead, which I may indeed do.
Only problem is, using the Aspire ones would take up two pci bracket
slots. OTOH, there are the Zalman fan mate rheostats. The Aspire is
so easy to mount so as to access without opening the case though.
Yes. Another consideration.
I suppose it's true that software could have a funky failure mode
wherein the fan speed was erroneiously lowered below the rheostat
tuned maximum, but couldn't similar failure be said of bios controlled
As for really needing, frankly, I don't, in fact, much use the
automated speed control aspect. That is, I don't allow the software
to automatically raise, and lower fan speed. At this point in
development of the software, it's a little too erratic in responding
to temp changes, and raising, lowering speeds. However, I *do* like
to be able to turn speed up or down, if resired, with a mouse click,
rather than reaching around to the back of the machine to do it.
Can you honestly imagine a situation where a software controller
failed, and set the speed below what it'd been set to before failure?
It's certainly within the realm of possiblity. I just don't think
it's very likely, and would appreciate any ellaboration you might be
able to add.
Re: Y Split Power Dist
On 25 Aug 2005 00:30:25 -0000, Bryan Hoover
It is a passive controller, it adds no additional current
draw so the max would be additive of the max current per
each fan, but of course having the controller presumes you
want to further reduce their speed, so the current will be
even lower. "Likely" you'll end up using under 300mA with a
couple of "typical" fans.
You might need to research the rating for the controller,
being a passive type it is possible (more easily) to
overload by having too many fans connected. The Newegg page
reads 1.2A, but personally I'd try to keep it under 1A.
Or use one of the front-bay mounted controllers which often
provide for 3 or more independant controls. They used to be
pretty expensive but are now under $20.
Not necessarily, the slot just has a hole drilled (or
possibly stamped in this case) into it, if you wanted to
combine two of them you'd just need to drill a 2nd hole in
one and move the rheostat over to it.
Do you plan on frequently adjusting the fan speed?
Personally I prefer an inline resistor on each fan, around
47 to 120 Ohms is typical... it's more reliable, lower cost,
and takes up no extra space. Downsides are of course that
you can't adjust it later, and along with that you need know
what value of resistor is needed to achieve the target.
I would put more trust in Award/etc bios not crashing than
Windows et al? If it's a buggy bios though, sure it's
another concern... actually to quote Bob Peace, in this case
"My favorite programming language is solder."
Sounds like you'd be as well off with a front-mounted
control. What I"d mentioned previously about drilling a
hole in the slot bracket, could still apply similarly to
drilling a hole or two elsewhere in the case, including in
an unused case front bay cover plate.
If the software were depended upon to raise fan RPM with
thermal rise and the temp rose but the software didn't
respond, wouldn't it then start going above the desired
I just don't see the need to introduce any additional
potential for problems. It's not really necessary to
maintain the exact same temp anyway, and having fans adjust
to rapid temp changes (like what might be seen on a video
card or CPU) can actually be more distracting than a
constant tone that is louder than the lowest heat RPM range
the fan would settle into. Constant tones tend to blend
into the background more if not "too" loud.
Regardless, your plan can work, I'd just more rigorously
test it since there is more potential for problems.
- » Exasperating graphics problem, motherboard related?
- — Next thread in » Home-built Computers