Need a decent cooler for a stock Athlon 64 3500.

> Do you have details, as to what RPM range the motherboard doesn't
> like ?

No, that's what ****es me off. I have gone over the manual, but it
doesn't give any reference to it's limitations. Actually, I noticed
that there is a hell of a lot more adjustable options with the MSI
BIOs, but I am using the HP version of the board with their stripped
down version of the BIOs.

I would think that only HP could tell me what BIOs presets that they
locked into their custom version. But HP customer support was less
than helpful. All they basically told me was that this is something
that only "qualified people" should deal with.


> It almost looks like the CPU fan header is a 3 pin. The AC Freezer 64 Pro
> uses a four pin connector (according to the description on Newegg). Which is
> not a problem.

Yes, there are two 3-pin fan connectors on the board, one CPU and one
case fan.


> If you plug a four pin fan into a three pin header, the PWM pin won't
> have an electrical connection. A floating PWM allows the fan to
> run at full speed, so should not interfere with normal operation.
>
> In principle, a four pin fan should not be fed a variable voltage on
> the 12V pin. So you'd want to keep the fan running at 100%, so that
> the motherboard feeds 12V to it. (The reason for this, is the PWM
> uses a transistor inside the fan hub, and to help keep the transistor
> saturated, it helps if the fan is being fed full voltage. The condition
> they're trying to avoid, is having the transistor partially conducting,
> which increases the internal heat in the transistor.)

Have you heard of anyone using the hotwiring method that I mentioned
in my last post? Would that solve the transistor problem, if I just
went ahead and hotwired pin-4 to pin-2?


> If you wanted to use the AC Freezer 64 Pro, having an adapter cable
> handy, may give you an option to help it run full speed. If
> it turns out to handle the lower fan speed without a problem,
> then perhaps the adapter cable would not be needed.

Yeah, I should probably go ahead and get one, just in case I need it.

But the first link you gave has a reviewer that says only the 5v line
is connected to the fan. That wouldn't give me full speed, would it?

> There is yet another alternative. You can take a separate fan,
> and plug it into the CPU fan header, to keep the BIOS happy.
> You would then need a means to power the real CPU fan, as the header
> would be occupied by an RPM equipped fan, whose purpose was
> just to keep the BIOS happy.

I'm using a micro-ATX case, so I don't really have any more room. lol

I will have to go over the measurement just to be sure that I can fit
the Freezer 64 in there. But then there will be even less room!

ShadowTek wrote:

>
> Have you heard of anyone using the hotwiring method that I mentioned
> in my last post? Would that solve the transistor problem, if I just
> went ahead and hotwired pin-4 to pin-2?

I wasn't aware that it was necessary. I thought the idea was, there
was a pullup resistor in the 4 pin fan, so that if PWM was disconnected
and not driven, the fan would run at full speed.

Also, I'd be careful of that hotwiring advice. As far as I know, PWM
is logic level (5V square wave). Pin 2, at 12V, would be more than that.

OK, one of the Wikipedia articles lists this, an Intel spec for
the four pin fan and PWM. Intel posts specs on their formfactors site.

http://www.formfactors.org/developer...e_PWM_Spec.pdf

"Absolute maximum voltage level: VMax = 5.25 V (open circuit voltage)"

That is the max level for the PWM pin. There is also text there
to suggest Intel is interested in the square wave being only
3.3V, to make it easier to drive directly by SuperI/O hardware
fan controllers.

It would suggest that tying pin 4 to pin 2 = smoke :-)

Well, maybe not (the signal might be feeding the gate
on a MOSFET or something, and maybe it is safe all the
way to 12V), but doing such a butchery would not be my first
choice. If you connect a multimeter to the PWM pin, while
the 4 pin fan is powered, and measure the voltage on the PWM
pin, you might get a clue from that, as to what provision
they've made for a floating PWM signal. I'm guessing you'd
see a 5V DC level.

I don't really see a reason for the design to do anything
other than failsafe and run at full speed. Any other policy
is stupid (i.e. 4 to 2 should never be necessary, for any
company designing to the Intel spec). And this whole concept,
the 4 pin fan, is Intel's stupid idea. It was never necessary
in the first place. The world is a nicer place, with only three
pin fans - fewer tech support issues.

>
>> If you wanted to use the AC Freezer 64 Pro, having an adapter cable
>> handy, may give you an option to help it run full speed. If
>> it turns out to handle the lower fan speed without a problem,
>> then perhaps the adapter cable would not be needed.
>
> Yeah, I should probably go ahead and get one, just in case I need it.
>
> But the first link you gave has a reviewer that says only the 5v line
> is connected to the fan. That wouldn't give me full speed, would it?

I happen to have a Molex cable lying on the desk, next to me. The
red wire of the Rosewill cable, goes to the yellow wire on the power
supply. So it is connected to the 12V side of the Molex, and not the
5V side.

>
>> There is yet another alternative. You can take a separate fan,
>> and plug it into the CPU fan header, to keep the BIOS happy.
>> You would then need a means to power the real CPU fan, as the header
>> would be occupied by an RPM equipped fan, whose purpose was
>> just to keep the BIOS happy.
>
> I'm using a micro-ATX case, so I don't really have any more room. lol
>
> I will have to go over the measurement just to be sure that I can fit
> the Freezer 64 in there. But then there will be even less room!

A 40x40x10mm fan is pretty small. And I don't really think it is
necessary. Just make sure the Freezer runs at 100% speed, and
you'll be fine. One way to try to do that, is disable any
fan speed control in the BIOS (if one is offered), as then
you get as close to 12V on the fan header as is possible.
But using the adapter cable, and powering the fan from a
Molex drive connector, eliminates all uncertainty about
whether it gets +12V or not.

Paul

> I happen to have a Molex cable lying on the desk, next to me. The
> red wire of the Rosewill cable, goes to the yellow wire on the power
> supply. So it is connected to the 12V side of the Molex, and not the
> 5V side.

Do you think the reviewer of the connector is full of it? Or do you
have a different model of connector that you are looking at?


>One way to try to do that, is disable any
> fan speed control in the BIOS (if one is offered),...

No, unfortunately there are absolutely no open settings for this.

ShadowTek wrote:
>> I happen to have a Molex cable lying on the desk, next to me. The
>> red wire of the Rosewill cable, goes to the yellow wire on the power
>> supply. So it is connected to the 12V side of the Molex, and not the
>> 5V side.
>
> Do you think the reviewer of the connector is full of it? Or do you
> have a different model of connector that you are looking at?
>

You'll notice he didn't buy the product. What I'd suggest to you,
is grab a Molex from your computer and look at it. The plastic
has two 45 degree angles on two corners, that function as a key
and only allow the connector to be plugged in one way. Now,
look at the picture of the Rosewill cable assembly. You should see
the yellow wire of your computer's hard drive cable, would end
up connected to the red wire on the Rosewill cable. And that
means +12V from the computer, is going to flow through the
Rosewill cable assembly.

The reason they didn't install the +5V and ground pin, is there
is nothing for them to do.

There are adapters like the Rosewill one, but with two Molex connectors.
One a male and the other a female. That kind of adapter is
designed for "daisy chaining". And when daisy chaining, all four
pins must be installed, to have continuity for the 5V and GND.

But when the connector is just feeding the fan, the fan only needs
+12V and GND, so the +5V and GND pins don't need to be present
for the thing to work.

What is a bit disconcerting, is the choice of wire colors. The
problem is, the convention for Molex, is yellow is +12V and red
is +5V. The convention on fans, is red is +12V, black is ground,
and yellow is tacho. So when they design an adapter to join
the two, the colors aren't going to match. That is why the red
on the Rosewill, ends up going to yellow on the Molex of your
computer. Thank goodness they all agree that "black is ground".
At least that is common :-)

Paul

> You'll notice he didn't buy the product. What I'd suggest to you,
> is grab a Molex from your computer and look at it. The plastic
> has two 45 degree angles on two corners, that function as a key
> and only allow the connector to be plugged in one way. Now,
> look at the picture of the Rosewill cable assembly. You should see
> the yellow wire of your computer's hard drive cable, would end
> up connected to the red wire on the Rosewill cable. And that
> means +12V from the computer, is going to flow through the
> Rosewill cable assembly.

Yeah, it's just that my PC is buried inside a cabinet and it's an
ordeal to disconnect everything just to open it up and have a look. I
figured since you already had a cable lying around that it would be
easier just to ask.

Thanks