Who Knows What To Do About Four-Pin CPU_FAN Headers?

[BonzaiDuck]

I'm carefully preparing an ASUS Striker Extreme board for initial configuration and installation in a test case.

I see that the CPU-FAN header has four pins: A "PWM" pin, a "Fan In" pin, a 12V pin and a ground.

I'll be using a different heatsink than the one which came with my Conroe E6600 processor, and I would want to control a 120mm fan from that motherboard plug.

What needs to be done -- if possible -- to accommodate a fan with tach-monitoriing to that four-pin fan headers?

 

[Ken90630]

I'm not familiar with the Asus board you're gonna use, but maybe something like this would do the trick?

Not to insult your intelligence, but it will enable your CPU fan to get its power from a Molex connector connected to your PSU, and you would retain your RPM monitoring capability by plugging the 1-wire connector onto your CPU fan header.

Of course, there may also be some 4-pin 120mm fans out there that would work by themselves without an adapter -- I haven't kept up with this stuff for awhile though, so I dunno.

EDIT: Before you try that adapter I linked to, see my next post below.

 

[natto fire]

Just hook the rest up except for the PWM pin.

As described here

And running 120 mm fans off the mobo is something that is still debated to this day. I personally think it will be fine, but you mght want to get one of those things Ken is talking about.

 

[Zepper]

You don' need no steenking mods! If you just looked closely at the connectors you would see that there is only one way that a standard 3-pin female fan connector will fit on that 4-pin fan header without forcing. And it's designed to do just that... I'm having difficulty witholding the D-word.

.bh.

 

[Binky]

There is only one 120mm PWM fan out so far.

http://www.performance-pcs.com/catalog/...ain_page=product_info&products_id=3977

Arctic Cooling has announced a series of PWM fans but I haven't seen them for sale yet.

 

[Ken90630]

Originally posted by: Binky
There is only one 120mm PWM fan out so far.

http://www.performance-pcs.com/catalog/...ain_page=product_info&products_id=3977

Arctic Cooling has announced a series of PWM fans but I haven't seen them for sale yet.

Well, my earlier idea about using that adapter cable might not work with your board. As I said, I haven't kept up with this stuff and these PWM (pulse width modulation) fans are new to me. That cable might work, in terms of enabling the fan to run, but maybe just wouldn't give you the PWM "functionality" that would make it run as quietly as it's designed to. Or maybe it would 'confuse' your mobo and not work at all, since the board would be looking for the PWM signal feedback from the fan itself. I dunno.

(I'm assuming that PWM is a 'new' technology and not just the same as rpm monitoring. If I'm wrong, someone please correct me.)

I'd prolly put in a quick call to Asus tech support if I were in your shoes and just ask what they think. Or just get one of those fans Binky linked you to.

 

[Zepper]

PWM fans have the PWM (regulates the power, not the voltage, to the fan by varying the pulse width of the voltage to the fan) chip right in them. The chip is controlled via that 4th line in response to the CPU temp. Some mobos will still control 3-pin fans connected to the CPU fan header too in a one or two-step method like many did before. The 4-pin CPU fan header will take either type of fan with no problem just as I said earler. And as I also stated above, there is only one way that a 3-pin fan connector will go on the 4-pin header without forcing - so it is very hard to get it wrong.

.bh.

 

[Ken90630]

Originally posted by: Zepper
PWM fans have the PWM (regulates the power, not the voltage, to the fan by varying the pulse width of the voltage to the fan) chip right in them. The chip is controlled via that 4th line in response to the CPU temp. Some mobos will still control 3-pin fans connected to the CPU fan header too in a one or two-step method like many did before. The 4-pin CPU fan header will take either type of fan with no problem just as I said earler. And as I also stated above, there is only one way that a 3-pin fan connector will go on the 4-pin header without forcing - so it is very hard to get it wrong.

.bh.

Thanks for the education. :beer:

I wish there were more hours in the day so I could keep up with all this stuff ....

 

[BonzaiDuck]

Thanks for your confirmations, folks.

I pushed this question here and in an e-mail to a tech-support guy for a specialist-reseller who's always willing to help. My caution relates to minor concerns about tach-sensing and motherboard "requirements" for running the stock cooler-fan off that plug.

On the 120mm fan angle, I would think most mobo manufacturers gravitate toward similar specs. With this ASUS board, the combined amperage across all mobo fan-headers is about 2A, or 24W total power consumption. Even a Delta Tri-Blade 120x38mm runs just above 0.80A. There will be five fans in the system excluding the proprietary fan on the VGA (which will soon be replaced by a heatpipe cooler and a less noisy fan): so, 0.80A + 4 x (0.30A) = 2A.

If I could trim the fan total to four, it would obviously be better, but this is a trade-off with the noise issue. For this, one intake will be on the case-bottom, another at the lowest point in the case-front, and a third behind the HD cage and at center of case interior.

The Tri-Blade, because it is fairly quiet in the low and mid-range, would be "variable" -- installed in case-rear for exhaust -- and could eliminate the need for two 120mm exhaust fans. The other fans I'd like to trim to a permanent 3/4 of their speeds -- no higher than 1,200 rpm each. I may decide to use a less powerful version or alternative to the Tri-Blade.

Here, choosing how close to come to the motherboard power-draw limit for fans may mean the choice between an auxiliary "rheostat" fan controller and not having one.

I've typically used the mobo fan-headers for up to 70% of the cumulative limit, without problems.

Your most humble and disparaged dummy --

Bonzai
[quack, quack]

 

[zagood]

With that many fans, my first thought would be to offload most of them to a rheobus or other fan controller. You'll be much happier with being able to change fan speeds on the fly. Also, it's usually better to get a higher RPM fan than you think you need and lower the speed on it. Better to have more CFM than you need than not enough.

You could use SpeedFan to try and control them through the motherboard, but I've yet to have a motherboard that SpeedFan works with on all sensors without some frustrating tweaking.

-z

 

[BonzaiDuck]

Through the development of things like "sleep" and "hibernate" capabilities, we've suffered through having mobo fan-control that doesn't extend to all fan headers.

I have guarded hopes for the latest generation of mobos.

The Striker Extreme manual [snicker at the model name] suggests all fan headers can be controlled by temperature through BIOS settings. We shall see . . .

At a minimum, I'll want to control two fans; at bare minimum -- I'll need to control one. But with the number of fan-headers on this board, I should be able to set some intake fans at a constant around 70 to 80% of their full speed, and vary the one or two fans pulling and pushing through the CPU HSF fins.

My guess -- just a guess -- is that those two fans won't ever need to run above 2,400 rpm, and a lower maximum may even be optimal.

ZAGOOD's note about finding fans with a wider range of speeds is well taken, if you can be satisfied that their dBA ratings at low and midrange are competitive with so-called "quiet" fans.

This "Conroe" build resurrects a 1994 Compaq ProLiant business-server case. At 8.5" x 26"(H) x 24"(D), and with 3" double-wheeled casters with "brakes," there is ample room on case floor for intake fans and a minimum of noise. I use a pressurized imbalance between exhaust and intake, with ducting past mobo and card components. Some of the pressure imbalance needs to remain for benefit of PSU cooling through its own 120mm fan.

I'll keep a photo record of my progress on the case, with some hope I can present it here later. I have much sheet-metal cutting and pop-riveting to do . . .

 

[zagood]

They couldn't get fatal!7y so they went with a porn star branded board. Extreme for sure!

The board says it can run the fans based on temps? Sounds like speedfan, couldn't have been too hard for a mobo manufacturer to do it in bios, surprised it took so long. That's cool. May as well start with that, then expand if you need to. Shouldn't have to worry about the fan headers if you're going for mid-low speed. Just look out for 120x38mm fans, some of them can suck over .5 amps.

Oh. Duh. Strike that last sentence.

Grab a fanmate2 or some other cheap inline resistor for the 5th fan if you want all of them on some sort of control.

Sounds like you've got a fun build on your hands, would love to see the diary.

-z