What's the point of 4-pin CPU fans?

[jrichrds]

All the motherboards I've purchased recently have 4-pin CPU fan headers, and it seems the only CPU fans that have 4 pins are the stock fans from retail-boxed CPUs.

I've read it was for fan speed control, but i've had older motherboards with PWM fan speed control of 3-pin fans. So what really is the purpose of the 4th pin?

 

[Keysplayr]

2 pins are for power and ground. 3rd pin is for rpm monitoring if supported. 4th pin is for speed control if supported.

 

[Denithor]

How come none of the fan producers (Panflo, Nexus, etc) put 4-pin plugs on their fans? I like the idea of using an Ultima-90 but don't like the fact that my motherboard then wouldn't be able to control the fan speed.

 

[Rhoxed]

i have a 3pin mobo and a 4pin Zalman cooler, my mobo still regulates fan speed ect. without the 4th pin

 

[BonzaiDuck]

My motherboard has five fan pinouts that can be regulated thermally: one of them is the CPU_FAN with its 4-pin PWM plug; a second one is a 3-pin CHA_Fan header; the remaining three are also three-pin plugs. The CHA_FAN plug allows regulation according to the user's choice of either the CPU or mobo temperature; the remaining three plugs are paired with 2-pin thermal sensor-pinouts for tape-on thermal sensors.

I don't see it as a significant draw-back that you are pre-empted from controlling the fan connected to three of the four CPU_FAN pins, if that is the case. If you're using a 120mm fan for the Ultima 90, you can simply choose the fan for top-end speed to give you minimum thermal resistance and minimum noise.

But with an hour's tedium, maybe some wire ties and a Dremel, you most likely could adapt the 80 or 92mm stock fan for use with the Ultima. (At one time, I thought these were 92mm fans, but I think they're 80mm.) If you build a trim little foam-board duct-box from the exhaust side of the Ultima to the case-rear exhaust fan, it won't matter if the 80mm fan gives less than optimal coverage of the Ultima's fins.

I, too, decry the availability of third-party 92 and 120mm four-pin "PWM" fans. But as far as the use of "wire-ties" -- it can be done with a certain neatness. And as far as "building duct-boxes," ducts are built by OEM computer manufacturers because they can mass-produce products built on similar motherboards, coolers and cases. So I hold the view that if DIY enthusiasts want to build computers, making little duct-boxes out of foam-art-board and glue is closer to "doing it right" than assuming the case-maker and board-maker have provided you as much in the way of optimal cooling as OEM system builders provide themselves.

 

[VirtualLarry]

Originally posted by: Rhoxed
i have a 3pin mobo and a 4pin Zalman cooler, my mobo still regulates fan speed ect. without the 4th pin

My GA-P35-DS3R has several fan headers, CPU_FAN is a 4-pin (PWM), and SYS_FAN which is a 3-pin. The mobo BIOS has a setting for either PWM or VOLTAGE to control the fans. Apparently, fan control over a 3-pin is possible, by varying the voltage supplied.

My HyperTX2 heatsink only has a 3-pin connector, so I set the BIOS to VOLTAGE.

 

[Cerb]

Long...but I had planned to use some PWM fans, and make a fan controller for them, before the Scythe Slipstreams got out and reviewed, so did my homework, and was nearly ready for a decent Mouser order .

With 3-pin fans, you have
+V
0V
Tach

Now, you can control it by changing voltage, or changing duty cycle (common PWM). Both can have negative affects on RPM monitoring. The circuitry needs a certain voltage to work, though, and your mobo sensors need certain voltages to register tach at. With PWM, you're turning power on and off--RPM monitoring flat don't work right. Tach works by going high for half a turn, low for half a turn...so you take off power, and it will go low. Then, low frequency PWM (most PWM control) can cause noise in the fans, from flexing the frame when it turns on/off. many mobos say voltage, but really mean PWM on the power (but some will do actual voltage, and I don't know about the GA-P35-DS3R, and have no intention to test it). Oh, and if you go too low, voltage or PWM, the fan will not run. Much higher than that, it won't start.

With 4-pin fans, you have:
+V
0V
Tach
Control

You send it +12V all the time, you recieve tach all the time, and a ~25kHz signal goes to the fan to tell it how fast to go, with nice linear response from 30% to 100%. The fan itself is responsible for handling control of the speed, and keeping it running. It's an ideal method for fan control. There is no downside to it. Even if it just amplifies, buffers, and passes the incoming signal, it's super-sonic (so no PWM noise). We're just getting a decent amount of these fans available, though.

 

[BonzaiDuck]

I wouldn't disagree with Cerb at all, but like the OP and Cerb said, the fans are just coming on the market, and choices are still limited. This does give me an incentive to go back and check around again after several months, though.

At least with the Ultima 90, the trick with the stock 80mm fan would work. But I have a TR-Ultra-120 Extreme -- 80mm is just not close enough to the cooler's recommended fan-size.

It sure would be nice if you could find an M1 Panaflo 92mm with the PWM/4-pin setup, though.

 

[BonzaiDuck]

RexFlo 92mm @ Sidewinder Computers

This might just be the thing for that ol' Ultima 90. I suppose I'd like to see the amperage spec at 0.28A instead of 0.22A, but the top-end of 3,000 rpm seems proportionately consistent with the approximately 55 CFM of airflow. The noise level seems acceptable at 35 dBA.

By way of comparison, a ThermalTake Blue LED 92mm has a noise rating in the 40's of dBA, and top-ends around 3,600 rpm and somewhere around 75 CFM. It's also a lot noisier at that speed and output in comparison to the high-end of the RexFlo.

If it were me, I'd get something like the RexFlo for the intake side of the Ultima and STIL build a trapezoidal foam-board duct between the Ultima exhaust side and a rear-exhaust case fan. [Which, hopefully, is reasonably aligned with the cooler so that your grade-school foam-board construction project is reasonably simple.]

 

[jonmcc33]

The same crap with PWM fans has been done for years with 3-pin fans. Your motherboard just needs to support speed control, such as Asus' Qfan, etc.

 

[Cerb]

Originally posted by: jonmcc33
The same crap with PWM fans has been done for years with 3-pin fans. Your motherboard just needs to support speed control, such as Asus' Qfan, etc.

Nope. Different thing.
http://www.formfactors.org/dev...%5C4_Wire_PWM_Spec.pdf

 

[firewolfsm]

Originally posted by: Rhoxed
i have a 3pin mobo and a 4pin Zalman cooler, my mobo still regulates fan speed ect. without the 4th pin

That's because the controller is on the fan, for the motherboard to control it, you need the fourth pin.