That would be news to all my users and the rest if the world who are all doing this. I'm currently controlling 16 fans via my MoBo fan headers on my own box. I'd be interested in your explanation on how my MoBo's 3 fan headers are controlling the 16 fans between 326 and 850 rpm on three separate channels if I'm incorrect. Also the 5 other builds here and all the other builds we do each month.
The only way to actually control the rotational velocity of a fan that does not have PWM control (3 pin chassis fans) is to raise and lower the supply voltage to the DC motor.
Exactly ..... and that's exactly what the 3 pin fan headers do. The three pin fan headers exist on the motherboard. The motherboard BIOS or companion utility (i.e Asus fan Xpert2) controls the fans by controlling the voltage on the header. It's one thing quoting a source, another understanding it....especially when the source is outdated and predates currently available technology.
The difference between PWM and DC voltage control in today's world can best be described by those self propelled spinning merry-go-rounds found in children's playgrounds:
DC Variable voltage control is comparable to Dad sticking the 3 kids on the playground thing and applying pressure to every bar that goes by, no stopping, grabbing every bar that goes by and pushing to keep the thing moving at a certain speed....to spin faster, Dad pushes harder.... to spin slower Dad pushes with less effort.
PWM control is comparable to Dad sticking the 3 kids on the playground thing and always pushing the bar as hard as he can. The way dad controls peed is to vary the time between each push. To go faster he decreases the interval between pushes.... the go slower he increases the time interval between pushes.
Years back when DC variable voltage control was less sophisticated, one problem with this approach was that when called upon to run at slow speeds, the low voltage applied was not enough to overcome the inertia of a non moving fan. Today, when implementing currently available technology, that problem is extinct. Today's variable voltage BIOS technology will test each fan in your system and determine stall speed, startup speed and prepare a list of data points from 0 to full rpm and from 0 to 12 volts. If it wants to run at 330 rpm but knows it won't start below 360 rpm It will start at a higher rpm and then ramp down.
The trouble with controlling fan speed via DC control is that lowering the DC supply lowers the torque applied to the rotor, lower it too much and it may stall out. If a DC motor stalls out it still draws current (the most current actually) and will eventually overheat and fail. The DC response of each motor is unique and all that we know is that they are supposed to rotate at a nominal RPM when supplied with 12 volts, we do not necessarily know how they will behave when supplied with less.
Interesting perhaps as a historical reference but doesn't apply today. This doesn't happen anymore if you are using currently available technology. Because of the testing that was done by the MoBo utility, it knows exactly where on the curve the stall point is. The utility will therefore create a setpoint just above the stall point, so that can **never** happen. And yes we do now exactly how they will be behave at each point on the curve because the test determined each of those points exactly.
Here's how I am controlling my box:
Channel 1
CPU_1 PWM Header => Water Pump 1
CPU_2 PWM Header => Water Pump 2 (PWM)
Channel 2
CHA_1 DC Variable Voltage Header => PCB-1 => (6) 140mm 1200 rpm Case Fans
Channel 3
CHA_2 DC Variable Voltage Header => PCB-2 => (6) 140mm 1200 rpm Radiator Fans on XT45-420
Channel 4
CHA_3 DC Variable Voltage Header => PCB-3 => (4) 140mm 1200 rpm Radiator Fans on UT60-280
1. Setup begins by running the AutoFan Tuning Mode. It runs the fan from 100% voltage and drops it down in 10% increments and records the actual rpm.. During this it determines min. speed, stall speed, startup speed and correlate each fan's rpm with % of full voltage
2. You then set the individual control points for each to determine at what temps you want to change speed..... utility will automatically ramp the speeds between the set control points.
3. The lowest set point is above the stall speed determined by the test. So the "stall scenario" you described is now impossible .
4. You have the option of running at the minimum speed above stall or turning the fan off at this temperature level.
5. You can also set the rate at which the fan responds to temperature changes from steep to gradual. This settings reduces sound as well as allows the fan to run at more stable speeds rather than ramping up and down chasing variable loads.
On my rig, the tuning mode determined the maximum fan speed is actually about 1238 rpm, tho I never see that speed. I can maintain an a < 10C Delta T under all conditions (outside of synthetic benchmarks) with a max speed of 850 rpm. The one thing I lose using 3 pin fans instead of 4 pin PWM is that the PWM fans will go down to 300 rpm but the tuning utility set the min. speed at 326 rpm. I am tempted to have the utility turn them off at 350 or 400 rpm when the puter is idle but since they are dead silent over the 326 - 850 rpm speed range, I haven't bothered.
Of course the caveat must be stated that if ya bought a $89 MoBo, it's not going to come with this technology. Speedfan may work for ya and you still can always just set the minimum fan speed in the BIOS so you never stall but the auto tuning and fine adjustments won't be available..