Noob Question. PWM Fan Control.

After browsing the forums, I came across a way to bump up the PWM frequency to 32Khz on Pin 10.

I am working a on project where I need to control a pretty high power cooling fan (12VDC, 1.8A) for a computer, but have run out of acceptable commercial products (cost effectiveness + high switching rate).

Having a high PWM frequency is important because the goal is having low noise. I'd like to get it to far end or beyond the human hearing spectrum.

Anyway, instead of using a motor controller IC, I wanted to use something like this:

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=IPP147N03LGIN-ND

A logic-level N channel MOSFET.

I had tried transistors before, but they seem to not want to power the fan at such a high frequency.

Would this part be adequate for my needs?

Yes that looks OK.

However,
PWM is not a recommended way of driving a fan for reliability and long life of the fan. The best bet is if you take the PWM signal, put it through a transistor to get a 12V signal and then smooth it with a large capacitor. Then run the output of that into the gate of a FET with the drain connected to +12 and the source connected to the fan (other end of the fan to ground). In that way you have a source follower, there is no noise from the PWM and the fan is driven in a much better way.

I am having a little troubling understanding how to implement that correctly.

If I hook up the transistor, I will get 12v. I understand that, but it that strictly necessary considering that its a logic level MOSFET?

If I hook up a capacitor to smooth out the voltage of the transistor output, won't that defeat the purpose of throttling the fan using PWM?

My limited knowledge of electronics prevents me from understanding completely.

I am not too worried about the fan. It is a Cold War Era Papst Fan (Made in West Germany!). The fan is completely metal (metal frame and blades) and the bearing in it feels like it can work was a lazy susan. I have 4 more sitting around and I bought them for about 4$ each.

I understand that, but it that strictly necessary considering that its a logic level MOSFET?

Yes you get an analogue voltage at the FET gate and it works in source follower mode.

won't that defeat the purpose of throttling the fan using PWM?

No you will get a variable DC voltage which will nicely control the fan speed.

I did a bit of research into how MOSFETs work, and figured out that they work (pretty much) as resistive current limiters when their gate voltage is low.

I dug around my parts pin and found the only MOSFET I could, an IRL1404, which can handle a very fair amount of current (over 100A continuous).

I tied a potentiometer between +12 and 0 and the center pin into the MOSFET gate and calibrated it until the fan was throttled to the level I wanted.

The MOSFET got HOT FAST.

It probably 4 seconds before I could smell the chip (and it was heatsinked to an old northbridge cooler.

On top of that, the fan kickstarter kept firing and caused the fan to vibrate on the table.

I tried pumping in a 500hz PWM from the Arduino and it ran cool to the touch. I couldn't feel the kickstarter either.

The problem is there is a buzzing sound because the PWM is in the audible range. Its not that bad on these fans, because they have so much rotational inertia, but its still annoying.

I tried pushing the frequency up to 8khz, but the fan just went full blast at that point.

I think the capacitance in the gate was too high.

Is there any way to figure out which MOSFETs are suitable for high frequency switching?

The MOSFET got HOT FAST.

Yes it will, it is dissipating a lot of heat because you are in the liner region of control.

the fan kickstarter kept firing

You didn't mention that the fan had one of these. Find out what triggers it. With the circuit I gave you are actually giving the fan a variable DC voltage so it could be being triggered by what it thinks is a low voltage.

I tried pumping in a 500hz PWM from the Arduino and it ran cool to the touch.

Again no surprise because the FET is either on or off, when it is off there is voltage across it but no current through it. When it is on there is lots of current flowing but not very much voltage across it. Power is voltage multiplied by current.

I couldn't feel the kickstarter either.

It probably kicks in at a certain speed or voltage so it wouldn't when running PWM.

I tried pushing the frequency up to 8khz, but the fan just went full blast at that point.

PWM only controls the speed for a loaded motor, it looks like the motor is going full belt and the mechanical inertia is keeping it going during the small off time. I would be very surprised if your FET was not able to switch at 8KHz.

Is there any way to figure out which MOSFETs are suitable for high frequency switching?

8KHz is not a high frequency, 300 to 500 KHz is high for a FET. Look at the spec sheet for the device and see the gate capacitance and switching speed parameters.