Controlling fans using PWM

[z3bb0]

So yeah I'm quite new to electronics and would love some help 

The project I'm working on is going to control the speed of 12VDC fans and since this might be sold as a project it needs to be cheap.
Right now there's a few things I need to know, how much will my darlington transistor(TO-126 NPN ) heat up during usage?( duty cycle 40-100%, 1A constant current, 12V input.) Will it need a heatsink?

Another thing that's not related to the switching transistor, I need to protect the circuit from overloading it. I was thinking about one of those self resetable fuses since it can't be a normal fuse since it's going to be sold.

Thank you for your time

[Delta2Wye]

Hi z3bb0,

To accurately determine if a heat sink is required, you must calculate the thermal impedance of your device and estimate the power dissipated. You can read about thermal resistance here:  http://en.wikipedia.org/wiki/Thermal_resistance. Dave did an entire episode on this exact material, EEVBlog #105 if I recall correctly.

Assuming your at a decently low switching frequency, <100 kHz, my gut tells me your going to be fine without a heat sink.

Lastly, I would recommend use a NMOS instead of a darlington pair as they are much easier to trigger that BJTs; although at your current level I don't think it will make a huge difference.

[z3bb0]

Thanks mate I'll take a look at this as soon as I get home

[T4P]

Do you want a simple 555 timer operated PWM DC motor speed controller? It works fine up to 20W (Didn't test it) without heating much on a BJT

[shebu18]

http://www.eeweb.com/project/circuit_projects/pwm-fan-controller-by-lm2902n-or-lm324n
This is a solution.

[z3bb0]

well no, I'll be using a MCU for PWM and this is my own design since I want to try to sell it.

[metalphreak]

http://www.pcbheaven.com/circuitpages/PIC_3-Wire_Fan_RPM_Controller/

Reading this is a good start. He has a few different projects on fan control and I like his style of writing

If you want to use the RPM feedback on 3pin fans you need to do pulse stretching where you apply a constant voltage for a short period while you read the speed. This is because the tach wire is powered by the same voltage as the fan motor.

4 pin fans don't have this problem, as you feed the PWM signal directly into the fan which has its own mosfet on the control board for the fan motor only.

If you use a mosfet you really don't have to worry about heat dissipation as you can run quite a low PWM frequency for fan control. Even a 3000rpm fan is only spinning at 50 rotations per second.

Many commercial fan controllers use linear regulation rather than PWM specifically so you can still use the tachometer wire for reading RPMs on motherboards etc (which have no idea when you are or are no "pulse stretching").

As always, Application Notes from the major IC vendors tend to be a treasure trove of information and theory!

http://www.maximintegrated.com/app-notes/index.mvp/id/1784

[Jeff1946]

I also vote for an NMOS.  If you select the right size resistor to connect the gate to your drive circuit you can slow down turn off and turn on.  This gives a slight power penalty but will cut down on electrical noise generation.