DIY Thermostat/PWM Fan Speed Controller

Here is a demonstration of the PWM variable speed fan controller I made. It varies the speed of the fan with temperature. This avoids using a resistor to impede the flow of current to the fan (which consequently results in the resistor burning off some of the current as heat).

This way, the fan doesn't have to run at full speed the entire time it's on. I used the Texas Instruments Hercules RM57L MCU (Launchpad RM57L843 to be exact) development kit (thanks to TI for providing it) to control it via a BJT darling transistor (2n6284).

This PWM technique is a great way to save energy and avoid the thunderous roar that accompanies high performance fans. For example, this is why a computer can be almost silent most of the time, and the fan will only get noisy if you give it a lot of work to do (i.e. playing Fallout 4).

While it's nothing new, it's still slowly making it's way into various industries because it is the quiet, efficient way to go (i.e. inverter air conditioners, inverter refrigerators, inverter washing machines, and your laptop fan is probably PWM already, etc).

Devices in which I rarely (or never) see PWM fan technology applied, but which may utilize it at some point include:

- Automotive radiator and condenser fans.
- 120 volt AC standing and desk fans.
- Attic fans (i've seen one or two which are variable speed, but i'm not sure if they used PWM).

This project will most likely be installed in the plug and play solar system to keep it cool (or for climate control in my vivarium, if I build it soon).

I will upload the RM57L source code for this project shortly. I want everyone to learn how to implement this in their projects, and to see just how easily it can be done with modern technology!

[ShaunDobbie]

That is a strange housing for the pwm controller

[laharlk]

did you compensate for the fan capasitance? some fans have a cap inside to filter some noise from going back into the line. Or are you just assuming a linear response? Btw what frequency did you use for the pwm? (if it's low you shouldnt need to compensate for reactive currents)