| Home | Job | Pinball | Photo Album | Automotive | Press/Awards | Contact |
My No-Hum Fan Speed Controller
Introduction
Thanks to David N. and Doug Smith for their help in getting me started on this project.The commercially available X-10 compatible fan speed controller is the 6381-U from Leviton. Unfortunately, this switch causes the fan motor to hum loudly. When I installed one in my bedroom, I found it nearly impossible to sleep with the fan on. This was very annoying since I have very tall ceilings in my bedroom, and a lot of warm air is trapped up high in the winter, while we are chilly a ground level. The room gets noticeable warmer when I turn on the fan. After receiving a copy of a circuit from one of the gentlemen mentioned above, I decided to build my own version.
Ceiling fans that are controlled by the switch like Leviton's buzz due to the PWM action of the wall switch's TRIAC. On the other hand, cheaper manual fan speed controllers that feature discrete fan speeds allow the fan to operate completely silently by switching capacitors in series with the fan. The project described here combines the two units. An X-10 controllable fan speed switch that allows the fan to operate silently by switching in capacitors.
Giving credit where it is due, I used the same approach for controlling the fan (capacitors in series with solid state relays) as a circuit posted on AOL, they however used a series of opamps and a discrete voltage divider. When I saw the schematic I remembered from my bag 'o tricks that there is a chip to do this job, the LM3914. It can even drive loads directly. As a result the circuit is more compact. I also used a real power transformer, instead of the resistor/capacitor type power supply of the AOL circuit.
Implementation
I used an 'old work' box as the enclosure, screwing a four terminal barrier strip onto the back. The whole unit can then be slid out of the wall, and disconnected for service or assembly. It can also easily be mounted back into its intended location. The box is then covered by a blank plate with a few holes drilled so the LEDs can stick out.The main challenge in construction is fitting the whole thing into a wall box. I used two circuit boards, one to hold the capacitors, solid state relays, and the power transformer. The other holds the LEDs, the IC and the dummy load for the inductive module. Each board stands vertically on the left and right side of the box, and the dummy load resistor is put in the back where holes have been punched for air flow. The power resistor should be rated for 10Watts, it dissipates less than 1.6 Watts in the circuit.
There is only one adjustment in the circuit. With the inductive switch fully on, adjust it until the hi-speed led indicator lights up. With the circuit in use, and an ON command is received, the leds will ramp in order red-yellow- green to full speed. An off command will also cause it to ramp down. Now you can send bright / dim commands to change the speed of the fan, and the LEDs will ramp up and down accordingly. Had I used a microprocessor, I could add a chime to click everytime a change of state occurred, but I resisted the temptation to overdo this job.
When you build this thing, I urge you to put together a project board like I did. Using a 2' by 1' piece of plywood, I put an GFCI outlet on it, with a mini-controller and a circuit bread board. Since you will be working at line voltage, you want to be able to unplug the whole thing when you touch it. The GFCI will help protect you from shocks. As you test and assemble, you can pick up the board to your ceiling fan for tests. Since there is a mini controller right on the board, you don't need to walk across the room for an X-10 transmitter. I would also add a fuse, something I did not initially do, causing me to ka-blow the Leviton switch during development (I repaired it).
wall switch on the left (along with the light switch).
Correspondence
>From: rod@mail.delnet.com (rod)>Ed. Did you modify your inductive load wall switch with a different
>TRIAC.
The design of the inductive wall switch allows the use of a 9k resistor as the dummy load. Power dissipation is low enough not to be a concern.
Long term update
The unit was installed in 7/96. After more than eight months of usage, especially thru a winter, I am extremely pleased to have built this unit. The fan operates completely silently, and we are saving energy. We can control the fan speed from bedside, and it is probably one of the best uses of X-10 in the home (3/97).September 2000 - I replaced the inductive dimmer module with a regular wall switch modified to add a neutral wire. The Leviton unit stopped working due to an ESD incident. The No-Hum Controller has performed flawlessly for these five years, and is used every night.
October 2001 - Kit and printed circuit board made available of the Fan Controller.
The value of the zener is 5.1Volts, and its cathode is the '+' terminal of the
power supply. The adjustment pot is 200k.
Video of one unit on the repair bench showing its operation (6/2012)
(c) Edward Cheung, 2013