• I split the power wire going to the variable speed switch and ran it to a toggle switch. In one position, power would be delivered to the variable speed switch as usual.
• In the other position, the toggle switch would reroute power to the guitar pedal. I connected this lead to one terminal in a two-prong female wall plug. The other terminal of the wall plug ran back to power the variable speed switch, and hence the fan motor.

Initial Attempt

On the guitar pedal side, I then connected the hot lead of the wall dimmer to one lead of a male wall plug, and the neutral lead to the other plug lead. In this way, I could run an extension cord from the fan to the dimmer and control the speed of the fan. Things seemed to be working great. I was able to control the speed of the fan with very good precision through the wall dimmer. In addition, I could toggle to operate the fan without running to the dimmer at all. It was at this point, however, that things began to become problematic.

First of all, the dimmer was larger than the potentiometer that was originally in the guitar pedal, so that the bottom cover of the pedal wouldn't quite fit on top of it. This seemed like a minor problem, since I could simply add some spacers to create some space between the pedal and the bottom cover, allowing room for the dimmer. Secondly, and more problematic, was that the position of the shaft of the dimmer was significantly different than the position of the shaft of the original potentiometer. Thus the mechanism of driving the dimmer from the gear rack would need to be altered. This was something which I was not comfortable in tackling, so I solicited the assistance of my father, who is very adept at mechanical things. He was able to mount the dimmer, and adjust the mechanism of driving it from the gear rack, and things worked pretty well.

However, at this point, another couple problems presented themselves. First, the dimmer would only deliver 100% power when it was rotated completely in one direction. Anything less than resting completely against one of its stops would result in noticeably less than full fan speed. And in fact, I am not sure that the fan was completely at full speed even when the dimmer shaft was completely turned in one direction. This problem became magnified when coupled with the fact that when driving the shaft from the gear rack, it was very difficult to allow the dimmer shaft to completely turn in this direction without putting significant strain on it when the guitar pedal was depressed. The second problem was that when the dimmer shaft was rotated in the other direction, below a certain setting the fan would stop turning. This result was predictable, and not necessarily undesirable, as I had wished for the fan to be able to turn completely off using the guitar pedal. However, before the dimmer clicked completely off, there was a range in which the fan wouldn't be turning, but there was an audible hum coming from its motor. I feared over time this may damage the motor.

Revised Plan

I decided to scrap the idea of using the wall dimmer and start over. I knew that the pedal would work fine to control a potentiometer, as this was its initial configuration, so I decided to use this potentiometer, coupled with a solid state relay and a microcontroller to control the fan speed. I could use a simple voltage divider along with the potentiometer to measure the position of the guitar pedal. Depending on the position of the pedal, I could use pulse width modulation through the relay to control the fan speed. This would solve all of the problems I was having earlier. First, no spacers would be needed to allow room for the wall dimmer, as the potentiometer which was originally used fit comfortably. Second, the shaft of the potentiometer wouldn't need to be rotated completely in one direction. I could just choose a position which was close to completely rotated, and call that 100%. Third, and similar to second, I could choose a position at which the fan was at a very low speed, and cut power to it below this position. These second two things would be possible because I was now using a microcontroller rather than a mechanical dimmer to control fan speed.

Components

Microcontroller

For the microcontroller I chose to use an ATMEGA328P with optiboot from SparkFun Electronics because I had an Arduino R3 Uno from a previous project which uses the same microcontroller. Thus I could program this chip by simply swapping it out in the Arduino and uploading code from my computer. For anyone interested, I originally used tried the ATMEGA328 without optiboot, figuring I could load the bootloader onto it. I fought with this for a very long time with no luck. When the new chip came with optiboot, it worked like a charm. I highly recommend the chip with the bootloader preinstalled. I used a 16MHz external crystal as an external clock.

Relay

For the solid state relay I chose the 8A version, also from SparkFun, as the current drawn by the fan measured at less than an Amp during operation, with a peak of about 1.5A at start up. For anyone choosing to control one of these relays from a microcontroller, be advised that it is necessary to place a current limiting resistor between the microcontroller and the control pin of the relay. When the microcontroller is part of a board like the Arduino, I don't think this resistor is necessary because the Arduino will limit the current going into the relay. However, when the microcontroller stands alone on a breadboard or printed circuit board, there is nothing to limit the current. The unlimited current flowing into the relay can be plenty to fry its internal LED.

Potentiometer

For the potentiometer I used the original one that came with the guitar pedal. It is a 470kΩ logarithmic potentiometer. I measured resistances ranging from about 30kΩ to 490kΩ. I wanted the fan speed to vary linearly with pedal position, rather than logarithmically, so I connected the potentiometer on the lower half of a voltage divider (i.e. between Vout and ground). For the top half of the voltage divider I used 122kΩ resistance by connecting a 100kΩ resistor in series with a 22kΩ resistor. I chose 122kΩ as the value for this as it would represent near the geometric mean of the extreme resistances of the potentiometer, and give a near linear output as the potentiometer resistance varied logarithmically.

Connections

It then occurred to me that in my current configuration, I would need two cords to run to the guitar pedal. One would run from the fan, the other would run from a 120V power source to supply the 9V power transformer. However, I already had 120V input coming from the fan. If I ran a 3 prong extension cord from the fan, I could carry a neutral line, along with the 120V source and the return. This would eliminate the need for two cords to be plugged into the pedal. I tapped into the neutral line from the fan and ran it to what would have been the hot lead in the plug. I attached the hot wire from the fan to what would have been the ground lead in the plug, and attached the return wire from the fan to what would have been neutral in the plug. On the pedal side, I split the hot lead to run to both the relay and to the hot connector on the transformer. I ran the neutral lead to the neutral connector on the transformer, and the return lead to the other terminal on the relay. I chose this configuration because I wanted to ensure that if the pedal were accidentally plugged directly into a wall socket, both inputs to the relay would be grounded, thus avoiding a short and potentially blowing up the relay.

Schematic