The Hall effect sensor is mounted to the frame so that a magnet on the pedals passes in close proximity once per revolution. the wooden stick was necessary for stability and to increase the surface area in contact with the bike. Epoxy was used to hold it all together.

Electrical Design

Information about the cadence of the rider (pedaling RPM) and whether or not the rider is currently pedaling are provided as sensory inputs to the PC104 stack. These physical inputs are captured by a Panasonic DN6848 Hall IC. The Hall element in the IC is activated by a magnetic field. The field is provided by a magnet mounted to the right pedal. The magnet we selected is a powerful rare-earth magnet to help ensure that the circuit switches its output each time the magnet passes the Hall IC.

As shown in the block diagram on the IC data sheet, the Hall element's output is amplified and used as base current to an open collector output stage. The IC is powered by +V and GND and the output is connected to +V by a pull-up resistor. The resulting signal is fed to a digital input on the PC104 stack.

The system uses feedback from the motor to determine how far to move the derailleur cable in order to shift gears. When first analyzing the various types of feedback mechanisms, we considered both potentiometers and Hall effect sensors. A problem we encountered with Hall effect sensors is that we would have to build and attach another part to the bike in order for the magnet to be mounted properly. We had also considered using an encoder with our motor, but determined that purchasing one would be too expensive and not offer a significant benefit over the potentiometer. Therefore, we came to the conclusion that a potentiometer would be a more practical alternative.

Typical potentiometers located in the lab were very bulky and required that we machine additional parts to adapt our design to work with one. Instead, we needed a potentiometer that would slide over the shaft of the motor and turn with the spindle. Various wholesalers did not seem to have that type of potentiometer in the necessary dimensions, so we proceeded to make one. We disassembled a spare potentiometer and drilled it out so it could be mounted to the motor shaft. The result is a flush design that was press-fitted around the motor shaft.

When powered, the potentiometer changes resistance proportional to the angle of the wiper (attached to the motor shaft). The corresponding voltage drop in the circuit can be used by the PC104 stack for motor feedback.

We also needed to create a circuit to power the motor, as the analog output from the PC104 stack is not capable of providing the current necessary to drive the motor. We determined that we only need to provide three states to control the motor: clockwise, counterclockwise, and zero. The motor is nonbackdriveable, so when current is removed, the motor comes to rest almost immediately due to the high gear ratio. This prevented us from having to design a complicated motor control system and waste power holding the motor at rest. The circuit itself consists of an op amp to provide the necessary current to drive the motor.

Because of the voltage and current requirements of our selected motor, using an LM348 op amp IC would not be able to deliver enough current to drive the motor. We explored several design alternatives, including using relay circuits and a power supply, before settling on our implementation. We used a power op amp (LM1875) that was capable of providing 20 W. We had to power it with +/- 15 volts from an external supply because the breadboard power supply could only deliver +/- 12 V.

Both the Hall effect sensor and the potentiometer leads were soldered to wires that ran along the bicycle frame to the breadboard. Each solder joint was covered with heat shrink for further protection and to ensure that adjacent leads didn't short. The Hall sensor was mounted to the frame with epoxy and the leads for both the sensor and the potentiometer were held in place with nylon zip ties.