Lab 5: Interrupts

The circuitry for this lab was even simpler than the last lab. I was able to connect the motor encoder directly to the MCU GPIO pins. A variable DC voltage +/- Vss was applied at the positive terminal of the motor using a power supply. A diagram is shown below along with a photo of the set up in real life.

I used interrupts in order to measure motor speed quickly and accurately. Interrupts allowed me to overcome limitations of direct polling, as my measurements were taken as soon as encoder signals changed, in comparison to polling. If I used polling, the speed of taking measurements would depend on the speed it would take other commands in my loop to evaluate. I could run into timing issues, potentially resulting in a reported slower-than-expected motor speed. By using interrupts, I ensure that my measurements correspond exactly to when the encoder signals change and that I can make measurements anytime the encoder signals change, even if other code is evaluating.

In order to get the most accurate measurement of speed, I used all edges of the pulse. I set up four interrupts, one for the rising and falling edge for each of the two signals coming from the encoder. The first signal coming from the encoder, "Encoder A", was used to reset counters that kept track of time between edges. This signal was treated as the first edge of our signal. The second signal coming from the encoder, "Encoder B", was treated as the second edge of our signal. It was used to check the counter values since they had been reset and to determine the direction of our motor spinning. The flowchart for the program I implemented is shown below.

I calculated speed of the motor by tracking how high a counter iterated between the edges of the encoder signals. Below is a graphic illustrating how the two encoder signals compare.

Based on this diagram, we can come up with two equations for pulse length in terms of clock periods: \[ l_{pulse, CCW} = \frac{4}{3} * 0.5 * (counter1 + counter2) \] \[ l_{pulse, CW} = 4 * 0.5 * (counter1 + counter2) \] These equations use an average between measuring based on distance between rising edges and distance between falling edges in order to be more responsive to sudden changes. Then, pulse length can be turned into a speed of revolutions per minute based on the specs of the motor, where there are 120 pulses per revolution, and converting pulse lengths into seconds. \[ v_{motor} = \frac{f_{clk}}{120 * l_{pulse}} \]