Conclusions

This project built a LED controller system for Cornell’s clocktower using Raspberry Pi Pico W microcontrollers. Our system successfully implemented a modular control architecture based on three types of boards (controller, sensor and face boards) and a custom CAN bus communciation, providing a user-friendly control panel for Chimesmasters to set and adjust lighting behavior of the clocktower.

Our final system achieved all of the core goals we set in the proposal:

  • The CAN bus supported data sending and receiving among boards based on Pico W’s PIO and DMA

  • On the core board, one switch determined LEDs to be turned on or off, one rotary switch determined the LED modes, and three potentiometer sliders controlled the brightness of LED’s R/G/B channels using PWM

  • The face boards responded accurately to R/G/B color updates when sliding the potentiometers on the core board

  • The sensor board monitored and returned sound data to the core board in the SOUND mode, and then achieving the control of lighting brightness by the external sound level

While the system performed well during demonstration, there were a few aspects of our project that could use improvement:

  • We suffered some hardware issues with the shorting of one of the face boards, as well as the unorthodox CAN bus behavior requiring jump starting, as discussed in Hardware Design

  • Our current potentiometer sliders grew strikingly warm during operation, specifically when the tapped output was at its highest. While this could be temporarily fixed by storing a color value when setting, then returning the sliders to the down position, a better solution would be diagnosing why the behavior happens. We were unable to determine the cause (as all measured resistances were normal), indicating that testing with other sliders may be useful

  • While we could communicate with our ambient light sensor, we consistently read values of 0 across the I2C lines, and were therefore unable to implement the light-based mode of the system. This isn’t a regression compared to the current system, but would be nice to figure out and implement in the future

Looking forward, our system provides a foundation for future expansion like WiFi/Bluetooth enables remoting control interfaces, and also the distributed hardware design allows more clock faces or connecting other buildings or devices.

In the end, there were no external standards governing our project. However, we reused Prof. Adams’ CAN bus implementation, and our team’s PWM implementation used in Lab 3.

The only safety concern was with the high-power draw of our LEDs, and whether our boards would be able to tolerate the large current draw. After careful design and testing in lab, our boards were able to withstand this current draw without noticeable temperature increase, indicating that they would be well-suited for the tower LEDs.