Carnegie Mellon ECE Capstone, Spring 2024: Twain Byrnes, Gary Bailey, Carson Swoveland
This week, we had some major successes with prototyping the motor-based energy harvesting scheme and finishing the breadboard prototype of the digital hardware.
We have no new risks that we are worried about, nor are there any changes we have made to the design of the system.
My primary accomplishment this week was designing the gearbox that will be used for the motor-based energy harvesting scheme. The gear ratio and general structure was based on the design by Cai, Liao, et. al, but I had to determine the exact dimensions, optimal gear profile, mounting strategy, etc. myself.
I printed using the “Extra Fine” slicer settings and standard print speed on the Robotics Club Bambu X1 Carbon printer in PLA. The result worked astonishingly well, spinning smoothly enough to keep going off of inertia alone for about a second. Even with the small amount of weight provided by three M6 nuts and minimal balancing, the gearbox would spin in my hand as I swung my arm. We were unfortunately unable to attach the motor for real power measurements, as we discovered that we needed M1.6 size screws, but even so this is very promising.
Working on the gearbox put me somewhat behind on the PCB design; I will need to work on that some more next week. Fortunately, the gearbox testing is doing much better than expected, so I should be able to catch up.
Next week, I will:
This week, I created a prototype for the piezoelectric energy harvesting. I cut one of the pieces of spring steel into one inch strips, prepped the surface, and attached three of the piezo tiles to the surface at locations suggested by the paper, leaving room to attach other tiles at the angles suggested to have more voltage output upon flexion. I then hooked each tile up to an oscilloscope and studied the outputs upon flexing the spring steel small amounts at the bend angles that would be likely to be seen when wearing the watch. Additionally, I began to figure out how to attach the watch straps together so that they can be reformed for our purposes.
My progress is on schedule, as this week was dedicated to creating a prototype of the energy harvesting.
In the next week, I hope to test this prototype more and to help Carson implement the other form of energy harvesting that we are looking into.
This week I finalized the barebones breadboard prototype of the GPS watch: a lot of the time was spent on debugging the UART failing to connect (the issue was that the baud rate I configured the GPS to had been reset, as the GPS doesn’t actually have persistent settings). In the end, we have a GPS setup that’s printing the coordinates out to the screen!
I’m about on pace for my tasks: by the end of this week I was scheduled to be done with the firmware needed to get GPS data on the screen, which is what I’ve done.
Next week I plan to work on more of the firmware functionality, mainly the timekeeping via the RTC. I’ll also help Carson and Twain iterate on the energy harvesting after the success we had this week with the motor-based design.
This week, I made more progress toward writing the boilerplate firmware needed to get the microcontroller to communicate with our various peripherals. This includes another pass at the SPI transaction handling for the display, and a preliminary serial setup for receiving GPS data.
Here’s a screenshot of my newly rewritten function for efficiently updating the display by sending only changed lines:
I’m a bit behind schedule in terms of producing our actual prototype board, but I have at least shown the fact that the various hardware we’ve been planning to integrate will work together, which was needed in order for Carson to start designing the board schematic.
Next week our GPS module should arrive, so I can actually properly do GPS integration rather than sending test data. After that point, I should finally be done with our “breadboard prototype” step, at least in terms of firmware.
My primary accomplishment this week was in making a first (very rough!) draft of the board schematic. It is missing a few connectors and features useful for testing (jumpers, test points, spare pins, etc.), but it’s very useful for checking if there’s any major design details we’ve overlooked. Thankfully, there weren’t any.
I am roughly on pace right now.
By next week, our components should have arrived. This means I can start testing the energy harvesting setup and experiment to find the optimal setup, and then get real numbers on how much power we can source.
The biggest risk we see right now is the timing on the shipping of necessary components. We have accounted for this as part of our general shipping time / slack in our Gantt chart, but we will mitigate similar future risks by attempting to figure out what we need earlier, and by utilizing Slack to communicate these needs earlier.
We did not make any specific changes to the existing, but we decided to make a design decision on our battery system: we will move forward with a LiPo battery. The only costs from this are in purchasing one such battery.
There have been no changes to the schedule.
For this status report, we were asked to answer additional questions. A was written by Twain, B was written by Gary, and C was written by Carson.
Part A: Our product solution addresses a critical need in the context of public health, safety, and welfare. The prolonged battery longevity ensures that individuals, especially those engaged in outdoor activities or requiring continuous location tracking, can rely on the device for an extended duration without the concern of sudden power depletion. Simultaneously, the watch’s emphasis on user privacy safeguards individuals from unauthorized access to their location data, mitigating potential risks associated with personal information exposure and contributing to the psychological well-being of the user. By prioritizing both functionality and privacy, this GPS watch contributes to the overall well-being of users, aligning with the broader goals of public health and welfare by providing a secure and dependable tool for personal tracking without compromising privacy or safety.
Part B: The most prominent existing solutions to the issue of a user wanting to keep track of their own location history involve sending the user’s location data to a companies who have stipulated in their privacy agreements that they are free to sell the data to others as they wish. To many people, this is an uncomfortable arrangement, as over time, more and more fears surrounding “big data” and data-selling arrangements like this have become prevalent. Our watch addresses this issue, as it allows the user to track their path while all of the data involved is stored on the watch itself, without ever being uploaded to the cloud.
Part C: Current watches that integrate both GPS and energy harvesting technology are very expensive (e.g. the Garmin Instinct II, which sells for ~$400). Other watches that have GPS technology (e.g. the Google Pixel Watch, at ~$200) don’t integrate energy harvesting, and thus have limited battery life as a result. Landhopper’s feature set is specifically designed to fill the niche of a wearable but long-lasting GPS, while trimming out features like Bluetooth and phone connectivity to avoid expensive transceiver modules and additional processing requirements.
This week, I spent significant time working on the slideshow presentation for the design presentation. I also helped create the list of parts we needed to order, and then submitted all of the ordering paperwork and dealt with the issues that arose in the order. Additionally, I helped Gary with the firmware for the GPS, looking at Rust crates for parsing NMEA 0183.
My progress is on schedule, assuming that the parts needed to test the energy harvesting show up quickly enough for me to get to prototyping, which is my main task for next week. I will need to figure out how to cut the metal, make it safe to wear, and attach it to the wristwatch strap and the piezo tiles to the metal.
Earlier this week, I worked with my teammates to review possible paths to follow for our energy harvesting design, and also worked on composing a purchasing list for our initial prototypes.
Later in the week, I started working on the breadboard prototype for the watch, along with the firmware: most of this time (around 9 hours) has gone into setting up our build environment and debugging the peripherals we need to use in the firmware. So far the display is implemented, and I’m also currently working on the UART for the GPS and the real-time clock. I am a little bit behind schedule as we had originally planned for me to finish with the breadboard prototype this week, but we’d failed to account for when we’d be able to order parts. However, I don’t think this should be an issue since I’ve been using my time to get ahead on the firmware (my task for next week), and progress should be faster on said firmware with Twain helping next week. Once the remaining parts we need for the breadboard prototype arrive (mainly the GPS), it should be pretty quick to finish the prototype (the wiring is pretty much just a few UART lines) so we can start testing the power systems.
The most significant risks we see right now are with the energy harvesting and its interaction with the GPS. The energy harvesting method we intend to employ may not be as efficient as we were hoping, since we were unable to find the correct kind of piezo tile, and generally we would not expect to get the same results as the paper. We are managing this risk by looking into alternative sources of energy harvesting. The other risk that we see right now is that GPS is a very finicky system, and we are worried that the voltage jumps from the piezo tiles will create a large amount of noise that may throw off the GPS signal. This is also being mitigated by looking into other potential sources of energy harvesting.
No set changes have been made to the system as of right now.