Carnegie Mellon ECE Capstone, Spring 2026 – Andrei Da Silva, Sam Istvan, Ben Berger
Our biggest risk right now is pieces breaking before or during demos, so we have continued testing each subsystem a lot to work out any bugs. We have also prepared back up motors and servos incase a critical part breaks in the next day to make sure we have something working, as well as collecting video of working parts as an extra backup.
No changes were made to the system this week, and our schedule remains the same.
This week I worked a lot on making finalize subsystems of the chip sorter. The two major accomplishments are full 3d printed and assembled chip hopper and and working color sensing. Both components are functional and are able to sort chips by color. Current work is focused on building the conveyor between the hopper under the table and the color sensing so that it is fully working for the demo.
Progress remains slightly behind the original schedule because of the 3d printing failure but the color sensing has been a big leap back to being on time.
Below are the working hopper with chips being passed through it and the color sensing with the arduinos to release chips into their proper storage.
This week I spent a lot of time working at TechSpark, where I worked on understanding the Raspberry Pi and how to connect it with the RFID, and working on using the playing cards with the RFID to give the central controller knowledge of the players’ cards to be able to run the game.
I had an issue with the Raspberry Pi that was given to me, which had an issue with the SD card reader, which heavily slowed down my progress this week.
The goal is to hopefully integrate the RFID code with the AI code this coming week, right before booth keeps us awake for days on end.
I have also been trying to dismantle the auto card dealer with no real progress, so I believe I will be pivoting to a different solution, which would require me take apart the start and stop buttons and connect those to the raspberry pi and that will be my goal in the next couple of days.
All infrastructure, hardware interfaces, bindings, and AI are done. AI is mostly trained, still training for the hardest difficulty. I have done significant debugging, as well as implementing more hardware integration interfaces and the entire graphics module for displaying game state on the screen.
I am working on designing the APIs to correctly target the other subsystems for easier integration. In addition, I am trying to optimize training since it is progressing quite slowly, and developing an easy way to checkpoint the system, since I currently do not have a way to save model weights and modularly swap them out for a “lower level” model.
The largest update is that there may need ot be some rethinking about what to do about the autodealer at the moment. Since we do not know if it will work with our current goals at the moment. I might have to make my own dealer and purchase the pieces.
In addition, I received the RFID equipment and buttons, and will need to test them out this coming week. Also we will begin gathering materials for the table to begin fitting everything in for the demo.
This week I made meaningful progress bringing the input of the chip sorter prototype together. The two key milestones reached were getting chips moving sequentially through the track reliably, and confirming that the magnet pickup is working consistently. Having both of these functioning together is a significant step forward from last week where I was still validating tolerances and adding magnets to the wheel.
Here is a step by step of the working prototype for the magnet pick up. This week I will be 3d printing all of it in a more solid piece.
Progress is still a bit behind schedule due to the printer downtime from last week, but the working prototype represents a real recovery. This week I hope to do some extra work and printing to catch up. My goal for next week is to integrate servos into the sorter for color-based chip sorting, which will move the project from a mechanical sequencing prototype into a functionally complete sorting system.
The most significant risk is the integration of parts before the interim demo. We are starting this work early in the week to hopefully mitigate the risk of it taking too long. Going to start working on the table to be able to put everything on it and testing RFIDs.
No changes to the design this week.
Schedule has stayed the same.
After returning from spring break, I completed the ethics assignment, which took well over a day. It was much longer than expected, especially since I have strong feelings about several of these topics, and I felt the need to express them.
I am still awaiting the RFID Readers and am excited to finally test them and implement them. I do feel that there is very little possibility for error or need for drastic change if it does not work as I am expecting them too. If anything, it is likely we would have to purchase a couple more readers to satisfy the use cases, but this will be decided once we have received the readers and tested them.
In addition, I have tested out the autodealer and have found it to work well with its own programming of Texas Hold ‘ Em. I now have to make a decision on whether to dismantle the power button so that it can turn on the dealer when one turns on the entire table. This will be a little design change, and the decision will be made by Monday this week.
This week less forward progress was made as the 3d printer was down. Much of my focus then was on reviewing and refining the centrifugal chip sorter design from last week rather than building the new components. After the transition from the vibration based sorter, I wanted to make sure I have this working consistently so I dont run into the same problems with the new design.
I worked on going back over the tolerances on the chip track to confirm that the dimensions are tight enough to reliably make the chips sequential without jamming. I also got some more magnets into the wheel to see how fast I can pick up chips. No major hardware was assembled but what exists was added to.
losing access to the 3d printer for a couple days was a big problem, its up now but I lost a week of new pieces. I am going to mitigate this by trying to get access to tech spark. My goal for the week is having the sorting system fully prototyped with new access to another printer.