History

Sprint 1

Click on the image to see the video!

At the end of Sprint 1, we had a small-scale board designed and built. The aims of Sprint 1 were to, connect with the RasPi to communicate with the Wii balance board, and to design and test our initial mechatronic system for the tilt mechanism.
Mechanically, the biggest insight of this sprint was with the ball joints used in mounting the tilting board to the base. We were using ball joints for all 3 connections, and this left an extra degree of freedom allowing the tilting board to spin rather than purely tilt. This made us realize we needed to use a U-joint in the center mount to prevent the rotation.
Electrically, the challenge was to give the servos 5V, when the signal from the PIC was only 3.3V. We developed the logic converters shown in the Electrical Design section during this sprint.

• Presentation

Sprint 2

Sprint 2 was a lot of experimentation and growth. By the end of this sprint, we had defined our MVP, built a full-scale model of the game, and had tested several actuation and data collection methods.
During the sprint, we Shopbotted our first life-sized hull; due to lack of foresight and changing designs, the panels had to be postprocessed nonetheless to accomodate our central maze and the spinner.
We also implemented our initial layout of the FSM, with which we could control flippers and the central spinner. We also had the coin mechanism working, and mostly accomplished lots of tasks in the way. A lot of work was spent on debugging; the linear actuator was malfunctioning, for instance, as well as the accelerometer library. After a period of efforts in vain, we prioritized on the integrated product, so both of these efforts were tabled.
Overall, we determined our MVP and were well on our way to making a complete game.

• Presentation

Sprint 3

Click on the image to see the video!

This sprint marked our Minimum Viable Game.
Mechanically, we successfully made the game robust with a U-joint designed for a car steering rack. This prevented twisting and made it substantially less likely that an errant 4th grader would damage the system. We additionally solved our issues with the center spinner clearance that we were having in Sprint 2 by increasing the gap between the walls to allow the ball to easily travel between them, lengthening the connection to prevent interference between the walls and the game surface, and integrating the blocking outcrop into our 3D printed part, rather than having a wood flap that fell off during testing.
Electrically, we wired improved the strength of the electromagnets by using a boost converter to step up the 12V power supply potential to 30V. However, we were actuating this signal directly with a simple push-button. Moving forward, we would need to use mosfets to electronically control the actuation of the electromagnets This allowed us to design the electrical system, even without being integrated in the code yet. We tested the pop bumper, so that we could integrate it in the next sprint, and wired up buttons for control as well, and started planning for moving our circuit to proto-boards.
For Software, we updated the state machine to be easily extensible (this turned out to be extremely helpful later on), and had all the game logic and active element control implemented and tested. This was also the sprint where we got the LCD screen with instructions working.

• Presentation

Sprint 4

Click on the image to see the video!

The final 2 weeks of our project were mostly dedicated to polish and robustness. We improved electrical connections, worked out software bugs, recut panels, painted the the game, and more. We took all the design insight we had learned prior and cut an entire set on new body panels out of MDF. We chose MDF over plywood this time around as it was less subject to warp (our 1st game area was a little askew).
The hardboard maze was designed to slot directly into the game surface as a press fit rather than needing to be fastened down by screws, as was the case with the first design. Lastly, a coat of black paint was applied to all the exposed surfaces and a layer of while vinyl was used to cover the countersunk screws holding in various components into the game area.
When we revisited the Raspberry Pi, it unexpectedly broke down, so we had to reconfigure another one. A lot of effort was put into making the startup sequence, as well as the following execution sequence, robust and reliable, so that the game would be a plug-and-play experience.

• Presentation