Sprint 2: Status update!
The sprint goal from last Friday was to perfect the dispenser, but a full list of what we wanted to do included making:
- A hopper that release sprinkles consistently and precisely
- A new one-axis gantry design to move the dispenser smoothly
- And finally, new programs in Arduino C and Python that use digital images to determine the stepper motor’s position.
Despite having created a one-axis gantry that moves smoothly with the help of linear bearings, hardboard pieces to replace the cardboard boxes from the last sprint, and a second metal rod to reduce friction, we found ourselves with a lot of work ahead of us. The version of the gantry for this sprint was still a one-axis gantry, and we needed a two-axis gantry to satisfy the requirements of our minimum viable product. Also, the gantry had some trouble successfully executing forward backwards movements as it got stuck as it moved towards the end of the rods in this direction. We added a full agitator and funnel system along with a conveyor belt meant to separate the stream of sprinkles into single-sprinkle outputs, but we still faced significant issues with the dispenser. Our sprinkle dispenser did dispense sprinkles according to an image output, but it did not do so consistently or precisely — the sprinkles were more likely to get caught in the back of the guides on the conveyor belt than to go where the team expected them to. We found ourselves with a lot of issues to resolve, so we’re going to focus solely on scaling up the gantry to two axes while doing some work to make the dispenser work next sprint. Of these two things, we identified the two-axis scaling as the biggest technical risk because doubling a system in one sprint is rather difficult and because without a two-axis gantry we have no chance of reaching our MVP goal, and we’d just have to forget the stretch goals.
This sprint we also developed successful black and white image processing! Anne put together code that processes low resolution images into “sprixels” or sprinkle pixels that the motor-related code could then process into movements for the gantry to execute. The team considered this to be one of the biggest accomplishments of the sprint despite that we’re unsure if we’ll reach a level of complexity at which this code might be used.
Of course, we couldn’t publish this post without posting a few videos! Here is our dispensing system working. If you look very closely, you’ll see a lone sprinkle come out!
