top of page

Automatic Ball Sorter

Background

The automatic ball sorter was a group project for my Intro to Mechanical Design course.    I was one of 3 team members. The project had a 5 week timeline.

​

Scroll to the bottom of the page to see a video of the device in action!

​

Problem statement

Design, fabricate, and test a device that will autonomously separate golf balls of two colors and ping pong balls of two colors by ball type and color and then deposit them in one-second intervals into four separate containers.

  • Ball characteristics:

    • 2 types: ping pong or golf ball

    • 2 colors: orange or white

    • This yields 4 possible unique combinations

  • All balls must end up in their designated containers

  • Must be at least 8 balls (2 of each color/type matchup)

    • Upper limit is 24 balls total​

  • Balls are initially loaded in random order​

​

My contributions​

  • Designed a space-saving turnstile and an elegant weight sorter

  • Programmed the Arduino to interface with the color sorter and control the motor

  • Designed novel color calibration algorithm to accurately determine ball color​

  • Designed a good-looking brochure for the project

​

The results

  • 1st place  (out of 31 teams)

  • Device performed flawlessly, sorting all 24 balls in one-second intervals

  • Commended by our professor for our fantastic use of cardboard

    • This helped reduce our weight significantly​

​

Our approach

We knew we wanted to maximize points, so we needed something capable of sorting 24 balls. We also needed to maximize a figure of merit, and we would lose points if our device was heavy, large, or unreliable. Therefore, we wanted to minimize size and weight. 

 

Every person first made an individual first design to get our brains going. Mine is shown below. Note that in particular for my design, the weight sorter is very unrefined. It is larger than it needs to be and would be unusable at high feed rates. However, it was a great exercise, because our final design was adapted from our combined initial ideas.

​

overview_light_annotated_extra_better.jp
main.png
mechanism overview.png
color mechanism.png
lever chute at equilibrium.png

Preliminary design

This concept has many flaws

  • Weight sorter is larger than it needs to be

  • Weight sorter would be unusable at high throughput

  • Color sensor wouldn't have time to get accurate reading​

  • Color lever would need to actuate very quickly

​

It was still a good exercise, though. From this design, we used:

  • Spiral loader

  • Turnstile (the basic idea, for rate control)

  • Cubbies built into base, doubling as a frame

Combining our designs

Our team decided on using a decision matrix to combine the best aspects of our designs.

​

Note that cost was weighted as 0 because we had some of the materials already and we knew the upper bound on cost wasn't very high. The rest of the numbers were arrived at by considering the equation for figure of merit (which we needed to maximize in order to earn the best sciore).

figure_of_merit.png
relative_weights.png

The device in action

Automatic ball sorter
Play Video
bottom of page