Designing a Skittle Sorter

Designing a Skittle Sorter

Nobel Peace Prize Ceremony 2015

Geneva, Switzerland

Tarung Bhimnathwala, Abhimanyu Choudhary, Christopher Decarolis, Alexander Yi

Role Breakdown

Writing Code

What exactly are we doing here?

Research Conducted

Building a Color Sensor

• Abhimanyu Choudhary: Shepherd and leader of Manalapan High School Science and Engineering Program, certified Actuary, as well as project leader, reincarnate of Buddha, Jesus, Mohammed(Tupac Shakur).
• Structural Design and CAD Manager
• Tarung Bhimnathwala: self-proclaimed Quantum Physics deity.
• Circuitry Manager and kid who never picked up the phone
• Christopher Decarolis: ASA(Acoustical Society of America) certified Acoustical Engineer, lead guitarist of 21 Grammy winning band Chris D. and the Baby Seals (pun intended)
• Coding Manaeger and Resident Salt Mine
• Alex Yi: Certified Actuary, head intern at Liberty Science Center, and part time olympic swimmer.
• Documentation Manager

• Color Theory
• How can colors be represented and distinguished by a computer?
• The Arduino Language
• How does it work, and how can it be used to sense color and control motors?
• Circuitry
• How do we operate stepper/servo motors?
• Dampening
• How do we stop things from bouncing and oscillating more than we like?

• Two sorts of code were written for the arduino:
• Interaction with the color sensor
• Stores an array of initial R,G and B values for each color of Skittle (Orange,Purple,Red,Green,Yellow)
• Reads in a voltage from the LDR in the color sensor circuit
• Checks the read in values against those stored
• If there is a match, the skittle is considered to be a certain color.
• Accounts for small variations in values using root mean square error.

• Basic Design: A LDR juxtaposed to an RGB LED
• The RGB LED was initially three white LEDS with red green and blue color filters, ended up being too clunky
• Shine each primary color (Red Green Blue) onto the skittle and obtain voltages from the LDR.
• Check these voltages with prerecorded values
• Based on results of check, find out the color
• The skittle must be isolated from outside light to make it usable everywhere
• A pvc casing was developed, painted all black to prevent any internal reflection/interference (Sorry didnt mean to trigger you)

How exactly are we going to do this?

The Beginnings of a Design

Writing Code (Pt.2)

Issues Encountered(Pt.2)

Changes Made and to be Made

• Wheel was propped onto a felt sheet to prevent friction
• Still has alignment issues
• Color Sensor was covered by small paper drapes to prevent ambient light interference
• Prevents need for calibration every time
• All plywood was screwed to dowels to prevent unintended movement
• Planks were made parallel using a level
• Funnel continues to jam
• Use a solenoid periodically to shake and unjam

Demo

Stepper Motor Circuitry

Issues Encountered

• Stepper motors were nightmarish for a reason
• Very fickle devices, looked to work on their own whim
• "Disgusting" according to Christopher Sanzi
• One poorly soldered connection (thanks Tarung) often destroyed the function of the motor
• Funnel would jam often
• Shaking the funnel with one's hands often did nothing
• Morale often dipped
• Not helped by being within 10 feet of a salt mine at all times

• Code was also written to deal with stepper motor interactions
• To refresh, two steppers were used
• One rotated the final container
• One rotated the disk/wheel that moved skittles
• Code rotated the wheel 90° after each skittle's color was checked.
• After the color was detected bottom container was rotated in increments of 72°(5 partitions for each color).
• Wheel turns, skittle falls into the partitioned container.

• Many redesigns were in order
• Number of Wheels
• RGB LED or Three LEDs with separate color filters
• Issues 3D printing various parts
• Filled in areas where holes should be originally
• Innumerable amount of issues involving imprecision
• A slight movement or change in conditions caused a complete breakdown
• Ex.
• Pressure due to weight of skittles slowed wheel and inhibited proper sorting
• Color Sensor operation varied due to ambient light
• Few measurements were taken.
• Lead to poor aesthetics
• Hundreds of retakes and retests were required until final function

• Perhaps the most nightmarish and traumatic portion of the project.
• We originally used a two-pin setup to conserve pins on the Arduino
• This proved extremely difficult to implement, with botched soldering jobs making the circuits unuseable
• We switched to a four-pin setup, which was easier to execute
• The Arduino, using the leads, excites one coil in the motor at a time
• This rotates the central shaft one step at a time
• Since each step is a defined angle, rotation through a precise angle is straightforward
• Circuit was controlled using an Integrated Circuit(IC)
• Uses a number of logic gates in order to control the aforementioned rotations.

Structural Design

Good Question! We've asked the same one through the past two marking periods.

The Problem-Design a Robot that can sort skittles into different groups based on their color, with no human interaction.

A Quick Meme Break (B4 the gud stuffz)

• Skittle Sorter consisted of 4(four) sheets of plywood stacked with space in between.
• Plywood sheets were held up by wooden dowels
• First layer : funnel and various pieces of graffiti(credit to Christopher Sanzi)
• Second Layer: the "brains" of the machine, contained wheel to move skittles, as well as bulk of color sensor
• Third Layer: a servo motor rotating aforementioned disk
• Fourth Layer: Partitioned container to hold all skittles

• We looked to take a "Systems based approach"
• An approach emphasizing not only each separate component of the project, but also interactions between components.
• What were our our components?
• Color Sensor- A device(s) to detect what color each skittle is.
• A set of mechanical constructs to move a skittle from its initial location to final location.
• A structure to house all above components.

• Near the beginning of the process, we looked to use a design involving two wheels and two servo motors.
• One wheel would move skittles through the color sensor and then to a second platform
• From this platform, the skittle would be positioned over a partitioned container and dropped into it.
• We chose to abandon this design however
• Servo motors are only capable of turning 180° (π rad) relative to their original position.
• Switched to servo motors instead
• Two disks made it difficult to coordinate the movements of each servo, and was an unnecessary abstraction
• Kept a one wheel design
• Instead one wheel would rotate to move skittles, the partitioned container at the bottom would rotate.