Você está na página 1de 21

3.3.

1 Recycling Project Report


Owen Benner
Bala Parthasarathy
Timmy Pollard-Grayson
Steve Szostak
Principles of Engineering
Block 2
Westwood High School
4 / 26 / 16

Design Statement
We will create a fully automated machine to separate marbles made of different materials.
The marbles will start out all in one container and must be sorted into separate bins by material.
The separation of 16 (or 20 for extra credit) marbles must take under two minutes. The materials
are: wood, opaque plastic, clear plastic, steel, and aluminum. The machine must be able to
separate four of these types (or five for extra credit).

Brainstorming

In my design, a motor turns to release a marble from a gravity feed down a ramp. The
marble rolls down the ramp and onto a platform. The electromagnet tries to pick up the marble
and put it in the steel bin. The marble is now sitting above the flashlight and next to the light
sensor, which has the LED behind it. The flashlight and LED flash, and the cortex records the
data from the light sensors reaction to the marble being illuminated. The program then decides
the composition of the marble based on its reflectivity and translucency. A motor knocks the
marble down one of two ramps, which run opposite each other below the flashlight. The servo
moves its arm across the two ramps if the marble is to be redirected down one of the two ramps
to the side, and if not it stays parallel to them. The marble rolls down one of the four ramps and
into its appropriate bin.

Decision Matrix

Each of us looked over all of the designs and graded them on a scale of one to four in the
categories of simplicity, cost in resources, projected success, flexibility, and efficiency. The
scores were totaled up, and Timmy's design had the highest score, so we chose it. I think
Timmy's design won because it used only one sensor to sort all of the marbles and the structure
was straight forward (it only had one ramp).

Final Design Solution

In the final design, a funnel feeds a marble into a sprocket by gravity. If the marble is
steel, it is pulled off the track by the electromagnet on its way to the sprocket. While in the
sprocket, the marble's reflectivity is measured by a light sensor (by ambient light). The marble
then hits a gate, and the optical encoder measures how much the gate turns from the momentum
of the marble. The cortex then determines the type of the marble based on its reflectivity and
momentum. One of three trapdoors (controlled by motors and servos) then opens to let the
marble down a shoot and into its respective bin, or all three stay closed and it falls down the final
hole.

Design Modifications
We practically redesigned the entire machine while building it. The changes made were
as follows:

The electromagnet that separates the steel marbles spins, picks them up, and scrapes them
off on a ramp rather than staying in one place and pulling them down a ramp. This was
done because the electromagnet wasn't powerful enough to be effective when stationary.
The steel container is next to the ramp rather than below the ramp. It is also a bag
rather than a hard container.

The marble material is determined by a flashlight and light sensor rather than a gate with
optical encoder. Opacity is measured rather than momentum. This was done because we
thought the gate wouldn't work. In hindsight this was a bad decision because the light
sensor is a nightmare to work with.
The flashlight is above the ramp and the light sensor is below the ramp.
A second solenoid stops the marbles so they can be read.

The marble feed is controlled by a solenoid rather than a sprocket. This was done because
we thought the sprocket would be too hard to build. In hindsight this was also a bad
decision because it is very hard to time the solenoid to only let one marble through at
once.
The solenoid is taped to the flashlight.

The marbles are sorted by a servo that splits the ramp into three paths that lead to three
separate buckets rather than three separate trapdoors. Wood is not sorted. This was done
because we didn't have the time or resources to build three trapdoors. In hindsight this

was not a bad idea, but the way it was executed made the structure of the ramp very
unstable because it had to lean to let the marbles down one of the paths by default.
The three marble containers are up next to the ramp, and are beside each other rather
than in a line.

The ramp is supported by four diagonal supports (forming three triangles) rather than five
vertical supports. This was changed to make the base smaller. This was a good change,
because trusses are stronger than vertical supports.

The machine does not sort wood. This was done because wood was too similar to
aluminum and we ran out of time to build a wood and aluminum sorter (I did design one
though).

In this idea, an aluminum or wood marble rolls down an insulated track with two exposed
wires. The wires are connected to a circuit with a battery, a resistor, and an LED. The
marble acts as a switch. If it is aluminum, the circuit is closed, but if it is wood, the
circuit is open. The circuit powers an LED, which signals a light sensor. When an

aluminum marble rolls over the wires, the cortex receives a signal from the light sensor
and determines that the marble is aluminum.

Final Design

The machine did not function well at all. The solenoid at the top would let several
marbles through at once. While it was supposed to be holding the marbles back, the spinning
electromagnet knocked them over the extended solenoid. The electromagnet stopped spinning
after 8 seconds for an unknown reason (it seemed to lose power slowly). The servo changed
position based on the light sensor reading, but it rarely sorted the marbles correctly.

Zero steel marbles were sorted correctly, zero clear plastic marbles were sorted correctly,
two white plastic marbles were sorted correctly, and one aluminum marble was sorted correctly.
15 marbles made it through the process, and the process took 47 seconds.

RobotC Program

More Legible Version


int sprockDelay = 250;
int whitePlasticThreshold = 225;
int aluminumWoodThreshold = 500;
int woodThreshold = 1500;
int type;
int clearAngle = 60;
int whiteAngle = -60;
int aluminumAngle = 0;
bool wood;

//Clear plastic: 0, white plastic: 1, wood or


aluminum: 2

void sprockPrep(){
motor[port3] = 127;
motor[port6] = 127;
}

//Closes solenoid gates.

void magnet(){

//Starts steel separation.

motor[port7] = -31;
}
void sprock(){
motor[port3] = 0;
delay(sprockDelay);
motor[port3] = 127;
delay(2500);
}

//Lets one marble into sorter.

void findType(){
int x;
motor[port2] = 127;
delay(200);
x = SensorValue[in1];
for(int f = 0; f <= 2000; f++){
if(SensorValue[in1] > x){
x = SensorValue[in1];
}
delay(1);
}
motor[port2] = 0;
if(x > whitePlasticThreshold){
if(x > aluminumWoodThreshold){
type = 2;
}else{
type = 1;
}
}else{
type = 0;
}
}

//Determines marble type using lightSensor


values.

void sortA(int t){


if(type == 0){
motor[port4] = clearAngle;
}
if(type == 1){
motor[port4] = whiteAngle;
}

//Sorts the marble into the appropriate bin.

if(type == 2){
motor[port4] = aluminumAngle;
}
motor[port6] = 0;
delay(1000);
motor[port6] = 127;
}
void findWood(){
}

//Not used.

void sortB(bool w){


}

//Not Used.

task main(){
magnet();
sprockPrep();
delay(5000);
while(true){
sprock();
delay(500);
findType();
sortA(type);
if(type == 2){
findWood();
sortB(wood);
}
}
}

//Does nothing.

Design Process

Define Problem: We carefully read the design briefing.

Generate Concepts: We did some research on the properties of the materials. We found
that while both aluminum and steel conducted electricity, only steel was magnetic. I
suggested using water to filter out wood at some point in the process (wood would float),
but Mr. Landers banned its use. We each generated a design.

Develop a Solution: We voted using a decision matrix. Timmy's design won, but then
we decided that the difference of mass between the marbles was too little for it to work,
so we redesigned the sensor assembly to use a light sensor instead. I also designed a
wood and aluminum sorter, as seen above, but it was never used.

Construct and Test a Prototype: We started construction later than we should have,
because we spent too long brainstorming. The construction started well, but then we
decided to use a single servo to split the marbles down three paths. This idea was fine, but
the execution involved having the structure lean diagonally, and because we didn't secure
it very well, the structure wobbled. This likely contributed to our machine's irregularity.
Additionally, the sensor assembly was very irregular. Because I couldn't get the flashlight
and light sensor to align well with the path of the marble by screwing them squarely into
the ramp or sides of the ramp, I had to hammer a strip of metal into a crooked shape to
hold the flashlight in just the right position. This made it very difficult to get regular
readings from the light sensor, which was one of the machine's major problems. We didn't
have time to test the full prototype until the presentation.

Evaluate and Present the Solution: We presented the machine in class. Our presentation
didn't go very well. The machine didn't work at all the first time. Mr. Landers came back

later to give us a second chance, which is the test I described above. The machine worked
slightly better, but still very poorly.

Team Evaluation
Owen: I solely developed the program and the system that read the translucency of the marbles
(though teammates helped a bit with the latter). The program worked as it was supposed to, but I
ran out of time to calibrate the solenoid that controlled the marble feed, so it let several marbles
through at once. I also ran out of time to calibrate the light sensor values. The light sensor system
did not work very well. I came in three times outside of class, but still was not able to get the
machine working properly.
Bala: Bala helped build the ramp, which was wobbly, and the servo gate, which worked well and
was well constructed. Bala helped me get the machine ready to run the program the first time
(before the machine was finished). The program didn't work because we plugged everything into
the wrong ports, but I wouldn't have been able to start the program if he hadn't helped. Bala
planned to come in out of class twice when I planned to. One time he didn't show up, and the
other time he did and did most of the clean up. Overall Bala did his job well.
Timmy: Timmy brainstormed the basis for our final design. He worked on the construction of
the ramp, which was not sturdy. He also helped to construct the servo gate that separated the
clear, white, and aluminum marbles, which worked well and was well built. He also designed
and built a steel separating magnet prototype, but it didn't work so we didn't use it. Overall
Timmy did his job well.
Steven: Steven worked on the ramp and servo gate, and designed and built the final magnetic
steel separator, which didn't work because we didn't have time to test it before our presentation.
Steven planned to come in out of class one time that I did, but didn't show up. I know he did
come in out of class at least once though. Overall Steven did his job well.

I do feel like I did a little more work (with both the entire program and most of the light
sensor system) than my teammates, but this was a massive project and they did what they could.
I do wish my teammates had been better able to help with the VEX sensors, because it would
have let me spend more time calibrating the program, but they did help me assemble the sensor
assembly when I needed more than two hands, and they adjusted the sensors a few times.

Reflection
If we could do this project again, we would find a way to separate the marbles based on
mass, or use the infrared line follower, which I now understand reads the marbles better when
used properly. We would also not use a design with one long ramp, because it was unstable,
bulky, and a lot more complicated than it looked.
The most challenging part of the problem was separating wood and aluminum, which we
eventually gave up on. If we could have used water, it would have been a lot easier. They had
very similar masses and wood would not give persistent light sensor or line follower readings,
because of the variety of colors on its surface. I had an idea to use electricity, but we ran out of
time to test it.
I learned that time is incredibly valuable when working on group projects. I also learned
that an unstable and poorly built structure has a massive impact on the function of a machine,
and that smaller designs, while they may appear to be more complicated, are in some cases more
sturdy and easier to work with.
Working with my team was challenging because I was tasked with the entirety of the
program and most of the sensor assembly. I did ask my teammates to adjust the position of the
sensors on a few occasions, but it didn't help much because the assembly was poorly built and
they crept out of position.

Você também pode gostar