3 - Final Prototype Preparation

Design

Cam Profile Design

In order to design better cams than we did for our first prototype, we decided to apply some of the techniques mentioned in the textbook. In short, some design changes that were made to our cams were the following:

• cam profiles were designed to be continuous across displacement, velocity, and acceleration functions
• pitch diameter of the cams was increased in order to keep all pitch angles between the cam's surface and the output rod less than 30°
• the maximum vertical output height of the tiles was reduced from 1.5" to 0.5" to reduce the size of the cams even further

Just like for out initial prototype, we utilized a graphing calculator to visualize the profile of the cams before manufacturing. An additional step we took this time around, however, was the use of Norton's Dynacam program to calculate pitch angles for the entire cam to determine if we had an appropriate prime radius. This program can be found in the CD of the "Design of Machinery" textbook. The connection between each of the dwells on the cams was chosen to be a modified sine curve, in order to achieve smooth continuity across displacement, velocity and acceleration.

Figure: screenshot of graphing calculator program to visualize cam profiles.

Live demo of graphing calculator program used: https://www.desmos.com/calculator/lsqv4ouwov

Figure: screenshot of Dynacam program used to compute pitch angles.

Once we had a method of making arbitrary cam profile designs, we set out to design the images that our tiles would output. Since we only had a 5x5 display, our image options were limited. We ended up designing the following images to display, where a color of white corresponds to an output height displacement of 0", and a color of black represents and output height displacement of 0.5". Gray colors are intermediate values between 0" and 0.5". For our final prototype, we only displayed the following four images: blank, one, checker, and spiral.

Figure: Potential image outputs for our 3D display (5x5 pixel grid).

Structure Design

For our structure, we ended up designing the entire assembly in SolidWorks, and then exporting those files to PDF drawings that were used by the laser cutter. All connections and contact surfaces between planes, rods, and shafts were toleranced to be either snug fits or loose fits, depending on their function.

Figure: Structure assemblies designed in SolidWorks.

In order to couple and drive all 5 camshafts together with a single motor input, we decided to design a drive train similar to that in train wheels. The drive train consisted of a series of parallelogram linkage mechanisms that rotated at the same speed.

Figure: Underside of assembly showing internal cams.

Video: 2015-12-03 09.17.14.mov

Materials

The large majority of our materials were 1/4" birch wood and 1/8" acrylic sheets that were used for making the various components in the laser cutter. Other materials used were 3/8" diameter Delrin rods for the camshafts, 1/4" diameter wooden dowels used to couple the revolute joints in the drive train, and hot glue for putting things together.

Base and outer surface: 1/4" birch wood

Structure sides, cams, drive train, output rods: 1/8" acrylic sheets

Camshafts: 3/8" black Delrin rods

Figure: Laser cut materials used in our final design.

Mechatronics

For our mechatronics, we used an Arduino board with a motor controller, and a AC to DC power supply. Our Arduino code was written to read the motor encoder position and rotate by 90 degrees with each press of the reseat button. This had the effect of switching between images whenever the button was pressed once. In addition, the display could be run in "continuous" mode by unplugging the encoder signal wires, since that way the motor would never find it's target position and keep turning.

Figure: Motor connected to the center shaft of 3D display. It is controlled via an Arduino motor controller.

// current code
#include "DualMC33926MotorShield.h"
enum PinAssignments {
  encoderPinA = 2,
  encoderPinB = 3,
  clearButton = 8
};
enum Positions {
  one = 1633,
  two = 3267,
  three = 4900,
  four = 6533
};
volatile unsigned int encoderPos = 0;
unsigned int lastReportedPos = 1;
boolean A_set = false;
boolean B_set = false;
DualMC33926MotorShield md;
void stopIfFault()
{
  if (md.getFault())
  {
    Serial.println("fault");
    while(1);
  }
}
void setup()
{
  
  pinMode(encoderPinA, INPUT); 
  pinMode(encoderPinB, INPUT); 
  pinMode(clearButton, INPUT);
  digitalWrite(encoderPinA, HIGH);  // turn on pullup resistor
  digitalWrite(encoderPinB, HIGH);  // turn on pullup resistor
  digitalWrite(clearButton, HIGH);
// encoder pin on interrupt 0 (pin 2)
  attachInterrupt(0, doEncoderA, CHANGE);
// encoder pin on interrupt 1 (pin 3)
  attachInterrupt(1, doEncoderB, CHANGE);
  
  Serial.begin(115200);
  Serial.println("Dual MC33926 Motor Shield");
  md.init();
}
boolean turn = false;
int target = -1633;
void loop()
{
  if (lastReportedPos != encoderPos) {
    Serial.print("Index:");
    Serial.print(encoderPos);
    Serial.println();
    lastReportedPos = encoderPos;
  }
  if (digitalRead(clearButton) == LOW)  {
    encoderPos = 0;
  }
  
  
  int i = encoderPos;
  
  if(i < (target-15)) {
    md.setM1Speed(-100);
    stopIfFault(); 
  } else if(i > ((target))) {
    md.setM1Speed(100);
    stopIfFault(); 
  } else {
    md.setM1Speed(0);
    stopIfFault();
  }
}
// Interrupt on A changing state
void doEncoderA(){
  // Test transition
  A_set = digitalRead(encoderPinA) == HIGH;
  // and adjust counter + if A leads B
  encoderPos += (A_set != B_set) ? +1 : -1;
}
// Interrupt on B changing state
void doEncoderB(){
  // Test transition
  B_set = digitalRead(encoderPinB) == HIGH;
  // and adjust counter + if B follows A
  encoderPos += (A_set == B_set) ? +1 : -1;
}

Constraints

Due to cam design limitations, we were not able to make a mechanism with more than 25 tiles output within a reasonable size. Even at this low resolution, our 3D display required a structure with dimensions of about 12"x12"x6". This was due in order to assure smooth motion as the cam rotates, which requires large cam diameters. In addition, we had to lower our output tile heights significantly in order to reduce the cams' diameters to a size that would fit within a cubic foot.

In terms of manufacturing limitations, there were certain limitations imposed by the laser cutter. Due to error in calibrating the laser head and the fact that the laser "eats up" material, tolerancing and fitting parts snugly with one another repeatably proved to be a difficult challenge. We were also forced to make a 3D display design that would fit in a cubic foot of volume, due to maximum size limitations of the laser cutter.

Overall, our design was made to mitigate many of these issues, but a better prototype could be manufactured in the future using metal pieces, which do not deform like wood and plastic do, and more accurate manufacturing methods, such as a CNC machine to make cams.

Photos and Videos