mechanical design
This is Gwen. After several design changes and modifications, this is her in her final form.
Quick Specs:
Dimensions- 11" Diameter x 11.5" Height 2 independent drive wheels 3" Skate Wheels 1/4" shaft 1/4" bearings 1/4" rigid coupler DC Motor Jameco 161382 2 ball casters Chassis 1/4" Duron and 3/8-16 threaded steel rod Shooter 1:10 scale model wheel for shooting balls 2" PVC tube shooter RC servo, sub-micro Hopper Stepper motor Jameco 155460 1/4" Duron |
Shooter Mechanism
This GIFF illustrates the range of motion of our shooter. Using one lever attached to the servo and a 2 force linkage, we were able to control the rotation of the shooter. The rear of the shooter rotates on a bolt that is threaded through a piece of wood attached to the PVC shooter tube. |
Shooter Calculations
Shooter calculations were performed in Microsoft Excel to determine the optimal shooter angle, the required initial velocity, and the corresponding required motor RPM. The first chart below shows the initial calculations of motor velocity and RPM required to reach a target distance of 86 inches (assuming shooting from the back of the board) at various angles, based on inputs of shooter height, target height, and wheel radius. The second chart was calculated based off of the velocities found in the first chart, and was used to determine the y position of a projectile in flight at any given x position and angle. From the second chart, we created a projectile motion graph to visually see which initial angles would be most effective.
Shooter calculations were performed in Microsoft Excel to determine the optimal shooter angle, the required initial velocity, and the corresponding required motor RPM. The first chart below shows the initial calculations of motor velocity and RPM required to reach a target distance of 86 inches (assuming shooting from the back of the board) at various angles, based on inputs of shooter height, target height, and wheel radius. The second chart was calculated based off of the velocities found in the first chart, and was used to determine the y position of a projectile in flight at any given x position and angle. From the second chart, we created a projectile motion graph to visually see which initial angles would be most effective.
The projectile motion trajectories graph showed the trajectories required to hit the target from a distance of 86". The dotted grey lines show the upper and lower bounds of the target. It is clear from inspection of the chart that angles around 25-30 degrees are optimal since the projectile stays within the target range well before and after the exact target point - meaning that if there is a bit of error in the shooter velocity the projectile will still likely reach the target. The 30 degree angle was preferred over the 25 degree angle, however, because it requires a lower initial velocity (highlighted in green on the first chart). Looking at the 25 and 30 degree angle trajectories, they also allow the projectile to hit the target from a distance of 48" using the same initial velocity - which would occur if the robot is at the wall closest to the target.
Based on the shooter calculations we selected a 15000 RPM DC motor, and built the shooter to have an approximate angle of 30 degrees (although we made it adjustable for testing).
Building the Shooter
Initial iterations included a cardboard and PVC prototype that used a "windmill" indexer (below, left), followed by a curved PVC prototype which allowed the AF's to roll smoothly through the tube when dropped in from above (right). The initial prototype functioned well, but synchronizing and positioning the indexer was a challenge. The idea behind the second prototype was to use a rotating hopper that would drop AF's in from above instead of relying on the windmill indexer. The second prototype caused the AF's to lose speed as they were ejected, since the curved portion of the shooter barrel redirected the AF's after they passed under the flywheel. The second shooter prototype also had vibration issues because the flywheel was attached to the DC motor using a spider coupler. To reduce vibrations and improve the AF's trajectory, we made our final shooter with a straight barrel and a direct-coupled motor. The flywheel position, the barrel angle, and the shooting direction were all adjustable.
Drive System
The drive system consists of two independently driven wheels. The wheels are mounted to the motor using a rigid shaft coupler and 3 pillowblocks. The pillowblocks on the axle include ball bearings for smooth axle rotation.
Initially, the L293 driver module was used to control the motors. However, we switched to the L298N which allowed us to set our own input pins and enable fast brake
The drive system consists of two independently driven wheels. The wheels are mounted to the motor using a rigid shaft coupler and 3 pillowblocks. The pillowblocks on the axle include ball bearings for smooth axle rotation.
Initially, the L293 driver module was used to control the motors. However, we switched to the L298N which allowed us to set our own input pins and enable fast brake
Drive Motor Sizing Calculations
Problems faced in drive system
1.) Axle misalignment
Originally, we used a 4 mm axle for the wheels and a spider coupler to connect it with the motor. Since there was no 4 mm coupler hub available in the shop, we took the smallest nearest size and drilled it out to 4 mm, but were unable to get the axles to lie straight in the hub. The errors introduced by the axle mismatch, by the use of hot glue to fasten the pillowblocks, and the overcut Duron pieces caused our wheels to experience severe internal torque and axial eccentricity.
Resolution: We switched to 0.25" shafts to match the motor, used rigid couplers, and allowed the motors to wiggle slightly in the wheel setup to relieve the internal resistances.
2.) Weight distribution and uneven playing field
During initial testing of our drive system with representative weights taking the place of the remainder of the robot, Gwen was not able to rotate in place and often pivoted instead. Being able to rotate in place is critical to navigation and orientation. We noticed that during these rotations, one wheel would not make contact with the surface of the Echo Chamber, either due to weight distribution or the unevenness of the playing field.
Resolution: We filed down the regions of the chassis where the ball casters are located. This ensured that at any point, the weight of Gwen would be distributed between the two wheels and one ball caster
Hopper Mechanism
The hopper is a rotating duron disk that holds 12 balls at a time, and is spun using a Jameco 155460 stepper motor. Below the top rotating disk is a stationary disk with only one hole in it, oriented so that a ball can be fed through the hole directly into the shooter. There are four supports on the underside of the upper disk, which glide along the lower disk as the upper disk rotates.
The side view of the hopper shows the two disks, the stepper motor, and spacers used for mounting the lower disk onto the threaded rods. Below on the top right the mounting mechanism is shown. The bottom spacer has a hole just large enough to accommodate the structural 3/8" threaded rods, while the two upper spacers and the disk have holes large enough to accommodate the nuts used with the threaded rods. The spacers are glued together, allowing the structure to be constrained by sandwiching the bottom spacer between two nuts.
The stepper motor is mounted to the lower disk, with a hole large enough to accommodate the stepper motor shaft and a coupler. The coupler was simply a single side of a spider coupler, which coupled the stepper motor shaft directly to the upper disk by having two small drilled holes in the upper disk that mated with the protrusions on the coupler.
The mechanical components of the hopper functioned well. The problems arose when integrating the stepper motor with the rest of the electrical systems, as described in the electrical design section.
The hopper is a rotating duron disk that holds 12 balls at a time, and is spun using a Jameco 155460 stepper motor. Below the top rotating disk is a stationary disk with only one hole in it, oriented so that a ball can be fed through the hole directly into the shooter. There are four supports on the underside of the upper disk, which glide along the lower disk as the upper disk rotates.
The side view of the hopper shows the two disks, the stepper motor, and spacers used for mounting the lower disk onto the threaded rods. Below on the top right the mounting mechanism is shown. The bottom spacer has a hole just large enough to accommodate the structural 3/8" threaded rods, while the two upper spacers and the disk have holes large enough to accommodate the nuts used with the threaded rods. The spacers are glued together, allowing the structure to be constrained by sandwiching the bottom spacer between two nuts.
The stepper motor is mounted to the lower disk, with a hole large enough to accommodate the stepper motor shaft and a coupler. The coupler was simply a single side of a spider coupler, which coupled the stepper motor shaft directly to the upper disk by having two small drilled holes in the upper disk that mated with the protrusions on the coupler.
The mechanical components of the hopper functioned well. The problems arose when integrating the stepper motor with the rest of the electrical systems, as described in the electrical design section.