Servo-Motor Cart

Project Overview

Project Overview:

Create a cart controlled by a servo motor and a proportional–integral–derivative (PID) controller that transports a cart a variable distance without tipping a vertical 8020 aluminum beam without supports in the shortest amount of time.

My Personal Duties:

Coded project to incorporate servo motor feedback into speed in rpm, transformed to linear distance traveled by cart and adjusted over time
Responsible for SolidWorks Motion Study and force analysis calculations to determine maximum acceleration
Designed main cart housing
Organized work tradeoff between group members and annotated working code documents for knowledge transfers

Ideation Stage

Because one of the criteria was to have the bar sitting on top of the cart without any side supports, I decided to have the base of our cart be entirely flat. This ensures that there is no tipping of the bar before the cart moves and no side support. The main challenge was designing the coupler between the shafts and the motor. I had two ideas in mind, namely a gear system or a pulley system. I decided to use the pulley system, as we could recycle the pulleys we used for the 2.5 DOF system.

Because we wanted the lowest center of gravity for the car to make the bar the most stable, I designed it so the platform was as close to the ground as possible.

Due to difficulty printing the parts via 3D printers, we decided on making our box out of laser cut MDF pieces cut into finger joints to ensure it was light but easy to construct. This design change was made after the motion study was constructed, so it is not reflected in the SolidWorks assembly.

SolidWorks Multibody Motion Study

I generated the basic assembly for the cart system in SolidWorks, as shown to the right. This system includes a motor, a pulley system, axles, a servo motor, and the unsupported bar resting on the top of the platform.

Based on this assembly, I attached the bar to the base of the cart and did a motion study to analyze the forces on the bottom edges of the beam. With an increasing acceleration of the cart, finding when the reaction forces were equal to 0 would be equivalent to when the bar is no longer fully in contact with the platform and will begin to tip.

Based on the motion study, the bar begins to tip at around 32.2 in/s^2, as the reaction force of the bar becomes 0. This was our expected maximum acceleration for the cart.

The cart theoretically would have achieved a forward motion of 7 feet in 3.5 seconds (or a total time of 7.0 seconds to return to start) assuming that there was no maximum velocity to the motor. Clearly this is not a reasonable answer, as the cart will likely not be able to achieve 60in/s and this time does not include a delay between the forwards and backwards motion in order to stabilize the bar.

Building/Testing the Car

Cart Construction

Testing

After coding the PID controller to incorporate the speed of the car to calculate the distance it has traveled, as shown below, we tested the cart with the aluminum bar and adjusted the maximum acceleration rate until the bar tipped. Based on experimental data, our car could travel at a maximum speed of 40 inches per second based on the maximum rotation of the motor. Using these two facts, we were able to determine the ramp up time and the dwell time at the maximum speed to achieve the desired distance.

For our construction, we created a finger joint pattern in epic and laser cut the parts. The wheels were FDM printed and had a GT-2 timing belt ring to allow for greater traction. This timing belt addition was created after we observed the cart not traveling straight due to uneven traction. Our cart includes a small shelf to hold our Arduino UNO board and power amplifier while balancing the beam on the lower portion of the car to maintain the lowest center of gravity. The electrical components were held onto the shelf via velcro for easy adjustment while the wood was glued together.

Determining linear speed of car

Programming car distance

Based on the physical tests, we found that the maximum acceleration for the car was around 34.5 in/s^2, which is very close to our estimated 32.2 in/s^2. This is likely because the bar in the model was higher off the ground than the physical bar was in our system, so the center of gravity would have been higher and the estimated acceleration would have been lower.

As you can see from this video, the bar begins to wobble but does not tip at the maximum acceleration of 34.5 in/s^2. Our record time was 11.12s to travel 7 feet, which is relatively close to our estimated time of 7.0s based on the simulation.

Motion Study Files

Zip file for cart motion study

Zip file for my cart assembly