Lunabotics Drum Excavator Test Bed

Senior Capstone Project

Fall 2021-Spring 2022

Animation of USR's bucket drum excavator operating while attached to the rover

Project Outline

Utah Student Robotics (USR) has built a lunar rover to compete in the NASA LUNABOTICS competition. USR wants to test multiple bucket drum excavator designs and determine the forces and torques acting on the drum while it is mining a lunar regolith. We built a test bed that provides analysis of X-Y forces and torques on the drum, monitors current drawn by the drive motors, and correlates these measurements to the excavator’s depth in regolith simulant. In addition to these engineering analyses, the test bed is IP60 rated (fully dustproof), user-friendly, and operable by one person.

User Needs

System accurately records horizontal and vertical forces on the drum
System accurately records torque acting on the drum
System records the depth of the drum over time
System records the current draw of the drive motors over time
System allows for multiple drum designs and drum widths
System is capable of being transported
System only requires one person to operate
System is rated for IP60 dust protection to minimize dust egress
System allows for video captures of tests
Stay under the allotted budget of $600

Initial design concepts made and voted on by team.

CAD Design

Final Product

Design Features

2 S-type load cells for vertical force measurements
2 bar-type load cells for horizontal force measurements
2 self made torque transducers
2 motors that rotate the bucket drum excavator
A linear actuator to move the lower hub vertically

Results

Our team was able to meet all the requirements except for the depth measurement accuracy. The test bed succeeds in all other areas including horizontal and vertical force measurements, torque measurements, motor current measurements, rotational drive speed, and system cost.

Table showing the specs of the final product.

My Contributions

On this project I was a lead for programming as well as instrumentation. A large part of my efforts went towards the design, fabrication, testing and calibration of the torque transducer and load cell subassemblies, as well as the programming of the MATLAB GUI and the master Arduino code that would send commands to the actuators and receive data from the load cells and bluetooth torque transducers.

Torque Transducers

Due to our small budget we were forced to design and build our own torque transducers. In order to reach the required sensitivity, we designed a full wheatstone bridge configuration with four strain gauges applied to the 8mm shaft. The raw output was run through an HX711 amplifier

Applying strain gauges to 8mm shaft

Wiring for strain gauges on prototype

Prototype torque transducer

Finished torque transducers

The transducers needed to be mounted to the drum of the excavator which means they would be spinning. This meant we needed to use a slip ring or transmit the data to the master Arduino via Bluetooth. We went with the latter and used and Arduino NANO BLE to transmit the data.
In order to power each transducer we included a battery and a power button that could be operated from the outside by turning a screw. The overall enclosure helped protect the transducer and its components as well as keep out the very fine lunar regolith.

Prony brake design concept

Prony brake experimental setup

Linear model of torque based on dataset averages a

To test our transducers I designed a Prony Brake Dynamometer setup which utilized one of our S-type load cells and a motor. Using this setup we were able to verify that our torque transducers one torque transducer was accurate to within ±0.449 kg⋅cm and the other to within ±0.746 kg⋅cm with 95% confidence.

Load Cells

Our system had two type of load cells to measure the forces acting on the drum. S-type load cells were used to connect the drum assembly to the frame and measured the forces in the vertical direction. Bar type load cells were used to measure the horizontal forces. Two of each load cell was used to help remove biases and account for readings due to forces in unwanted directions Each load cell was tested and calibrated in an experimental setup that hung varying known loads from the cell.

Experimental setup for load cells

Horizontal load cell subassembly

Amplifier breakout board

Amplifier breakout board soldering work

The signal from each load cell was sent through an HX711 amplifier. I fabricated the amplifier breakout board so it would be easy to replace components. The amplifiers were mounted on header pins so they could be easily replaced if they were faulty (which some of them were). JST connectors were attached to all of the wires so they could be plugged and unplugged from the breakout board.

Programming

Final MATLAB GUI with randomly generated data plotted

The goal of the UI was to provide the users with an easy and foolproof way to connect to and operate the test bed. The UI is able to scan for the connected Arduino, and lights the connected light if it gets the correct message back. Once it's connected it can connect to each torque transducer. If the transducers are off or don't respond correctly then it lets the user know. Once everything is connected you can set the desired depth of the linear actuator and turn on/off the motors. To record the data you can set a test duration and run a test. The test can be paused or ended at any point. Once the test is complete you can save all the raw data into a single file or save specific data individually. The data plots update in real time as the test is being run to give the users a better idea of how the test is going while it's on it's way.

Special Thanks

Special thanks to my team, team AR^4M, for all the hard work they put in over the year we worked on this project, and to Dr. Mark Minor for being our advisor and helping us out throughout the process!

Page updated

Google Sites

Report abuse