By: Marcus Blaisdell
During the spring semester of 2015, two architecture professors, Saleh Kalantari and Ebrahim Poustanchi came to the Robotics Club to ask for help in creating an interactive art project they called “Robobble“.
This project was described as a changeable, amorphous blob that could be controlled by a cell phone app.
I, Marcus Blaisdell, together with Austin Bonnes and Tim Pizzino all volunteered to help.
The team started meeting and discussing what was required to create this project. The architects first envisioned one hundred, a meter long, telescoping rods that would be arranged in a sphere and control the shape of the piece. We investigated our options of what could actually extend and retract. Some of the ideas considered were hydraulic extenders and linear actuators.
The team decided on linear actuators. Austin located several options online and the team considered them and decided to order two different versions for evaluation. The first actuator to arrive was notably heavy but performed well. The speed was pretty slow for what the architects wanted and they asked for other ideas but no one had any.
For control, Arduino Mega seemed the obvious choice. It has 54 digital I/O pins and when the size of the linear actuators was taken into consideration, the overall piece was reduced to only twenty actuators arranged in a duo-decagon. Each actuator would require two I/O pins to control the bidirectional movement, which the Mega can handle.
In terms of wireless communication and control, bluetooth technology seemed to be the best option since all modern smartphones have it natively.
To power the actuators with the Arduino, we would need some type of motor controller. The architects wanted to keep the entire project under $3,000 and the linear actuators were priced at $140/ea so most of our budget was taken right there. I consulted with an electrical engineer at Schweitzer Engineering Laboratories, Andrew Gulbrandsen, who suggested we use transistors in an H-Bridge configuration. This was investigated and the cost seemed very reasonable, at approximately $4 per motor. I proceeded in creating a parts list, which were then ordered.
When the new linear actuators arrived and were tested, they were all found to be defective. They would wobble terribly. It appeared that they were not perpendicular in their bases and would rotate in a circle that was quite pronounced at the end of the arm. The pieces were returned and as we were at the end of the semester, the project was put on hold.
As the Fall 2015 semester began, the project was started again and new actuators were ordered. Once they arrived, I attempted to construct the H-Bridge controllers but found that the incorrect transistors were ordered and they were too underpowered for this application. We needed them to fully turn on with the 5V output from the Arduino but they required 12V to fully turn on. Marcus then consulted with two electrical engineers from SEL, Andrew Gulbrandsen and Doug Bruns, as well as EE students, Matt Foreman of the Robotics Club, and Ryan Summers of the RoboSub Club. Everyone agreed that the transistors would not be able to handle this task. We were left with the option of either trying to order more transistors and continue with the H-Bridge or order regular motor controllers. I found some motor controllers, DROK L298N, that could handle two 12V motors per and were only $8.20/ea. Ten of them were ordered and tested, which we have found to function perfectly.
The original power supply for the project was a laptop power supply that worked very well for testing a single linear actuator. Due to a communications error, seven additional laptop power supplies were ordered with the expectation that they would be used in combination to power twenty motors. This proved to be unreliable. The power supplies were not able to provide enough current and so two Robotics Club members, Connor Cole and Matt Foreman, suggested a PC power supply. I contacted VGH and found that they had a 900W in stock that was capable of 40A at 12V. This was purchased and hooked up and had more than enough power for all twenty motors simultaneously.
For the cell phone app, nobody had any experience or knowledge of how to build one. However, Matt Foreman suggested that we should try using the MIT App Inventor. This was found to be very easy to use and was utilized to create the app.
The architects handled the construction of the physical piece and I did the wiring and programming. The first version of the firmware and phone app was trying to use a slider bar to set the position of each actuator and since the actuators have no position sensing or any sensors at all, this was being approximated with timing. In the end, this proved far too difficult and so it was modified to be simply out, stop, or in with the operator deciding when to stop each motor at the position they desired. This worked much better. Gabriel de la Cruz provided feedback on the Arduino code suggesting to use arrays to handle all of the variables instead of assigning each individually. His suggestion reduced the overall size of the Arduino file by almost 30%!
This project required a lot of help from a lot of people and was truly a group effort with contributions from several members of the Robotics Club, RoboSub Club and Schweitzer Engineering Laboratories. Special thanks go to Ace Hardware for their help with advice and equipment to wire the project together.
The Robotics Club was also mentioned in the WSU News Article.