Mecanum Wheel Drawing Robot
October 2021 - December 2021
This was the final project for an Olin course called Principles of Integrated Engineering (PIE). We had six weeks to complete a project of our choosing. My team's goal was to create a robot that would draw line art without using a gantry so that it could be used to draw images at a wide variety of scales.
I was one of two mechanical engineers on our team of four. My first task was to design the system that would hold the marker. We wanted the robot to be able to draw discontinuous lines, so I decided to use a 3D-printed rack and pinion mechanism to allow it to raise and lower the marker. The holder was toleranced to be a press-fit for the markers that we were using, and it included a spring to help keep the marker tip pressed against the paper with a consistent pressure even on uneven surfaces.
The marker holder was supported on each side by an L-bracket, which fit into a slot on the side of the rack and allowed it to slide up and down, but not side to side. The pinion was driven by a small servo, which was another reason that the spring suspension was important--we didn't trust our cheap servo to return to the perfect position each time we lowered the marker.
However, when the marker holder was connected to the chassis that one of my teammates had designed, it became clear that there was a problem. Although the rack and pinion mechanism worked well and the press-fit was perfect, the marker's maximum height was still too low to be able to clear the paper.
We also realized that the stepper motors that we were using to drive the mecanum wheels were larger than the diameter of our wheels, making it impossible for the robot to gain any traction. We decided to solve two problems in one by ordering new, smaller motors and giving the robot "legs" that would give it enough clearance underneath for the marker to lift off of the paper.
This change also allowed me to alter the design of the marker holder so that the servo and pinion rested underneath of the chassis. That helped to clear up some space for the microcontrollers on top of the chassis.
When our new motors came in, we quickly discovered another flaw in our design. The motors got hot enough under normal operating conditions that they melted the 3D-printed parts that they were attached to. We didn't have enough lead time to order new motors again, so we instead added a three-layer fiberboard shield to protect the plastic. It included spacers between two of the layers to further decrease the heat transfer between the motors and the legs.
Here's what those design changes looked like when we finished fabricating and assembling the new parts.
And here's what the robot looked like fully assembled.
It was then time to start testing the robot's drawing abilities. Here you can see an early test of the robot drawing lines in different directions. Notice that the robot couldn't get enough traction to move sideways when the pen was lowered.
This is a Paragraph. Click on "Edit Text" or double click on the text box to start editing the content and make sure to add any relevant details or information that you want to share with your visitors.
With a few tweaks, including weighting the robot with some large bolts, we were able to get the robot to draw while moving in any direction, although the lines were much less consistent when it went sideways.
This is a Paragraph. Click on "Edit Text" or double click on the text box to start editing the content and make sure to add any relevant details or information that you want to share with your visitors.
By demo day, our robot was able to draw recognizable shapes, although they were often shakey in the corners. We did not achieve our goal of creating a robot that could reproduce an image accurately in line art, but it was able to create some interesting pieces of art.
This is a Paragraph. Click on "Edit Text" or double click on the text box to start editing the content and make sure to add any relevant details or information that you want to share with your visitors.
If we'd had more time, we might have switched out the mecanum drive for something more steady, and would hopefully have fixed the rounding error in the software that prevented the translation of an image to lines.