ToasterBot

Test platform for tank drive control scheme

The very initial inspiration for this design was to build a robot to fit inside the shell of a 60s chrome-plated toaster, but feature creep set in and I didn't love the additional mass of the toaster shell. The name remains, but the current build is largely driven by parts I had on-hand, and min-maxing various specifications.

I settled on a differential drive or "tank drive" scheme as I was working with other robots of the same style. It's fairly tricky to become proficient with this platform, versus the Ackermann steering of a normal or RC car. Something cheap and portable would be helpful in training new robot operators, and that's where the current design takes many of its constraints from.

The majority of the parts used are surplus from old robots, like the motors and speed controllers, as well as chain, sprockets, bearings, etc. The control system comes from the RC world, with a cheap PWM transmitter and receiver.

The main construction is rotationally symmetric, allowing me to build two identical modules then bolt them together. While this version is 3D printed, the main structure was designed such that it could be machined from flat pieces of plastic that are sandwiched together during assembly. This makes replacements easy, and lowers the overall design, prototyping, and manufacturing effort.

Gear reduction from the motors is accomplished with a single stage belt reduction, specifically using the smallest motor pulley I could purchase, and printing largest wheel pully that wouldn't hit the ground. This results in a top speed near 20 miles per hour, which is borderline unreasonably fast, but the additional effort from a second reduction stage is negated by overpowered motors that don't need further speed reduction.

The wheels on each side are surplus from the end effector of a larger robot, and connected by #25 chain. Roller chain is very overkill for this application, but it acts sort of like a tread and drags the robot over small bumps it may be stuck on. I considered a long belt, but given the wear opportunity chain seemed more sustainable.

You may be wondering why I've been referring to this is a "robot" the entire time, even though it doesn't appear to have any robot-like features. That curiosity is well placed, as this is better described as a telerobot or unmanned vehicle. All the actions it can preform are directly controlled remotely by an operator or "teleoperated", similar to most so-called bomb-disposal robots. A true robot would be capable of operating autonomously in some capacity, which is not the case for ToasterBot at this time.