Today I had the idea to share a story under the theme of “the combat robot engineering you don't see”.

Normally my team works on the Battlebot HUGE and similar small versions, but this post involves my 1lb fighting robot, “Drinky Bird”. Drinky Bird is a hammer-style robot, aiming to hit its opponent with a vertically swinging axe. In particular, it’s designed to have a fairly low-power hammer while being incredibly defensive, mostly to entertain people at a youth-focused combat robotics league. And since most of the parents are familiar with those old Drinking Bird toys, everybody gets a kick out of it.

Over its first few competitions, we struggled to get enough power into the hammer arm to consistently flip the robot back over when upside-down. If Drinky Bird fell over with the hammer arm in the swung position (forwards), it could easily correct itself with the retraction of the hammer. But if the hammer was already retracted, it didn’t have the motor power to lift the whole body upwards over the long front forks. And in a game where the length of forks can generally determine your level of success, shortening them was not an acceptable tradeoff.

In times like these, it’s nice to consult the notebook of neat ideas and see where to pull inspiration from. Another Battlebots competitor called “End Game” is a conventional 4wd vertical spinning robot, with a rear motor-operated self-righting arm. In recent seasons, they implemented a large axial spring on the arm to provide a spring-assist to the self-righting motion. Because it’s an actuator that only does work in a single direction, they can bias the motor torque in that direction more heavily with the spring. Retraction of the arm charges the spring, and it sits compressed at-rest. Then when necessary, the extension of the arm utilizes the spring-assist to more quickly flip the robot over, and to take stress off the motor and gearbox.

Like any good idea, it’s quickly stolen and implemented in a new place and a new way. For Drinky Bird, we hoped that an axial spring was the answer to our self-righting woes. The goal here was to bias the motor torque in the forwards direction, giving it enough power to self-right without removing its ability to correct itself from the rear flipped position. As an added bonus, this means more power going into the forwards swing to hit the opponent with.

A spring was specified on McMaster-Carr to around 40% of the motor’s output torque (normalized for its placement within the gear train between the motor and the hammer arm). Not so strong that it cannot easily retract, and not strong enough to backdrive the motor and move its position. But enough to give a big torque boost to the forward swing.

Recesses were designed into the support uprights and hammer gears to hold the spring legs in place, and the axial spring’s center hole was wrapped around the hammer’s main axle. The spring is held in place by a small piece of tape so it doesn’t fall out during disassembly, and passively held in place by the overall hammer assembly once the robot is all together.

The result was… great success! Drinky Bird has never since had a problem with self-righting, and this fix was significantly lighter (less than 1 gram) and cheaper (96 cents per spring) than installing a larger motor. In a sport where the weight limit is law, I’d argue that the outcome here was even better overall.

So next time a motor in a project is just a smidge too weak, maybe keep this idea in mind to help get your system over the final hurdle.

To end, enjoy a fight from Drinky Bird's recent competition at ColossalCon Anime Convention in Ohio!