While I generally have been strong with my software and electrical skills, I always lacked mechanically. So in early 2018, I started building combat robots. I picked myself up a Shapeoko and Harbor Freight Lathe, hit the books, and taught myself some basic machnining skills.
Bloodsport is my entry for the Battlebots 2019 season with Team Bots ‘n’ Stuff. Progress and details TBD.
Deathstroke is my 2nd 30lb featherweight overhead spinner, taking what I learned from Mowferno and applying some new design features. The new focus for this bot is self righting and designing a reliable electrial system.
Going 0-2 at AVC2018, Mowferno was my first featherweight and the most challenging combat robot I’ve built yet. It featured a 8lb, 25” aluminum overhead bar with 4140 teeth. The drive was composed of 4 seperate drive modules, each with a p60 gearbox, sk3 3536 1200kv brushless motor and 2” colson wheel. Each motor was driven with HobbyKing 40A ESCs flashed with Simonk firmware. Two Scorpion HK-4015 1070k brushless motors powered the weapon, totalling 4.9kW peak power.
This first match I was against a low profile midcutter. While at first look seemed like it might slip below my overhead spinner, it was fortunately just tall enough for me to hit. This was the only low profile bot at the event, so figures I get matched with it first round! The match started and after the first hit, 3 motors came unplugged. Here I discovered the hardway you need tape or heatshrink to keep bullet connectors together and all electronics need a “home” so they don’t slide around so much. Taping down electronics isnt enough for these upper weight classes like it is with my beetle.
The first match was ended in my own tapout, when the opposing robot ripped off my wedge. This sheared 6 bolts and ruined 3 threaded holes. The way you mount a wedge is just as important as the wedge itself.
This match, I was against another solid horizontal spinner. During the twitch test, I discovered my weapon ESC randomly crapped out and stopped working, meaning I didn’t have a weapon for the rest of the match. Just my luck. I still haven’t discovered why this ESC died, however it’s likely a signal issue. This match ended the same as the first. The wedge, now duct-taped on, lasted about 45 seconds. The wedge was violented ripped off, sent into the roof, straight through a pinata hanging above. An exciting end for Mowferno.
A video of a spinup test on Mowferno can be found here
This was a 2 week reiteration on Zero Traction V1. This iteration included 4 20mm brushed drive motors and a 500W brushless weapon motor, along with a much larger base. I also included a mixer that was constructed from an old quadcopter IMU. This was to start testing some hardware for the autonomous combat fighting class at Robogames.
The first match was against SMEE, a largely 3d printed drum spinner. After a few good hits, the weapon stopped working. Inexperienced, I tried the rest of the match to get the weapon back up and running, however during this time, SMEE demonstrated better control, leading to a judges decision in SMEE’s favor.
The reason the weapon stopped spinning was due to a screw a few mm too long cutting into one of the windings of the motor. The screw must of cut through the winding’s enamel after a few hits. The chassis was grounded due to using the Fingertech 3.5mm charge jack, which ties the battery’s ground to the chassis. Thus, one windings on the motor was tied to ground. Unfortunate.
The second match, now in the losers bracket, was against El Shaddai who was another drum spinner. El Shaddai’s wheel was knocked off after my first hit! This gave me a great deal of control over the hits I gave and took, so I continued the rest of the match expirementing with how the robot took and gave hits. In general, I discovered gyroscopic precision is a great friend to overhead spinners, as the robot tends to pop straight up instead of flail itself around the arena uncontrollably, like Zero Traction V1. I won this match by judges decision, but I shouldn’t ignore the fact I cornered myself and ended up flipping in the last few seconds of the match.
This third match was against Thunder Child, a strong kit-bot that I knew I would struggle against. Before the match I discovered some damanged that caused my drive train to only be able to move forward and to the left. V3 of ZT would need to have far greater clearences on the drive, as wheels rubbing against their mounts ended up creating these issues. Knowing this, I went into the match to try to KO, as Thunder Child would definitely beat me in control. The match went predictably, with TC taking a few hits but mostly getting under my bot and ramming me into the wall. Edventually I tapped out as I had lost all control of my bot, with the spinner at full speed. After reseting my RX several times, it was determined there was no way to stop the robot, so we had to wait about 5 minutes for the battery to die before removing the robots from the arena.
Losing control of the robot was later determined to be as following. Before this match, I had moved the motor wires to make some adjustments. These wires ended up being exposed to the weapon during the fight and one phase got clipped. This caused shorting and the IMU hardware to brownout and not recover correctly. New software would need to be written for the “quadcopter mixer” that handles this error. However, the first mistake was allowing the wires to have the ability to be clipped by the weapon :)
This was my first combat robot I designed, taking roughly 3 weeks to complete. The concept here was to a fail fast and reiterate (which failing it did!). In competition I found the drive base to be far too small, causing it to flip over far more then staying upright. The weapon and drive system was far too underpowered and the chassis was overbuilt by quite a bit.
A melty styled bot is in the works.