Servo - August 2016

34 SERVO 08.2016

sensorless motor starting algorithms and rapid reversing. The robot community quickly began modifying RC controllers to drive smaller robots with a moderate success rate.

One builder, Lucas Grell, wrote a guide on how to use SimonK in particular for robot applications. Go to https:// goo.gl/3wQTVX to check it out. I was curious if the same firmware programmed into a larger RC controller could handle a much bigger robot.

The enemy of sensorless controllers is inertia. If the controller can't prod the motor into moving to detect its voltage signal, it will never run the motor. Big robots have a LOT of inertia! Luckily, both SimonK and BLHeli are very configurable; they are written to be modified for a range of RC controller models. I decided to begin with SimonK because of my familiarity with the common Atmel ATmega chips found in many RC controllers.

My experimental subject was the HobbyKing Turnigy DLUX 250A HV controller which I selected out of a few candidates after buying them and taking them apart. The DLUX 250A had an ATmega microcontroller, good construction, and — most important — wads of transistors and a massive heatsink to handle the motor current.

To load SimonK onto this controller (which was not in the list of premade configurations), I spent several hours tracing the signal circuitry to find out which microcontroller pins were used for what purpose.

It worked on the bench, but I needed to put it in a 250 lb robot to really get a sense of what I had to tune. Luckily, I had built a "practice bot" prior to BattleBots Season 2. This test dummy (named Sadbot) was welded out of six inch steel C-beams and 1/4" thick plates, with Overhaul's spare wheels bolted in. Quite literally made from a piece of highway overpass, it was a crude way to get 250 pounds on four wheels, and was exactly what I needed.

To interface the fast RC airplane motors to Overhaul's spare wheel assembly, I had to mount them to BaneBots P80 gearboxes, for which they were a near perfect fit. The custom gearmotor dropped in place of the F30-400 motors, and I was in business.

While Sadbot could move, I realized that extensive tuning and modification was needed to get "DC motor-like" behavior. First, I discovered the controllers were blowing up capacitors because of the current spike of starting a 250 lb load. I replaced all the capacitors with high performance power-supply grade ones as a result. Next, I found that the robot's driveability was improved significantly by enabling the braking feature of Some large RC airplane motors, like the kind I would end up using in Overhaul 2.