Driving around my beetleweight (3lb) combat robot prototype using my custom xbox controller radio. The prototype is made out of acrylic, the final version will be made out of aluminum or possibly magnesium if I need the extra weight.
There are several problems:
There's wires hanging all over the place. The biggest problem is power distrubution due to the thick wires on everything. It really needs a custom board with carefully aligned connectors to be managable at all.
The raw tank steering is difficult to control. I plan to integrate an IMU to have guided directional movement.
The electronic speed controllers (ESCs) do not start smoothly. This is a programmable setting but...
I cannot find a compatible programming card for the HK-SENS-35A ESC. The recommended one (HKSS) is no longer available.
The ESCs do not switch immediately from forward to reverse. Once the motor is stopped, you have to return the throttle to neutral to begin reversing. I have an afro ESC that does not have this problem, but it doesn't support hall effect sensors. I'm going to order a second one and try driving on them. It may also be possible to work around this by programming the receiver to spam the neutral-to-reverse signals.
It drifts slightly to one side because I modified one of the motors to test fit hall effect sensors. I got the sensors working in a bench test, but given the other problems with the ESCs I may not bother building the sensor mounts.
It has too much drive power. I programmed the radio to have stick movement be ~15% power and the right trigger to act as a multiplier. When you see the bot spin out like crazy, that's me barely goosing the trigger. I guess this does avoid the brushless motor stalling problem. It can easily spin all four wheels when driving up against a wall.
Acrylic construction limits how much I can tighten the screws, so things tend to vibrate loose.
The gearing is too tall. My original plan was to use a worm gear to have 20:1 reduction, but I opted for the current system (6:1 reduction) for easier construction using parallel plates instead of right angles.
The next step is to test some different motors and ESCs, and then program some alternate control modes that use the IMU.
oh wow that's pretty nice you should probably find a way to muffle all that noise though because it sounds like a dying car when it drives also name it betelgeuse
There's maybe $200 worth of parts in the prototype bot here, then another $150 or so for the controller. The whole project including mistakes, spare parts and tooling was probably around $2000
If you were going to make a just-for-fun kit, you could probably reduce the price by using plastic gears, a smaller battery, and a simpler way to put the frame together. There would still be too much hand labor involved in attaching the gears to the wheels though.