try out stepper instead of dc motor

[Armandpl]

dc motor issues:

• needs a gearbox to have enough torque + have a thick enough axle to clamp the robot arm onto (axle straight out of those motors are super thin). gearbox is not durable, i broke one already by sending gaussian noise as an action. stepper doesn't have a gearbox and afaik will not be damaged if we skip steps or send garbage inputs.

• needs a slip ring (or a well designed and tuned position controller) to avoid tangling the cable. ideally we don't want low level controllers, we just want the algo to send raw low level actions. seems easier to implement safety limits + resets with a dumb auxiliary controller on a stepper.

• order: nema17, drv8825 breakout, breakout encoder with magnet as5600, breadboard+cables, logic level shifter breakout, some way to provide power to the motor

[Armandpl]

lol just found this https://github.com/ferrolho/rotary-inverted-pendulum/tree/main basically the same
only one encoder makes sense dk what i was thinking when i put two. rl does make the stepper skip steps tho maybe we could have end stops?
printing the arm also seems better, no need to bend and cut aluminium, only need bearings and screws. also no need to print a separate adapter for the encoder's magnet

[Armandpl]

usb-c PD?

[Armandpl]

also, sheet metal vs. printing?