I designed a robot arm in Python

A code-defined arm exposes the choices a render hides: link lengths, servo brackets, cable routes, and where the dock’s pad and cradle sit. The torque budget drives them. On the course’s example arm, the stretched pose with the 230 g dummy drone at 32 cm needs about 12.8 kg·cm at the shoulder. Derated by 11/12 for the voltage the cables lose, the 12 V STS3215 has 27.5 kg·cm, a 2.2:1 margin, where the 7.4 V version’s 1.4:1 fails the gate. Its rated torque, the most it can hold without overheating, is 10 kg·cm, so the arm never holds the drone stretched out: on the example arm the pad and cradle sit within about 20 cm of the base axis.

Draw every part for how it prints: pick its orientation while drawing it, lay links flat so their load runs along the printed lines, and fillet inside corners, because a sharp right angle printed upright snaps at a layer line. Draw the arm in its zero pose, pointing straight up with every servo at position 2048, so the Module 7 calibration only trims small offsets. Print one joint first, the sixth servo with its bracket and a weighted lever, which stays on as Module 8’s pendulum bench. The gate checks that every printed part fits a 180 × 180 × 180 mm build volume and that the gripper reaches 0.30 m from the base axis at table height.

Every pitch joint needs a margin of at least 2:1 with the dummy drone at full horizontal reach, from stall torque derated by 11/12, and holding the drone must need at most the rated 10 kg·cm at every joint at the pad, the cradle and every pose the dock path holds. The Module 2 controller must draw the figure-eight with the CAD arm in sim within the nominal gate, and the trainer must review the WireViz plan before the first power-on. Keep cad_to_params.py, which builds link masses and inertias from per-part slicer estimates and weighed servos and writes them into the sim’s Params; solid CAD would overstate the mass of infilled parts. Then dock the drone in sim, from the pad to the cradle and back.