Delivery System¶
Task 5 of the project: the drone must carry a small payload to the detected landing pad and release it there. This section documents how the release mechanism was chosen, how it is built and driven, and how the mechanical parts around it get manufactured.
The chosen design in one paragraph¶
The payload is held by a 9 g micro servo acting as a pin release: a servo
arm blocks the payload (or a small trapdoor) in the closed position and swings
clear to let it fall. The servo's signal wire is driven directly from the
Raspberry Pi companion computer (GPIO18, 50 Hz PWM, 1100 µs closed /
1900 µs open) using gpiozero with the lgpio backend, while its power comes
from a separate 5 V BEC — never from the Pi's 5 V rail. The companion
software (release.py in the Pi-Code repo) hides the servo behind a small
ReleaseMechanism interface, so the same mission code that drops via an FC
output in simulation drops via the Pi GPIO on the real aircraft. This won over
grippers, winches and electromagnets because on a 3.5-inch indoor CineWhoop
every gram and every millimetre of prop-guard clearance counts — see
Research & Concepts for the full comparison.
Pages in this section¶
| Page | What it covers |
|---|---|
| Research & Concepts | The release options we compared (trapdoor/pin, gripper, winch, electromagnet, passive hook) and why the servo pin release won |
| Servo Mechanism | The implemented mechanism: wiring, PWM values, software stack on the Pi, the FC-output alternative used in SITL, and how the mission triggers the drop |
| 3D Printing | The Tinkercad → Cura → university-printer workflow and the design requirements for the servo/payload housing |
Related pages elsewhere in the docs:
- Frame Extension — the mounting platform the housing will attach to
- Raspberry Pi — the companion computer that generates the servo PWM
- Incident 2026-08-21 — why the aircraft is currently grounded
Status¶
Status (2026-08-22)
- Concept and electronics: decided and implemented. The servo drive
pattern (gpiozero/
lgpioon GPIO18, separate BEC) is verified on the bench, and the software path is fully validated in SITL via the FC-servo variant. - Open: a real drop during flight has not happened yet — the drone is not flightworthy after the 2026-08-21 crash. The 3D-printed servo/payload housing is still to be designed (requirements here), as is its mounting on the frame platform.