Skip to content
kellanwilliams— home

FPV Camera Drone

current

A custom FPV quadcopter designed, built, and flown to film ultimate frisbee — three full iterations from hand sketches through Onshape CAD, 3D-printed frames, and hand-soldered electronics to controlled flight with 30-minute endurance.

  • Onshape
  • CAD
  • 3D Printing
  • Embedded
  • PID
  • FPV

I wanted aerial footage of ultimate frisbee games. Buying a drone would have solved that. Building one taught me considerably more.

Three full iterations over two years, each one a rebuild rather than a patch: hand sketches → full Onshape CAD → 3D-printed frame → every wire and joint soldered by hand.

v1 — the one that didn’t fly

A high-school attempt that never got airborne. Worth listing, because everything the third build got right came from understanding why the first two didn’t.

v2 — airborne, uncontrolled

Version two flew. It did not fly where I pointed it. Getting a multirotor off the ground is a solved problem you can buy in a kit; getting one to hold an attitude you commanded is a control problem, and that’s where the actual engineering starts.

v3 — the rebuild

Version three was re-engineered from the ground up against the specific failures of the first two:

Change Why
Airfoiled arms Reduce drag through the frame instead of treating the airframe as dead weight
Rainproof enclosures It is used outdoors, at fields, in the Pacific Northwest and in London
Working battery cooling Thermal headroom is what turns a short flight into a long one
100 g lighter Every gram is thrust budget and flight time
Stronger motors Authority to actually execute a commanded correction
Bluetooth PID tuning Retune gains at the field without tethering to a laptop

That last one changed how fast I could iterate more than anything else on the list. Tuning a PID loop is empirical — you fly, you watch the oscillation, you adjust. Doing that over a USB cable means a full landing and teardown per attempt. Doing it over Bluetooth means adjusting between throws.

Result: fully controlled flight, 30-minute battery life, and automatic return-to-sideline so it brings itself back for a battery swap instead of being walked down.

Why this project

It sits at the intersection of most of what I do — CAD and fabrication, embedded control, computer vision, and the sport I’d be at the field for anyway. It’s also the project where nobody handed me a spec: the requirements came from the footage I wanted, and every constraint after that was mine to discover.