Youssef Abdallah

Builds/001

Autonomous eVTOL Drone

Field-tested

A 2-meter autonomous eVTOL with a fully 3D-printed airframe

Logged
2024.05
Timeframe
May 2024 – present
Role
Team Lead — UAS2030 @ KFUPM
Stack
3D printingFlight-control hardwareElectronics integrationSolderingCAD
Source
—
FIG 01 — The airframe on the flight line — 2 meters tip to tip, nearly every structural part printed
Contents
  1. 01Mission brief
  2. 02Airframe: printing a 2-meter wingspan
  3. 03The 3-propeller configuration
  4. 04Electronics integration
  5. 05First autonomous flight
  6. 06Leading the team
  7. 07What I’d do differently

Mission brief

UAS2030 is a student team at King Fahd University of Petroleum and Minerals building an autonomous, lightweight eVTOL — an electric aircraft that takes off vertically like a multirotor, then flies missions on its own. The design goal: a 2-meter-wingspan platform for autonomous surveillance flights, buildable and repairable with a 3D printer.

I lead the team. That means I own the airframe design, direct the electronics integration, and — the part no org chart mentions — I’m the person gluing the schedule back together when a wing snaps two days before a flight test.

Airframe: printing a 2-meter wingspan

Nearly every structural component of this aircraft comes off a 3D printer: wing sections, fuselage, propulsion mounts, and the electronics housings. That’s not a gimmick — it’s the core engineering constraint that shaped everything else.

FIG 02 — Wing sections mid-fabrication — a 2m wingspan doesn't fit on a print bed, so everything is designed in joinable segments

Printing the airframe forces honest engineering:

  • Everything is segmented. No print bed fits a 1-meter wing, so each wing is a chain of interlocking sections with load paths designed across the joints.
  • Weight is a budget, not a suggestion. Every gram of infill trades directly against flight time. We tuned wall counts and infill per part based on the loads it actually sees.
  • Crashes become iterations. When a landing goes wrong, you reprint the broken section overnight instead of rebuilding a composite part for a week. The design improved fast because it broke sometimes.
FIG 03 — Fitting an assembled wing — the joints carry flight loads, so tolerances here matter more than anywhere else

The 3-propeller configuration

Instead of a conventional quadcopter layout or a single tractor prop, this aircraft uses a custom 3-propeller configuration — a layout chosen to give vertical takeoff capability without carrying four full-size lift motors through the entire mission.

FIG 04 — The tri-prop layout — lift where you need it for VTOL, efficiency where you need it in cruise

The tradeoff logic: every motor you carry is dead weight in whichever flight phase it doesn’t serve. Three propulsion units was our balance point between VTOL authority, cruise efficiency, and the redundancy/weight budget of a lightweight airframe.

Electronics integration

I directed the electronics integration across the aircraft: propulsion wiring, the flight-control hardware, and the onboard systems that support autonomous operation.

FIG 05 — The electronics bay — flight controller, ESCs, telemetry, and power distribution in a printed housing designed around them

The interesting problems were never individual components — they were interactions:

  • Power architecture: three ESCs plus avionics on one battery means brownout protection and clean power for the flight controller are non-negotiable.
  • EMI discipline: high-current motor leads and a GPS receiver do not want to be neighbors. Routing was planned into the printed housings, not improvised.
  • Serviceability: connectors and bay covers everywhere, because a field repair with gloves on is the real test of an electronics layout.

First autonomous flight

The aircraft flies autonomous surveillance missions — takeoff, waypoints, loiter, return — with the ground station watching telemetry rather than a pilot holding sticks.

FIG 06 — Watching the aircraft fly its own mission plan. Terrifying, then boring, in that order — which is exactly the goal

The first fully autonomous flight is a strange experience: months of work compressed into a few minutes where your only job is to not touch anything. When it landed itself, the team was louder than the aircraft.

Leading the team

The technical work is half the job. The other half:

  • Scoping to reality — a student team’s velocity is set by exams and print times, not ambition. Plans that ignore that fail.
  • Making integration everyone’s problem — airframe, propulsion, and avionics people all made better choices once they saw how their part loaded someone else’s.
  • Documenting like the team will change — because it does, every semester. Build logs like this one started as an internal habit.

What I’d do differently

  • Freeze interfaces earlier. We let the wing-to-fuselage joint evolve while electronics packaging was also evolving, and each change rippled into the other. Locking mechanical interfaces first would have saved reprints.
  • Overbuild the landing gear from day one. Every early failure was a landing. The airframe was optimized before the part that touches the ground was — backwards, in hindsight.
  • Weigh everything, always. A shared weight spreadsheet from the first print would have caught the creep we later had to diet away.