Propulsion · electroaerodynamic thrusters · specs published, build pending

Thrust with no moving parts.

A thin wire at high voltage above a grounded aerodynamic collector ionizes the air around it; the ions drift across the gap, drag neutral air with them, and the electrode assembly is pushed the other way. Enclose it in a duct, stack the stages, and you have a solid-state thruster you can build from foam, wire, foil, and 3D prints. Two build specifications reproduce the current experimental state of the art.

Reference for both: Gomez Vega, N. (2023), Advances in electroaerodynamic thrusters for aircraft propulsion, PhD thesis, MIT AeroAstro. The thesis found that Coulomb force per unit volume is mostly a function of power draw per unit area, and that smaller gaps give better volumetric performance at lower voltage — a 15 mm gap produced up to 80% more force per unit volume than 30 mm at the same areal power. Small gaps mean staying under 20 kV instead of the 40 kV used on MIT's 2018 airplane.

Phase 1

Single-stage ducted thruster

One emitter-collector stage in a 180 × 180 mm duct. Goal: a working, instrumented thruster with clean current–voltage–thrust data. Target 50–300 mN.

15–20 kV DC · positive corona · NACA 0010 collectors

Phase 2

Multistaged ducted (MSD) thruster

N stages in series inside a duct with inlet and nozzle; each stage adds pressure rise like a compressor stage. Alternating polarity lets ten stages run from one supply at one voltage.

benchmark 0.84 N · 26 N/m² · 5.0 N/kW

Design parameters — thesis test article vs. this build

ParameterThesis valueYour build (Phase 1)
Duct cross-section180 × 180 mm square, StyrofoamSame — XPS foam board or foam-core
Stage duct height~51 mm50–80 mm
Emitter wire56 µm tungsten50–80 µm tungsten (0.1 mm stainless to start)
Collector profileHT14 airfoil at 10% t/c, 20 mm chord, PLA + foilNACA 0010, 20 mm chord, PLA print + foil wrap
Gap spacing15 / 20 / 30 mm testedStart at 20 mm; holders for 15 and 30
Electrode pairs1–13 tested7–9 pairs, uniform spacing
PolarityPositive corona (emitter +, collector grounded)Same — quieter, higher sparking voltage, less ozone
Operating voltageUp to ~18–20 kV15–20 kV DC

Phase 2 prerequisite: a completed Phase 1 thruster with clean I–V and thrust data. Do not start at the multistage build. The thesis's ten-stage device measured 5.6× the thrust of one stage, with diminishing returns as internal flow velocity — and thus pressure loss — rises.

Documents

Single-Stage Ducted EAD Thruster — Build Specification

Concept, design parameters from the thesis test article, bill of materials, high-voltage supply sourcing, safety protocol, test procedure, and the I–V–thrust data you need before Phase 2.

EAD Propulsion, Phase 1 · 10 pp · PDF

Multistaged Ducted (MSD) EAD Thruster — Build Specification

Stage modules designed once and replicated, the alternating-polarity wiring trick, inlet and nozzle, inter-stage electrical connections, and the benchmark to beat.

EAD Propulsion, Phase 2 · 10 pp · PDF

See also: MTS Build Sheet 002 — droplet thruster

Beamed-thermal propulsion bench work lives under MTS, not here: an acoustically levitated droplet as a micro-thrust balance under laser illumination. Different project, same TinyLev.

cross-reference · MTS Phase 0

Status

  Specifications published — Phase 1 unbuilt next: HV supply sourcing → Phase 1 build → I–V–thrust data

Safety

These devices run at 15–20 kV DC. The specifications' safety sections — supply selection and current limiting, discharge procedure, enclosure, ozone ventilation — are load-bearing. Read them before powering anything.

Authorship & method

This research is drafted in human-AI collaboration: analyses are AI-drafted under the owner's direction, then re-derived, red-teamed, and adopted or rejected by the owner. All decisions, inventions, and commits are Christopher Brotherton's; commits credit AI drafting via Co-Authored-By trailers. The dated commit history is the record of human contribution.

Technical critique is welcome and useful — open an issue. Errors that survive review get corrected in a dated commit.