Lorentz Aerospace

The question

Can a compact electromagnetic atmospheric actuator transfer enough electrical power into directed neutral-air momentum to support useful flight?

A helicopter answers that question with blades. It throws air downward, and the air throws the helicopter up. XR-1 asks whether the same job can be done with fields instead of blades: no rotor, no propeller, no exhaust, and the air around the aircraft as the working fluid.

The atmosphere is the reaction mass.

Nothing in the idea escapes Newton. To stay up, a 10-tonne aircraft must push about 98 kilonewtons of air downward, every second it hovers. The only question is what does the pushing. Here the chain is long, and every link in it is ordinary physics:

1. Electrical power from a turbine-driven generator,
2. ionization, which makes a region of the surrounding air conductive,
3. current driven through that conductive region,
4. a magnetic field crossing the current,
5. the J × B body force on the charged particles,
6. collisions that hand the force to the far more numerous neutral molecules,
7. bulk atmospheric momentum: a column of air moving downward,
8. reaction thrust on the vehicle that holds the magnets and drives the current.

The chain is sound. Whether it is efficient enough to fly is the open question, and it is the only one this program exists to answer.

The force

The magnetic field does not push ordinary air directly. Lorentz first creates mobile charge carriers in a controlled region of the atmosphere. Electric current flowing through that region interacts with the magnetic field. The resulting Lorentz force acts on charged particles, which transfer momentum through collisions to the much larger neutral population.

The objective is not to accelerate a tiny plasma jet. The objective is to turn a large volume of surrounding atmosphere into the effective working fluid of a distributed actuator.

Summed over a volume, the force density becomes a force the vehicle can feel:

XR-1 is an attempt to engineer that integral.

The integral is only the start of the accounting. Useful vehicle force depends on where the current closes, on pressure stresses, on boundary terms and above all on how much of the momentum actually reaches neutral gas. The currents and magnets are carried by the vehicle, so internal forces between them cancel; a force that stays inside the machine lifts nothing. Thrust exists only as momentum delivered to external atmospheric mass. Established physics

Electric fields alone can move air. Ionic wind, or electrohydrodynamic thrust, is real atmospheric momentum transfer: ions dragged through the air by an electric field collide with neutrals and set them moving. It has flown a small solid-state aeroplane. Lorentz investigates a different route, current crossing a magnetic field, and asks whether magnetic coupling opens a substantially different regime of body-force and thrust density. Any measurement must separate the two. Established physics Experimental question

The working fluid

The atmosphere is both propellant and problem.

Air at sea level is an excellent insulator. That is why lightning is rare and power lines work. Before any magnetic force can act, the machine must first make the air electrically accessible.

Very little ionization is needed to make air conduct. At the reference density used on this page, one molecule in about twenty-five million carries a charge. The charged particles collide with neutral molecules about a trillion times per second. Those two facts decide almost everything downstream. The collisions pass forces from the charged minority to the neutral majority within nanoseconds, which is what makes momentum transfer possible at all. The same collisions turn electrical energy into heat, which is what makes it expensive. Established physics

And the conductivity does not last. A free electron in sea-level air is captured by an oxygen molecule within nanoseconds to tens of nanoseconds, depending on how hot it is. Conductivity is a perishable good: it must be made continuously, and paid for continuously.

Heat electrons to create conductivity

The candidate mechanism is millimetre-wave electron heating. Microwave fields energize the few free electrons already present; energetic electrons knock further electrons loose by impact; the ionization must be sustained against attachment, recombination and transport. The useful ionization fraction is an experimental result, not a design percentage. Established physics for the mechanism; Experimental question for the yield.

A gyrotron is one source option

A gyrotron converts energy from a gyrating electron beam into millimetre-wave radiation inside a magnetic cavity. Megawatt-class tubes exist: ITER's 170 GHz gyrotrons have delivered 1 MW for 1000 s, and Wendelstein 7-X heats its plasma with ten 140 GHz gyrotrons of nearly 1 MW each. Those are fusion-laboratory installations with their own halls, supplies and cryoplants. Packaging one for an aircraft is a design problem, not a physics one. Its high-voltage supply, launcher, transmission and cooling belong in the power and mass budgets, and its cavity magnet is distinct from the external magnetic field where the atmospheric plasma would interact with the beam. Established physics for the source; Design target for carrying it.

A 60 GHz photon carries 0.00024814 eV; oxygen's ionization energy is about 12.07 eV in the NIST Chemistry WebBook. A microwave photon cannot directly ionize an oxygen molecule. The route is electron heating followed by impact ionization, not photoionization. Established physics

XR-1

A conventional aircraft moves air with surfaces. XR-1 moves air with a volume.

XR-1 is a ten-metre, ten-tonne research configuration built around one idea: an annular electromagnetic actuator on its underside, divided into thirty-two independently driven sectors. Every drawing on this site uses the same vehicle, the same annulus and the same sector numbering. Design target

The shape follows the fields. A flattened disc gives a long circumference for actuator area, radial symmetry for distributed control, a natural home for a large ring coil, short structural paths and a big internal volume. It is a candidate, not a uniquely correct solution: linear segments, partial annuli and other open-flow geometries stay in the comparison. The geometry study covers roughly 7.5–10 m and 10–12 tonnes, with hover-scale reaction thrust of 100–120 kN. Design target

The aircraft is inside the engine.

A helicopter's rotor is a fixed object. XR-1's rotor would be a region of air: wherever the excitation makes it conductive, wherever current flows through it, wherever the field crosses it. At the level of momentum theory, the actuator performs the same job as a rotor: it accelerates atmospheric mass downward. The difference is that a distributed electromagnetic body force replaces mechanical blades, and its shape can be changed in milliseconds.

XR-1 system architecture

Six subsystem groups. Each one is recognizable engineering; the program's risk lies in how they couple. Design target

Propulsive airflow beneath the hull is distinct from any aerodynamic boundary-layer or shock effect; the two cannot be counted twice as thrust, and no elimination of sonic booms is assumed. Structure, shielding, intakes, landing hardware, fuel, thermal rejection and emergency behaviour all compete for the same mass budget.

Momentum

Moving enough air to hover is not inherently the impossible part. Efficiently coupling electromagnetic power into that air may be.

Start with what any hovering machine must do, whatever it is made of. Actuator-disk theory, the momentum balance behind every helicopter rotor, gives the minimum power needed to hold a weight up by pushing air through a disk of a given area. It does not care whether the disk is made of blades or of fields. It follows NASA Glenn's actuator-disk explanation; use $\rho=1.225\,\mathrm{kg/m^3}$ and $g=9.81\,\mathrm{m/s^2}$, sea level, out of ground effect. Established physics

About 2.2 MW of ideal air power. A 10 MW electrical study architecture.

That is the benchmark. A perfect actuator of 10 m effective diameter needs 2.2 MW delivered to the air to hold 10 tonnes; the XR-1 study carries 10 MW of electrical input. So the architecture closes on paper only if at least 22.15% of all electrical input ends up as directed kinetic power in the neutral air. A good helicopter rotor turns most of its shaft power into that kind of flow. The central challenge is plain: can a distributed electromagnetic actuator approach useful fractions of conventional momentum-transfer efficiency? Ten megawatts does not prove feasibility. It sets the target. Design target Experimental question

One reconciliation belongs here. The 10 m diameter is the hull; the physical actuator annulus (2.6 to 4.9 m radius) covers 54.19 m², not 78.54 m². If only the annulus carried momentum, the ideal power would rise to 2.667 MW. Whether the induced flow fills the whole disk, or reaches beyond the hull as the envelope concept hopes, is a measurement. A 10 m hull has not been shown to actuate a 20 m disk. Field extent, visible plasma and entrainment do not establish effective area; only the measured neutral-air momentum flux does. The table assumes ideal flow; nonuniform loading, swirl and other losses raise the requirement. Experimental question

Where the power goes

The field equation is easy. The power budget is not.

Ten megawatts go in. Some fraction comes out as air moving downward. Everything else becomes heat, light or motion that lifts nothing. The program is, at bottom, a hunt for that fraction.

Two costs attach to any steady interaction volume, and they pull in opposite directions. Established physics

The first is Ohmic. For a fixed force $F$ carried uniformly by a volume $V$ in a field $B$, dissipation falls with conductivity, with the square of the field, and with volume. The second is the cost of keeping the gas conducting at all: every free electron in sea-level air is soon lost, mostly by attachment to oxygen, and each replacement costs at least the energy per ion pair, about 34 eV in air (NPL summary of the recommended W-value). That cost grows with volume. Between them sit the three knobs the design actually has: how large the active volume is, how dense the ionization is, and how strong the field is.

Read those rows together. With textbook-order inputs, a steady, uniform, sea-level, 50 m³ interaction volume does not close on 10 MW; it misses by more than an order of magnitude on each cost. That is not a verdict on the concept. It is the specification of the first experimental gate: which combination of pulsed rather than continuous excitation, altitude (three-body attachment scales with the square of gas density, so it falls faster than the air thins), electron temperature (hot electrons attach less readily and detach more), field strength (Ohmic cost falls as $1/B^2$; a REBCO magnet has held 20 T) and geometry (thin, strongly driven regions rather than a uniform volume) brings the two costs down together. The Princeton group's argument that the energy cost of ionization governs cold-air magnetohydrodynamics was made for hypersonic inlets (Macheret, Shneider and Miles); it applies here with full force. Anyone who quotes a thrust number before that measurement is guessing. Experimental question The coupled models must reproduce both costs for a measured discharge before their predictions for a vehicle are believed. Simulation target

Two more losses sit in front of the gas. RF output is not wall-plug input: four 1 MW sources at an assumed 50% efficiency draw 8 MW before anything else runs. And a static superconducting field stores energy without dissipating it, but refrigeration, controls and AC losses are real operating loads. Do not add plasma watts on top of an already all-in coupling efficiency; count every input once.

The atmospheric actuator

One sector, drawn completely. Thirty-two of them make the annulus; one of them is the first thing Lorentz would build. Engineering hypothesis

Plasma generation and current drive are separate channels. Plasma generation establishes conductivity. A slower current-drive or traveling-field system uses that conductivity to accelerate air. Changing one affects the other, but separating them makes the experiment identifiable: hold the plasma state roughly constant while varying the magnetic drive. A phased sequence of coils acts like a distributed atmospheric linear-induction drive, with phase speed $v_{\mathrm{phase}}=f\lambda$. Phase speed is not the exhaust speed. Slip, penetration, current phase and collisional transfer determine how much the neutral flow follows the field. Engineering hypothesis

The field is where you put it

In weakly ionized air the magnetic Reynolds number is far below one. The applied field penetrates the plasma and the plasma does not carry it. That is bad news for anyone hoping to confine air magnetically and good news for anyone trying to control it: the magnetic topology is a coil-design choice, known in advance, not a plasma equilibrium to be discovered. It also means a vacuum-field calculation is a fair first map of the field the plasma will see. Established physics

For a static field, $p_B=B^2/(2\mu_0)$ is 9.95 MPa at 5 T and 159.15 MPa at 20 T. These are local magnetic stresses on the winding and its structure. They are not available vehicle thrust, and if B is a sinusoidal peak rather than a static or RMS value, the time-average pressure carries an additional factor of one-half. Large-bore REBCO magnets have reached 20 T in a fusion test (MIT and Commonwealth Fusion Systems, September 2021); carrying such a coil set, its cryostat and its refrigeration in a flying vehicle is an open design problem. Established physics for the stress; Design target for the coil set.

Beneath the equation

From here the page is written for plasma physicists and aerospace engineers. The introductory relation $\mathbf J=\sigma\mathbf E$ is where everyone starts, and in magnetized, collisional air it is not enough. Established physics

Collisions set everything

The electron–neutral collision frequency is $\nu_{en}=n_n\,\sigma_m\,\bar v_e$. At sea level $n_n=2.5\times10^{25}\,\mathrm{m^{-3}}$; the momentum-transfer cross section of nitrogen for electrons of a few eV is of order $10^{-19}\,\mathrm{m^2}$; the mean speed of 2 eV electrons is about $9.5\times10^{5}$ m/s. Together: $\nu_{en}\approx$ 2.37 × 10¹² s⁻¹. A collisional multifluid model therefore keeps electrons, ions and neutrals distinct rather than treating the mixture as one ideal conducting fluid, with collision rates, Hall parameters and a conductivity tensor; quasineutrality, sheath conditions and current continuity constrain charge separation and current closure. Collective plasma behavior does not establish collisionless or ideal MHD. The NRL Plasma Formulary supplies the scales. The ionization fraction at the reference density, $x=n_e/n_n$, is 4.0 × 10⁻⁸.

Momentum exchange between species

Pairwise exchange obeys $\mathbf R_{st}=-\mathbf R_{ts}$; reaction and injection terms must also conserve total momentum. Summed across species, internal exchanges cancel. Charged-particle force and the same momentum passed to neutral air are not two independent thrust contributions. Electrohydrodynamic ion wind also transfers neutral momentum, by a different route, and must be separated in any measurement. Established physics

Generalized Ohm's law and the Hall parameter

Once electrons gyrate between collisions, current no longer follows the electric field. Neglecting electron pressure and ion slip, the electron momentum equation gives

or, with $\mathbf B$ along $z$, $\mathbf J=\boldsymbol{\sigma}\cdot(\mathbf E+\mathbf u\times\mathbf B)$ with a tensor conductivity

The sign of the off-diagonal terms follows the carrier charge. At 5 T and $\nu_{en}=2\times10^{12}\,\mathrm{s^{-1}}$, the electron Hall parameter is 0.440: electrons are marginally magnetized, so the Pedersen conductivity is 0.838 of $\sigma_0$ and a Hall current of 0.368 $\sigma_0$ flows across the applied field. Ions, with cyclotron frequencies near $10^7\,\mathrm{s^{-1}}$ at 5 T and collision rates of $10^9$–$10^{10}\,\mathrm{s^{-1}}$, have Hall parameters of order $10^{-3}$–$10^{-2}$: they are not magnetized at all, which is what lets them drag the neutrals. Hall currents are not free. Unless electrode segmentation or geometry lets them close usefully, they carry Joule heat and push in the wrong direction. Established physics

Stuart number, magnetic Reynolds number

Two dimensionless groups decide what kind of flow this is. The Stuart number, or interaction parameter, compares the electromagnetic body force with the inertia of the flow; the magnetic Reynolds number compares field advection by the flow with field diffusion:

N increases as U decreases. At hover speeds a given conductivity and field produce a larger interaction parameter than the same hardware at flight speed; slow air is easy to push around once it conducts. The difficulty at low speed is somewhere else entirely: the gas is cold. Hypersonic MHD borrows its conductivity from the shock layer, where the air is already hot and partly ionized. At twenty metres per second nothing supplies it, and every electron must be bought with ionization power. That is the two-cost problem again, from a different direction. Established physics

The magnetic Reynolds number at hover, with the same inputs, is 2.84 × 10⁻⁵. Frozen-in-field intuition does not apply. Small $R_m$ invalidates ideal-MHD assumptions without forbidding externally driven electromagnetic force; it is the regime of electromagnetic pumps, not of stars.

Attachment, frequencies and field penetration

Three-body attachment to oxygen, $e+\mathrm{O_2}+M\to\mathrm{O_2^-}+M$, has a rate coefficient near $2.8\times10^{-30}\,\mathrm{cm^6\,s^{-1}}$ for thermal electrons (IAEA INIS record). At sea level that gives an attachment frequency of 3.68 × 10⁸ s⁻¹, a lifetime of 2.72 ns. Heated electrons attach more slowly and detach, so the effective loss rate in a driven discharge is lower, and the $10^7\,\mathrm{s^{-1}}$ used in the two-cost table is an optimistic assumption that Gate 01 must replace with a measurement.

For $n_e=10^{18}\,\mathrm{m^{-3}}$ and 2 eV electrons the Debye length is 10.51 μm, with about 4,867 particles per Debye sphere. The electron plasma frequency $\omega_{pe}=\sqrt{n_ee^2/\epsilon_0m_e}$ is 5.641 × 10¹⁰ rad/s, or $f_{pe}=$ 8.979 GHz: in the collisionless limit, below it the layer reflects, above it the layer transmits; at sea level collisions blur that cutoff into absorption, and millimetre waves sit well above it. The fundamental electron cyclotron coefficient is 27.99249 GHz/T:

At sea-level pressure, electron-neutral collisions can broaden or suppress resonance: the collision rate exceeds these angular frequencies, so absorption is collisional rather than resonant, and resonant heating becomes a low-pressure, high-altitude mechanism. A matching-field contour does not establish a stable closed plasma sheet. Keep three frequencies separate: cyclotron frequency depends on field, plasma frequency on density, and molecular oxygen's absorption near 60 GHz (ITU-R P.676) is a property of the molecule that heats gas rather than making plasma. Established physics

Poynting flux at the window

Energy reaches the air as Poynting flux, $\mathbf S=\mathbf E\times\mathbf H$. One megawatt through a 0.1 m² window is 10 MW/m², a free-space RMS field of 61.4 kV/m, far below the megavolt-per-metre fields at which sea-level air breaks down unaided. Starting the discharge therefore needs focusing, seeding or a separate initiator; sustaining it, once free electrons exist, is the part that costs power. Established physics

Current scales

These are dimensional checks, not operating points. For a uniform useful volume of 50 m³ and the 98.1 kN reference load, $J=F/(BV)$ gives 392.4 A/m² at 5 T or 1,962.0 A/m² at 1 T. In a uniform, perpendicular, coherently directed interaction $F_B\approx JBV$; that is a force estimate, not an electrical power budget. Conductivity, current closure, return currents and opposing regions determine whether the useful integral can be sustained. Magnetic pressure is field stress, not available vehicle thrust.

How large can the active region become?

The baseline actuator is a thin layer under the annulus. The ambition is larger. Three regimes, each harder than the last. Engineering hypothesis

Regime I — active propulsive volume. Weakly ionized regions under the annulus couple to neutral air; everywhere else the air is ordinary. Lift is momentum flux beneath the hull; control is its distribution around the ring. This is where the program begins. Experimental question

Regime II — full-body plasma envelope. The active region grows around much of the aircraft and the same system participates in propulsion, force allocation and near-body flow. Boundary-layer behaviour, drag, heating and shock interactions become research questions, not assumed benefits. Plasma actuators have altered boundary layers at laboratory scale and magnetic fields have moved bow shocks in hypersonic test flows; neither establishes what a metre-scale envelope does to an aircraft. Simulation target Unverified regime

Regime III — field-dominated cavity. Electromagnetic and plasma pressure would reorganize the gas around the hull. Hypothetical. Requires experimental validation. The magnetic pressure $B^2/2\mu_0$ at 0.5 T is 99.47 kPa, about one atmosphere, which is why the idea is tempting. The reason it is hard is equally simple: a magnetic field exerts no force on neutral molecules, and in sea-level air the ions are collision-dominated, not magnetized. It belongs to the research horizon. Unverified regime

Build the sector before the aircraft

XR-1 begins as a stationary piece of the propulsion system.

The first experiment is not an aircraft. It is one sector on a thrust stand.

Sector S03, full scale, bolted to a six-axis balance in a qualified laboratory with high-voltage, RF, magnetic-field, cryogenic and thermal safeguards. Everything it does can be measured: the power going in, the plasma it makes, the force on the stand and the air leaving it. Progress means passing the next physical gate, not making the apparatus more elaborate. Experimental question

1. GATE 01

   Plasma

   Can the required electron density and conductivity be created in atmospheric air at acceptable power?

   MEASURE electron density · electron temperature · conductivity · spectral emission · attachment and recombination · spatial uniformity · sustaining power, continuous and pulsed

   This gate answers the two-cost table with numbers instead of assumptions, with repeatable diagnostics and stated uncertainty.

2. GATE 02

   Force

   Does energizing the magnet produce a repeatable force correlated with the predicted J × B direction?

   CONTROLS B-off · field reversal · current reversal · dummy loads · cable-force controls · thermal controls

   Explicitly separate magnetic thrust from EHD, thermal buoyancy, vibration, cable forces, corona wind and measurement artifacts. B-off and reversal controls distinguish ion wind and thermal effects from the intended magnetic contribution; bound cable, ground, nearby-structure and pressure forces. A single thrust-stand reading is not enough. Glowing plasma does not prove thrust. Plasma luminosity is not evidence of useful coupling.

3. GATE 03

   Neutral coupling

   Is electromagnetic force becoming bulk neutral momentum?

   MEASURE PIV and schlieren · wake velocity · mass flow · pressure field · momentum flux · temperature

   Map the neutral momentum flux over a closed control volume and compare it with the force on the balance. No mechanism is validated merely by naming the flow.

4. GATE 04

   Efficiency

   How much electrical input becomes directed atmospheric kinetic power? This is probably the single most important number in the program.

   MEASURE directed-air kinetic power · every electrical input · effective area and loading · heat and other losses, with uncertainty

   Newtons per kilowatt is area- and loading-dependent; numerical efficiency bands are screening scenarios, not universal viable/dead thresholds. Repeat across operating conditions and seek independent reproduction.

5. GATE 05

   Control

   Can multiple sectors produce controlled vector thrust?

   TEST differential sector power · pitch · roll · yaw · vertical force · lateral translation · dynamic stability

   Include actuator limits, transients, saturation, failed sectors and safe loss of power on a constrained rig. This is the gate at which the allocation diagram in section 11 becomes either a control law or a drawing.

Useful thrust is not the first gate. Useful thrust per electrical watt is.

Only after these gates close does the aircraft become the experiment.

Then tethered lift, then an aircraft whose power, mass, cooling and reserves close together. The decisive result throughout is measured neutral-air momentum and kinetic power against complete electrical input. If it is useful and reproducible, it earns the next engineering step. If it is not, this page will say so.

Ten megawatts changes the aircraft

The power architecture is jet fuel → turboshaft → generator → DC bus → actuator sectors. Nothing exotic, and nothing small. Design target

An hour of hover burns almost three tonnes of fuel.

That is the honest shape of the machine: about 33 MW of fuel power flowing through the airframe to deliver 10 MW to the bus and, at the reference point, at least 2.2 MW to the air. It is a heat engine with an electromagnetic output stage. The numbers make XR-1 an engineering study rather than a free lunch, and they point at where the leverage is. Design target

• Generator specific power. Aircraft electric machines are a design task in their own right; NASA's electric-machines work shows why installed efficiency, specific power and cooling dominate. Shaft rating is not net electrical output.
• Turbine efficiency and conversion. Generator, gearbox, rectifier and converter losses all sit between fuel and bus.
• Cooling. Every lost watt becomes heat inside a ten-metre airframe that must also hold cryogenic magnets.
• Transient power and startup. A 15–25 MW peak is an unclosed study target, not installed capacity. Two 5–7 MWe sets provide 10–14 MWe, so a peak needs more capacity or a specified buffer. Engine-out hover needs real reserve: four 2.5 MWe units leave only 7.5 MWe after one fails. Breakdown and magnet ramp are startup loads of their own.
• Superconducting loads. A superconducting magnet is not a power source. It stores field energy and returns it; superconducting magnetic energy storage (SMES) is storage, not generation, and a field energy density $B^2/2\mu_0$ is not a system energy density once winding, structure, cryostat and quench protection are counted. Refrigeration, controls and AC losses remain operating loads.
• Signatures. Intake and exhaust momentum and turbine thrust must be measured separately; heat, exhaust and acoustic signatures remain part of the system.

For the same 10 t / 10 m point, 15–30% all-in efficiency implies 7.383–14.767 MW of electrical input. Across 10–12 t and 10–20 m effective diameters, ideal power spans 1.1075–2.9117 MW; combining that range with 15–30% gives 3.692–19.411 MW. The same power budget cannot cover every geometry and efficiency combination.

Control without surfaces

If the sectors can be driven independently, the distribution of force around the ring becomes the control surface. Segmented thrust vectoring is a research objective of the program, not a demonstrated capability; Gate 05 decides it. Engineering hypothesis

Each sector could vary its RF excitation power, local conductivity, current amplitude and phase, traveling-field phase, trim-coil field, duty cycle and pulse timing. Not every variable can be commanded arbitrarily or instantaneously: the sectors share power, share the magnetic topology and interact through the gas between them. Design target

The pilot does not fly the plasma

The pilot commands vehicle motion. A hierarchy of control systems translates desired forces and moments into a spatial electromagnetic interaction pattern, in the spirit of fly-by-wire, except that the actuator is a distributed plasma and field system. Inner loops regulate measured current and inferred plasma state; a constrained sector allocator distributes commanded forces and moments; an outer loop closes on vehicle motion. Allocation must include moment balance, saturation, reserves and failed sectors. Design target

The controller must continuously estimate an unobservable high-dimensional plasma state from incomplete sensors, then choose the sector commands whose predicted effect best matches the demand within every limit. Interferometry, emission spectra, probes, electrical measurements, pressure and airflow sensors give calibration- and model-dependent state estimates; inertial units measure vehicle motion. Offline coupled plasma-chemistry, electromagnetic and gas-flow models inform reduced-order onboard estimators, which must be shown to track the full models before they are trusted in a loop. The same class of problem, on a very different plasma, was solved on a tokamak with a learned controller in 2022 (Degrave et al.). The transfer is of method, not of result. Simulation target

The propulsion problem is inseparable from the computation problem.

Why this is difficult

RF carrier or switching frequency is not feedback bandwidth. Sensor latency, gas response, stability margins, thermal and current limits, magnet quench and safe loss of power set the achievable rates. Modern control cannot rescue poor coupling. Established physics

What would kill the idea?

The architecture makes claims that can fail. These are the ways it fails.

1. If atmospheric conductivity requires substantially more power than useful momentum transfer can ever recover, the architecture fails.
2. If magnetic force primarily accelerates charged species without efficient momentum transfer to neutrals, the architecture fails.
3. If discharge instability prevents a large uniform interaction volume, the architecture fails.
4. If achievable magnetic field and current density produce insufficient body-force density at practical mass, the architecture fails.
5. If thermal management overwhelms aircraft mass and drag budgets, the architecture fails.
6. If sector forces cannot be dynamically controlled without destabilizing the plasma, practical flight control fails.

These are experiments, not philosophical questions.

What would change the program?

A research organization follows the measurements. These results would move it, in either direction.

1. A measured step-change in atmospheric conductivity per ionization watt.
2. Unexpectedly strong neutral coupling: more wake momentum per unit J × B than the collisional models predict.
3. Stable distributed discharges at atmospheric pressure over areas the size of a sector.
4. Higher-than-expected newtons per kilowatt on the sector thrust stand.
5. Major advances in HTS field strength per unit magnet mass.
6. Major advances in generator specific power.
7. Traveling-field geometries that materially reduce Joule losses.

Each gate on the test program maps to one or more of these lines. A negative result is a result; it will be reported here as prominently as a positive one. Experimental question

The problem is older than the hardware required to solve it

People could conceive of pieces of this machine decades ago. They could not control, power or simulate it. What follows is documented work by other people and other institutions: the technical ancestry of the problem, not an institutional lineage. Lorentz Aerospace inherits none of it, claims none of it, and implies no endorsement by any of it.

1. 1832Faraday at Waterloo Bridge. A copper wire on the parapet, electrodes in the Thames, and an attempt to measure the voltage induced by the river moving through the Earth's field: the first magnetohydrodynamic experiment. Bakerian Lecture, Phil. Trans. 1832
2. 1937Hartmann's "Hg-Dynamics". Laminar flow of a conducting liquid in a magnetic field, and the dimensionless number that still carries his name. Kgl. Danske Vid. Selsk. 15(6)
3. 1942Alfvén's waves. "Existence of electromagnetic-hydrodynamic waves"; the 1970 Nobel Prize in Physics followed for founding magnetohydrodynamics. Nobel Prize citation
4. 1955Kantrowitz founds Avco Everett. Arthur Kantrowitz establishes the Avco Everett Research Laboratory in Everett, Massachusetts, for reentry gas dynamics; shock tubes, MHD power generation and laser propulsion follow. Dartmouth archives, Kantrowitz papers
5. 1958Resler and Sears, "The prospects for magneto-aerodynamics". The first systematic look at what electromagnetic forces might do to the air around a flight vehicle. J. Aeronautical Sciences 25:235
6. 1959Ziemer moves a bow shock. An imposed magnetic field increases the bow-shock standoff distance in a laboratory hypervelocity flow: the germ of every later claim about fields reshaping near-body flow. ARS Journal 29(9):642
7. 1960sNASA Langley's MHD accelerators. A sustained program accelerates caesium-seeded nitrogen with J × B in segmented Faraday channels; one 2.54 cm channel triples flow velocity, 2 to 6 km/s, at 1.15 T and about 360 kW, and a 20 MW facility follows. Arnold Engineering Development Center runs parallel programs. NASA TP-2003-212285, historical review
8. 1961–62High-field superconductors. Niobium-tin at Bell Labs, then niobium-titanium; superconducting magnets stop being a curiosity. History of Nb-Ti (Lee)
9. 1963Avco Everett Mark V. The first large MHD generator that worked: 32 MW for seconds. Plasma, magnets and power conversion at industrial scale, in a single building. NASA TP-2331
10. 1964The first gyrotron. Invented, designed and tested in Gorky, USSR: the millimetre-wave source fusion laboratories would later push to a megawatt. Nusinovich, Thumm and Petelin, 2014 review
11. 1964–67King's magnetohydrodynamic propulsion apparatus. James F. King, Jr., of Winston-Salem, files on 30 September 1964 and is granted US 3,322,374 on 30 May 1967: ionize the surrounding air, drive it with a traveling polyphase magnetic field, steer with segmented ionizer sectors. The architecture appeared before the enabling hardware. US 3,322,374 · prior-art note
12. 1969Thomson scattering settles an argument. A Culham team measures electron temperature on the T-3 tokamak with a laser; plasma diagnostics become quantitative. Peacock et al., Nature 224:488
13. 1985The first commercial FPGA. Deterministic control at the sensor rate becomes something you can buy. IEEE milestone
14. 1987Superconductivity above liquid nitrogen. Wu, Chu and colleagues report a 93 K transition in a yttrium-barium-copper oxide; REBCO is born. Phys. Rev. Lett. 58:908
15. 1991Digital plasma control. DIII-D regulates tokamak plasma position and shape in real time; the loop in Fig. 23 has an ancestor. Ferron et al., IEEE SOFE 1991
16. 1992Yamato-1 goes to sea. The first superconducting MHD-propelled ship makes its trial runs in Kobe harbour in June 1992: J × B thrust in seawater, a conductor nature supplies for free. Sasakawa et al., OSTI 135352
17. 1999The ionization cost of cold-air MHD. Macheret, Shneider and Miles argue that the energy cost of creating conductivity governs magnetohydrodynamic flow control in cold air, the constraint Gate 01 must measure. AIAA 99-4800
18. 2000sREBCO as an industrial tape. Coated conductor reaches production lengths; a magnet becomes a matter of tape and structure. IEEE Trans. Appl. Supercond.
19. 2006–07General-purpose GPU computation. CUDA moves multiphysics onto a desk. Announcement
20. 2010Plasma actuators reviewed. Dielectric-barrier discharges have measurably altered boundary layers at laboratory scale: effects real and small. Corke, Enloe and Wilkinson, Annu. Rev. Fluid Mech. 42:505
21. 2011Wide-bandgap power electronics. The first commercial silicon-carbide power MOSFET; megawatt converters get lighter and faster. Announcement
22. 2018An aeroplane flies on ion wind. MIT's solid-state aircraft: electrohydrodynamic thrust, no moving parts, sixty metres. Proof that electric fields can move enough air to fly something, and a calibration of how small that something is. Xu et al., Nature 563:532
23. 2021Twenty tesla from REBCO. MIT and Commonwealth Fusion Systems hold 20 T in a large-bore high-temperature superconducting magnet. MIT statement
24. 2022A learned controller shapes a tokamak plasma. Estimation and allocation of a high-dimensional plasma state, in the loop, on real hardware. Degrave et al., Nature 602:414
25. 2023A megawatt for a thousand seconds. ITER-class 170 GHz gyrotrons demonstrate 1 MW for 1000 s. ITER Newsline
26. 2026Magnetoplasmadynamic thrusters, sixty years on. Researched since the 1960s and never flown operationally, a lithium-fed MPD thruster reaches 120 kW in a JPL, Princeton and NASA Glenn test: Lorentz-force acceleration of plasma, in space propulsion. NASA, April 2026
27. NowLorentz XR-1. Nobody has measured useful, efficient, controllable electromagnetic momentum transfer to sea-level air at vehicle scale. That is the experiment. Experimental question

Every entry is a published result by named people; the links go to the sources. The implication is ours: the conductors, the sources, the diagnostics, the converters and the computation each arrived separately, mostly for fusion and space, and only recently did all of them exist at once at the scale the problem needs. Engineering hypothesis