Incapp.org · Propulsion

PMP:
A Solar Wind
Propeller

Pepin Magnetic Propulsion (PMP) is a magnetic traveling wave that accelerates ambient solar wind plasma rearward, producing thrust. It is not a solar sail, and the difference is not cosmetic — it changes what the vehicle can do. This page sets out the mechanism, the arithmetic, and the experiments that would prove it wrong.

John Pepin · [email protected]
Comment, correction and refutation all welcome.
The original proposition, June 2018: PMP.

The distinction everyone skips

Tell someone you want to use the solar wind for propulsion and they hear "solar sail." Then they explain, correctly, that you cannot outrun the wind you are pushing against, and the conversation ends.

That objection is sound. It applies to sails. It does not apply to this.

A sail is pushed by the medium. A propeller pushes on the medium, using its own power, and does not care what the medium is doing. A propeller aircraft in a tailwind exceeds the wind speed without difficulty. A paddle wheel goes upstream, downstream, or across. The current is something it works in, not something it rides.

PMP is the second kind of machine. The solar wind is reaction mass, not a push. That is the whole of what distinguishes Pepin Magnetic Propulsion from the magnetic sail, the electric sail, and every other solar wind concept it gets mistaken for.

Why a static magnet cannot do this. The Lorentz force qv × B is always perpendicular to velocity, so a static magnetic field does no work on a charged particle. It steers, it never accelerates. A permanent magnetic structure in the solar wind is therefore a magsail — it deflects, it gets drag, and it is capped at wind speed.

To add energy to the plasma you need an electric field, which means a changing magnetic field. The line is sharp: static field is a sail. Traveling wave is a propeller.

The mechanism

Think of a coilgun run backwards. A coilgun fires a sequence of coils inward to accelerate a projectile along an axis. Reverse the sequence so the wave travels outward, and instead of launching a slug you push on whatever plasma the wave passes through.

The induced currents in the plasma feel J × B from the traveling field and are driven rearward. The reaction arrives at the spacecraft through the field structure. In miniature this architecture exists and works — it is the pulsed inductive thruster, which accelerates plasma with a time-varying coil field and no electrodes. PMP is that principle at kilometre scale, running on ambient solar wind instead of injected propellant.

Which of the two 2018 geometries survives. The original essay proposed either a cylindrical field reflector that would gather and accelerate the wind, or an oscillating field that would instead knock particles, affecting a larger volume. The second is the one that holds up.

A design that gathers and accelerates is a ramscoop, and it pays for stopping the flow: net thrust requires exhaust velocity above the relative wind speed, and above that speed the device becomes a brake. An oscillating field acting on plasma passing through is a propeller, with no such ceiling in any direction. Everything below assumes the second architecture.

The arithmetic

Because the plasma passes through rather than being swallowed, the accounting is a propeller's, not a ramjet's. There is no term for bringing the flow to rest:

F = ṁ · ΔV
P = ṁ · ΔV · (vrel + ΔV/2)
η = vrel / (vrel + ΔV/2) where ṁ = ρ A vrel is the mass flow through the swept area

Two things follow immediately. Thrust never goes negative, so there is no speed at which the machine becomes a brake — upstream, downstream and cross-stream all work. And efficiency rewards a large mass flow given a small kick, which is exactly the design rule that makes a propeller beat a jet at low speed.

At 1 AU, taking 8 protons/cm³ with 4% alphas, a wind speed of 440 km/s, and a 100 km swept radius:

ΔV (km/s)Thrust (N)Power (MW)EfficiencyΔv/year*
204.31.90.9854 km/s
367.73.50.9697 km/s
5010.75.00.95135 km/s
10021.410.50.90270 km/s

*continuous operation, 2,500 kg vehicle. Scale by the square of the aperture radius: a 300 km structure gives 69 N and roughly 875 km/s per year.

What continuous thrust buys

Thrust that never stops permits a flip-and-decelerate profile: accelerate to the midpoint, turn over, decelerate the rest of the way. Arrival is at rest relative to the destination, with no capture burn and no propellant reserved for one. Transit times at the two apertures above:

Destination100 km aperture300 km aperturePeak speed*
Mars, 0.5 AU114 days38 days46 km/s
Ceres, 1.8 AU216 days72 days87 km/s
Jupiter, 4.2 AU331 days110 days132 km/s
Saturn, 8.5 AU1.3 years157 days188 km/s
Kuiper Belt, 40 AU2.8 years340 days408 km/s

*at the 300 km aperture. These assume wind density at 1 AU throughout, which is optimistic — density falls as the inverse square of solar distance, so outer-system legs would run at reduced thrust. Treat the Saturn and Kuiper figures as upper bounds.

A worry that turns out not to apply. The 2018 essay assumed strong fields would disrupt radio and proposed pulsed lasers for communication. That appears unnecessary. The plasma frequency at ambient solar wind density is about 25 kHz — five orders of magnitude below an S-band downlink — so conventional radio passes through the structure unimpeded. Blocking a 2 GHz signal would require compressing the plasma roughly ten billion times.

No propellant is carried. The delta-v is bounded by mission duration and reactor life rather than by a tank.

Why it must be a plasma magnet

Thrust scales with swept area, so aperture is the entire game. A static dipole reaches a radius proportional to the cube root of its magnetic moment, which means area grows only as the two-thirds power. Doubling the aperture costs eight times the moment. Reaching 30 km with ordinary coils would require something like 10¹⁰ times the moment of a laboratory electromagnet. That is not an engineering gap. It is the wrong curve.

Slough's plasma magnet escapes it. A polyphase antenna of a few metres drives currents in the surrounding plasma, and the structure inflates under its own self-repulsion until magnetic pressure balances the wind's dynamic pressure. Radius is set by drive power, not by onboard moment. Tens of kilometres become reachable with kilowatts.

It also happens to be a time-varying-field device already, which is what the mechanism requires. Coils were the wrong answer twice over.

The cavity argument

There is a reason the two concepts fit together that is better than convenience.

The ambient solar wind at 1 AU is strongly super-Alfvénic — roughly Mach 12 against an Alfvén speed near 36 km/s. Waves cannot propagate upstream in such a flow, so wave-driven momentum transfer has nowhere to work.

But pressure balance, B²/2μ₀ = ρv², implies an Alfvén speed inside the structure of v√2, which puts the interior Alfvén Mach number at 1/√2 — just below unity. This is not a tuning choice. It is a consequence of what pressure balance means.

So a plasma magnet carves out, by construction, a marginally sub-Alfvénic cavity inside a super-Alfvénic flow: a region where waves propagate and can couple, surrounded by one where they cannot. The structure is not merely the aperture. It is the cavity the mechanism needs, and it builds itself out of the wind it acts upon.

The inversion is the whole idea. In a magsail, the wind shapes the plasma. Here, the plasma shapes the wind.

Precedent. Alfvén wave pressure doing work on plasma is not exotic. It is a leading explanation for how the fast solar wind is accelerated in the first place. The mechanism proposed here is one the Sun already runs at stellar scale.

How to prove this wrong

PMP is not one claim but four, and they fail independently. Three can be killed on a bench without going to space.

  1. Does a traveling wave impart directed momentum to tenuous plasma? This is load-bearing. Everything else is engineering around it.

    A plasma source producing a drifting stream in a vacuum chamber, an RF antenna launching an axial traveling wave, and a retarding potential analyser or laser-induced fluorescence measuring the downstream ion velocity distribution.

    Fatal result: the distribution broadens without its mean shifting. That is heating, not acceleration — energy in, temperature up, momentum nowhere. It is what most wave-plasma coupling does, and if it is what happens here, the concept is dead.

  2. Does that momentum reach the hardware? Accelerated plasma is not the same as a vehicle feeling recoil. Field lines can slip and the reaction can vanish into the plasma's own dynamics.

    Torsion balance on the antenna assembly, measuring force directly rather than inferring it from ion data.

    Fatal result: the plasma accelerates and the balance reads zero.

  3. Does the axial wave destroy the current ring? The rotating field that sustains the structure and the traveling wave that does the pushing must coexist. Testable in particle-in-cell or MHD simulation before any hardware exists.

    Survivable if it fails — a frequency-separation problem, not a physics problem.

  4. Does net thrust beat bow-shock drag at scale? Simulation of the full structure in super-Alfvénic flow. Also survivable: a geometry question.

Controls, and why they matter more than the measurement

Thrust measurements in plasma chambers have an unhappy history. The EmDrive produced a real, repeatable, carefully measured force that turned out to be thermal expansion and interaction with the apparatus. Any positive result here deserves the same suspicion, applied in advance.

Three nulls, all cheap, and the first is worth more than a dozen positive runs:

Measure scaling, not a number

A single thrust reading proves little. The model makes four quantitative predictions, and a spurious force will not satisfy them together:

If a force is measured but does not scale with B/√ρ, the model is wrong even though something pushed.

And if it flies

A flight demonstrator must sit outside Earth's magnetosphere — Sun-Earth L1 or lunar distance — because there is no solar wind inside it and residual atmospheric drag in low orbit swamps everything. It needs solar power and months, not a battery and minutes: the same hardware that yields an unmeasurable millimetre in five minutes yields some 96 km of displacement in thirty days.

Magsail drag and solar radiation pressure both push anti-sunward and both exceed the expected thrust, so a vehicle simply drifting outward demonstrates nothing. The clean test is to reverse the wave every week. Drag is identical in both directions. Thrust changes sign.

What I am asking for

Criticism, in the specific. The arithmetic above is back-of-envelope and the assumptions are stated so they can be attacked. I am most interested in hearing from anyone who can say whether the wave coupling in Claim 1 is plausible at these densities, whether the cavity argument holds up under a real MHD treatment, and whether the bench experiment has already been done by somebody and failed.

If this is wrong, I would rather know which step is wrong than be told it sounds like a solar sail.

Prior work

None of the components here are mine. The arrangement is.