Concept Paper  /  Speculative Engineering  /  2026

The Reactionless Drive —
A 60-Year Thought Experiment
That May Not Be Wrong

Can a completely sealed vessel move through deep space — gaining speed, changing direction, decelerating — with nothing expelled, nothing external, nothing but what is inside?

Closed SystemGyroscopic Propulsion Deep SpaceNo Reaction MassHypothesis
▼   scroll to explore
01 — The Concept

What this paper proposes — and what it does not claim

This is not a claim that reactionless propulsion has been achieved. It is a carefully constructed argument that a specific configuration of internal gyroscopic machinery has not been definitively ruled out — and that a simple Earth-based test could answer the question once and for all.

Classical physics is unambiguous: in a closed isolated system, the center of mass cannot move. Internal forces cancel. Newton's third law holds. This paper does not dispute any of that.

What this paper proposes is that a sufficiently complex choreography of nested gyroscopic masses — spinning up and spinning down in precise synchronization with their arc position — may produce an asymmetry between the power stroke and the return stroke that does not cancel cleanly. A residual. Tiny. But in deep space, over years and decades of continuous operation, even a tiny residual accumulates into real velocity.

“The question is not whether conservation of momentum holds. It does. The question is whether the accounting closes cleanly for a system of this specific complexity — where the effective inertia of the moving mass is itself cycling in synchronization with its arc position.”

The concept has been in continuous development — mentally — for 60 years. Every component exists today. The only question is whether the choreography produces what the inventor suspects it might.

System typeClosed — nothing crosses the vessel boundary during operation
Drive mechanismCounter-rotating gyroscopic flywheel clusters on three axes
Energy sourceElectrical — solar, nuclear, or other onboard source
Reaction mass expelledNone
External fields usedNone
Testable on EarthYes — pendulum string deflection test
StatusUnproven hypothesis — awaiting physical test
02 — The Foundation

Building the concept from first principles — no laws broken yet

01
The weight in the box

A person inside a sealed box moves a weight toward the rear. The box nudges forward. When the weight stops, the box stops. The center of mass of the entire system has not moved. Net displacement zero. No laws broken.

02
Replace the weight with a spinning flywheel

The center of mass still cannot move. But something has changed: the flywheel has opinions. It resists reorientation. Move it through an arc and it does not behave like a dead weight — it precesses, pushing back in geometrically specific ways determined by its spin state.

03
Cancel all reactions with counter-rotating pairs

Spin up a single flywheel and the box reacts. Add a counter-rotating partner: the reactions cancel. Extend this across all three axes. The box remains completely inert while internally very busy.

04
The rowboat analogy

A single oar on one side of a rowboat makes it spin. Two oars pulling symmetrically move it straight. The entire rowboat — sealed inside the box. The oars are the flywheels. The geometry of which side pulls harder and when determines what the box does.

05
Honest energy accounting

Every spin-up and spin-down event consumes electrical energy and generates heat. Regenerative braking recovers energy on spin-down. Magnetic bearings eliminate friction. This is not perpetual motion — it needs a power source and generates waste heat. Both are accounted for.

At this stage, no laws have been broken. Everything described is orthodox physics. The open question begins in the next section — precisely identified.

03 — The Drive System

Architecture of the oar-flywheel propulsion system

Three primary oar-flywheels arranged in a Y pattern — 120 degrees apart — at the center of the vessel. Each spins on a different plane at constant RPM during straight-line travel. RPM changes only for course corrections.

Fig. 1 — Y pattern top-down view — three oar-flywheels at 120° from center hub
SEALED BOX Oar A Oar B Oar C Oar A — vertical, front-back ↻ Constant RPM Oar B — 45° tilted plane ↺ Constant RPM Oar C — horizontal plane ↻ Constant RPM 120° spacing — all reactions cancel Box feels nothing from rotation alone

Every oar-flywheel has a counter-rotating partner. Every sub-flywheel cluster has a counter-rotating companion. The box experiences zero net torque — a neutral stage. The only forces that matter are produced during spin-up and spin-down transitions at the power stroke position.

Fig. 2 — Sub-flywheel cluster detail (single oar-flywheel, face-on)
const RPM Cluster A — 12 o'clock Cluster B — 8 o'clock Cluster C — 4 o'clock Sub-flywheel axes X axis spin Y axis spin Z axis spin 3 clusters x 3 axes = 9 independent flywheels per oar-flywheel

The three planes — animated

All three oar-flywheels run simultaneously at constant RPM. Green pulses show sub-flywheel clusters at the power stroke position — peak gyroscopic energy.

Fig. 3 — Three oar-flywheels animated, constant RPM, three planes
04 — The Power Stroke

How the sub-flywheels create the asymmetry

Power stroke — front to rear (0° to 180°)

Sub-flywheels spin up from zero. Peak at 90 degrees — maximum gyroscopic energy, maximum resistance to arc movement. Spin back to zero at 180 degrees. Gyroscopically heavy throughout this arc.

Return stroke — rear to front (180° to 360°)

Sub-flywheels remain at zero. Dead weight only. No gyroscopic activity. The return stroke is physically not the same situation as the power stroke.

Fig. 4 — Sub-flywheel spin speed profile — bell curve peaking at 90°
zero max arc position — one full 360° rotation 0° FRONT 90° PEAK 180° REAR peak gyroscopic energy spinning up ↗ ↘ spinning down

Three sub-clusters per oar-flywheel means three power stroke peaks per revolution. Three oar-flywheels across three planes means nine power stroke events per cycle. No dead spots.

05 — The Open Question

Where the physics becomes genuinely uncertain

Status: unresolved — awaiting physical test

Everything described so far is orthodox physics. The open question begins here, at the boundary between confirmed mechanics and genuine uncertainty.

Conservation of momentum is not in dispute. The center of mass stays put. This is one of the most verified principles in all of physics.

The question is more subtle: does a gyroscopically active mass moving through an arc produce the same reaction on the box as an identical dead weight moving through the same arc — and does that difference cancel perfectly over a full cycle in a system of this complexity?

The answer to the first part is demonstrably no. A spinning flywheel moved through an arc creates more resistance than a dead weight moved through the same arc. The gyroscopic rigidity is real and measurable.

The answer to the second part — whether that difference cancels perfectly in this specific nested multi-axis configuration — has not been definitively answered.

“The law is not in question. The question is whether the accounting closes cleanly for a system where the effective inertia of the moving mass is itself cycling in precise synchronization with its position in the arc.”

The relativistic footnote

In relativistic mechanics, E=mc². The rotational kinetic energy stored in a spinning flywheel has a mass equivalent. In this system the effective mass at the oar-flywheel tip cycles between slightly higher on the power stroke and rest mass on the return. The implications of this cycling mass on momentum accounting — performed thousands of times per minute — have not been formally analyzed for this configuration.

Fig. 5 — Three oar-flywheels in 3D perspective — distinct planes
06 — Research Context

What has been studied — and what gap this concept addresses

The Gyroscopic Inertial Thruster (GIT) concept has generated over 278 academic papers and more than 536 patents since the 1960s. The general conclusion of review literature: gyroscopic forces cancel over a full cycle in the configurations tested.

Those configurations involve relatively simple arrangements — a single gyroscope on an arm, contra-rotating pairs, eccentric masses with variable radius — with one to three degrees of freedom. Notable contributors include Eric Laithwaite's Royal Institution work (1974), Sandy Kidd's forced precession research, and Roy Thornson's eccentric mass concepts. None achieved verified net thrust in controlled testing.

On the solid state side, the Mach Effect Thruster — developed by physicist James Woodward and funded by NASA NIAC in 2017 — proposes mass fluctuations in piezoelectric stacks can produce propellantless thrust. Independent replication at TU Dresden found results in the sub-micronewton range, but increasing measurement sensitivity found the claimed effect decreased, with vibration artifacts likely contributing. Unresolved.

1
Continuously rotating oar-flywheels

Prior GIT designs use reciprocating or oscillating mechanisms. The oar-flywheel concept uses continuous rotation at constant RPM — the return stroke is a gyroscopically dead portion of a continuous cycle, not a separate mechanical reversal.

2
Nested 3-axis sub-flywheel clusters

Prior designs use single or dual-axis gyroscopes. This configuration uses three orthogonally nested flywheels per cluster, three clusters per oar-flywheel, across three oar-flywheels on distinct planes — 27 independently controllable sub-flywheels whose cross-axis interactions have not been formally analyzed.

3
Bell curve synchronized spin profile

The spin-up and spin-down of each sub-flywheel cluster is precisely synchronized to arc position, peaking at exactly 90 degrees — the maximum leverage point. This specific timing relationship is not addressed in reviewed literature.

“The general conclusion applies to the configurations that have been studied. The specific nested multi-axis variable-spin-state choreography described in this paper has not been identified as a tested configuration in available literature. The door is not closed. It is simply not yet open.”

07 — The Test Method

A simple Earth-based experiment that answers the question

No laboratory, no grant, no exotic instrumentation. A box, some strings, a beam, and a careful eye for sustained deflection.

01
Establish zero reference

Hang the test box from strings attached to a rigid overhead beam. With all systems off, all strings hang perfectly vertical. This is the zero reference.

02
Run the system

Activate the internal drive. Box vibration and oscillation are expected and not interesting — they are noise. What matters is the sustained average string position over time.

03
Look for sustained average deflection

If the strings maintain a consistent average angle away from vertical while the system runs — not oscillating around zero, but displaced from zero — that is a result. A pendulum on long strings is sensitive to extraordinarily small sustained forces.

04
Stop, return, repeat

Stop the system. Strings return to vertical. Start again. Deflection returns if real. This repeatability distinguishes a genuine result from vibration artifact.

The test is self-validating in both directions. Positive result — deflection while running, return to vertical when stopped, repeatable — is compelling evidence. Null result is also valuable after 60 years of wondering.

String deflectionHigh-speed camera, angle measurement, long strings for sensitivity
Internal monitoring3-axis accelerometer logging — correlates internal activity with detected motion
Timing controlComputer-controlled sub-flywheel choreography — contact switches insufficient
EnvironmentIsolated from building vibration — concrete floor preferred
08 — The Solid State Alternative

A second path — and possibly the simpler one

An accelerometer converts physical movement into an electrical signal using a tiny internal proof mass. The logical inverse: a device that converts an electrical signal into directed physical movement. Solid state. No moving parts.

Piezoelectric actuators already demonstrate the principle — apply voltage, the crystal physically deforms. MEMS actuators produce controlled microscale movement from electrical signals. Both exist and work reliably. What does not yet exist — to this author's knowledge — is a device purpose-engineered to produce directed net linear movement from an electrical input at a scale useful for propulsion.

“If the solid state path is viable, every advantage of the mechanical system is amplified. No moving parts. No bearings. No wear. No mechanical timing. The path to the stars may be a crystal responding to a waveform.”

Moving partsNone
Friction lossesNone
Timing controlWaveform — no mechanical synchronization required
ScalabilityStack more units for more thrust — fully modular
Current statusConceptual — no known implementation at propulsion scale
09 — Mission Architecture

From a box on strings to an interstellar probe

If either system produces even a tiny net thrust — the implications for deep space travel are profound. Not because the thrust is large, but because of what happens when a tiny continuous force operates in an environment with no drag, no friction, and unlimited time.

Continuous acceleration — no fuel limit

In deep space there is no drag. The only consumable is electrical energy. Even millinewtons accumulate to velocities that dwarf chemical propulsion given sufficient time.

Deceleration — same system reversed

Flip the effective thrust direction by changing axis emphasis. No additional propellant. No staging. No expendable hardware.

Direction control — axis emphasis

Emphasize one oar-flywheel axis over others. The vessel curves. Return to balanced RPM — straight line resumes. Full three-dimensional navigation with no thrusters and no propellant.

Inner solar systemExternal solar cells — proven, reliable, no moving parts
Deep spaceSmall nuclear source — electrical output drives the system
Waste heatDirected rearward — photon emission provides additional thrust
Energy recoveryRegenerative braking on all spin-down events

The AI navigator

A vessel of this type does not need human biology on board for the cruise phase. The onboard AI operates in deep sleep low-power mode during cruise. Periodic wake cycles for navigation correction, system health checks, and compressed status bursts back to Earth. Continuous operation near a star for science collection.

“Humans will almost certainly never travel between stars. Biology is too fragile, time too long, distances too vast. But intelligence need not be biological. The natural explorers of the galaxy are patient, do not require oxygen, and do not get bored on a 400-year journey.”

The shape of the vessel does not matter

The drive system inside a cube functions identically inside a sphere or any other geometry. The sphere is structurally optimal for space. And a certain other shape has appeared in human imagination for a century — now with a plausible internal explanation.

Fig. 6 — Sealed vessels in formation: box, sphere, flying saucer — same drive, different shells
NO EXHAUST — NO REACTION MASS — SEALED
10 — A Personal Observation

North Idaho — late afternoon — approximately 2010

This section is included not as evidence for any claim, but as context for why the question in this paper matters personally to the author. Included because the engineering details are precise enough to be worth recording.

While driving home on a rural two-lane highway in North Idaho — good afternoon light, clear visibility — the author observed an object moving slowly and deliberately above a set of high-voltage power transmission lines. The sighting lasted long enough to accelerate to approximately 100 mph to close the distance. The following observations were made in real time by someone with 46 years of experience in electronics, networking, and antenna systems:

Shape: Elongated oval — football-shaped in side profile, likely circular from the front or rear. Estimated 40 to 55 feet in length. Clearly three-dimensional based on light reflection behavior.

Surface: Highly reflective silver. Smooth. No visible seams. No marker lights, no FAA lighting, no illumination from portals. Purely reflective.

Portals: A row of evenly spaced rectangular openings with rounded corners along the midsection — approximately 3 to 4 feet wide and 2 to 3 feet tall. Dark black. No portal at the very front or very rear of the row.

Top: A cluster of vertical antennas of varying heights — 3 to 18 feet — grouped within approximately 6 feet of each other at the center top. The mix of lengths is consistent with omnidirectional antennas for different frequency bands. A recognizable functional antenna array to anyone who installs and maintains them professionally.

Bottom: A dish antenna approximately 3 to 5 feet in diameter, mounted on what appeared to be a 2 or 3 axis gimbal, pointed forward in the direction of travel. Point-to-point communications, data relay, or forward observation would be consistent with this configuration. There were transmission line towers in the direction of travel.

Flight behavior: Moving slowly. Perfectly tracking the power transmission lines — directly above the wires by approximately 5 feet. No drift. No wobble. Controlled straight-line movement. Completely silent.

Assessment: Not a balloon. Not a conventional aircraft. Not a known drone type. After 15 years of reflection: something functional, purpose-built, operating with intent on that stretch of road above those power lines. What it was built by, or when — the author does not know.

Animated reconstruction — driver's perspective — North Idaho highway — late afternoon

The connection to this paper is direct. A person who has spent 60 years thinking about how a sealed vessel might move without reaction mass — and who once followed one at 100 mph down a two-lane Idaho highway — has a particular kind of stake in whether the question is worth asking.

The answer is yes. It is worth asking. And it is worth testing.