PULSED OFFSET GYRO
A mechanical device that simulation says produces a directional force bias. The next step is building one and finding out if reality agrees.
240,000+ simulation samples show a consistent 15.5 N directional bias with a 60x signal-to-noise ratio. Control runs show nothing. The simulation work is published openly with full data — but simulation isn't proof. Only hardware can answer the question.
Counter-rotating masses on an offset circular race track. Pulse the angular velocity — faster through the wide side, slower through the narrow — and the centripetal force integral doesn't cancel over a full revolution. The simulation shows a net directional bias on the frame.
Three counter-rotating pairs arranged at 120 degrees, pulsed at three cycles per revolution, produce a clean unidirectional force with zero off-axis components in simulation. The control runs (constant speed, zero pulse) show no bias. The effect scales linearly with pulse strength and vanishes when you turn it off.
This isn't reactionless thrust. Newton's third law holds — the reaction force goes into the race track surface. Think of it as controllable vibration-driven locomotion: time-asymmetric internal forces that create directional movement when coupled to a surface.
The simulation predicts a directional force bias. The physical hardware might not reproduce it. Friction, manufacturing tolerances, vibration coupling, thermal effects — any of these could eat the signal or create false positives.
If the force doesn't appear, that's still a result worth having. A null result with proper instrumentation tells us exactly where the simulation diverges from reality — which assumption breaks, which parameter matters. That data has value.
If the effect is not confirmed, a full null result report gets published — data, analysis, and conclusions. Nothing hidden.
I'm Gavin. Single parent. Engineer by trade — precision machining, fabrication, electrical installation, automotive diagnostics. Self-taught programmer — Python, C++, TypeScript, embedded systems, GPU compute. I built everything on this site.
No academic backing. No institutional funding. No lab. Just a home workshop, an AI collaborator called Nyx, and a simulation that keeps producing a number that shouldn't be there.
Every contribution goes directly into hardware, tooling, and instrumentation. The work is documented openly — you can follow it all on the journal.
If you think this is worth finding out, chip in.
Every contribution goes straight to hardware and tooling. No amount too small.
GET IN TOUCH
Want to know more about the research, the prototype design, or the physics? I'm happy to talk.
gavinviney@indigo-nx.com