Power-balance laboratory · D–D / D–T / D–³He

FusionSim

Every fusion concept is an energy ledger: joules invested in fast ions or hot plasma against joules returned by fusion. FusionSim runs that ledger live for three machines — the electrostatic fusor you can build in a garage, the gridless orbit trap that swaps the fusor's grid for a planetary orbit, and the pulsed field-reversed configuration that is the most credible bet for small-scale net electricity. Move the sliders. Watch where every joule goes.

The gap, on a log scale

gain vs. what breakeven demands

Each row shows a device's fusion gain (filled dot: fusion energy out per unit of energy put into the plasma or beam) against the gain it would need for net electricity given its own recovery scheme (open ring). The fusor and orbit-trap rows assume a 40%-efficient thermal plant bolted on; the FRC's finish line moves with your recovery-efficiency slider — that is the whole trick.

Bench 1 — Electrostatic fusor

steady state · D–D beam-target

Ions fall through the cathode voltage and recirculate until something eats them: a grid wire, a charge-exchange collision with background gas, or Coulomb upscattering. Fusion has to happen before that — and the odds per pass are about one in ten billion.

Vacuum chamber = anode (0 V) Cathode grid, −V₀ ~95% open area — until it isn't Star-mode core: converging D⁺, where fusion happens Recirculating ions D₂ inlet, a few mTorr To vacuum pump HV feedthrough (−V₀)

The whole machine: a spherical shell held at ground, a centimeter-scale wire cage at tens of kilovolts negative, and dilute deuterium gas. Ions born in the gas fall inward, overshoot the center, and oscillate through the core — the pink "star" — until a grid wire or a charge-exchange collision claims them. Elegance is the appeal; the grid and the gas are the verdict.

Presets

Machine

Neutron rate
D–D neutrons / second
Fusion gain Q
fusion power ÷ input power
Input power
Fusion power
Mean core transits per ion
Orders of magnitude to breakeven
Where each ion's energy ends up

Grid interception deposits the ion's full energy as heat on the wires; charge exchange hands the energy to a fast neutral that hits the wall. Not modeled (both make reality slightly worse): power carried away by secondary electrons and X-rays, and the modest fusion contribution of fast neutrals after charge exchange.

Bench 2 — Pulsed FRC with direct recovery

per-pulse ledger · the Rider loophole

Compress a field-reversed plasmoid, burn for milliseconds, expand, and push the energy back into the circuit as electricity. Because the pulse is over before the plasma fully thermalizes or radiates away, marginal fusion gain can still mean net electricity — if recovery efficiency is extraordinary. This is Helion's wager, reduced to its arithmetic.

Capacitor bank — pulse out, electricity back formation acceleration compression coil = generator formation acceleration 1 2 3 4 5 6 form accelerate merge & compress burn ~ms expand recover

One pulse, left to right in time: plasmoids formed at both ends are electromagnetically fired inward, merge at the center, and are squeezed by the compression coil until fusion conditions hold for a few milliseconds. The fusion-heated plasmoid then expands against the field, pushing flux — and current — back through the same coil into the capacitor bank. The coil is both the piston and the generator; there is no turbine anywhere.

Presets

Fuel

Pulse

Engineering gain Qeng
electricity out ÷ electricity in, per pulse
Net electric power
at your pulse rate
Fuel gain Qfuel (Efus/Eplasma)
Fusion yield per pulse
Plasma energy at peak
Switched energy per pulse
Recovery η needed to break even
Gross fusion power
Fate of one pulse's energy (switched + fusion)

The vertical marker is the energy you paid for the pulse. If the two electricity segments together reach past it, the machine is a power plant; if not, it is a very expensive capacitor heater. Note how the marker chases the bar as recovery efficiency changes — breakeven here is a property of the circuit, not just the plasma.

Bench 3 — Gridless orbit trap

single-ion orbit tracer · Orbitron-style

Bench 1's verdict was the grid and the gas. This bench deletes the grid: a bare charged spike sits at the center, and an ion given enough sideways momentum orbits it like a comet — never touching metal. One test ion's equation of motion is integrated live below for your settings. The questions that decide the concept: how deep is the well, how long does the ion recirculate, and what finally claims it?

The actual integrated trajectory, not a cartoon. In a perfect spherical well the orbit is a closed Kepler ellipse that would repeat forever; in the cylindrical (Orbitron-style) well it precesses into a rosette. Either way nothing solid stands in the ion's path — so the losses move elsewhere: a charge-exchange collision with background gas, the slow drift of the periapsis into the spike, or Coulomb upscatter.

Presets

Well geometry

Trap

Neutron rate
D–D neutrons / second
Recirculation lifetime
one ion's expected survival
Peak ion energy (periapsis)
Orbital period
Mean orbits before loss
Periapsis clearance
Fusion gain Q
Orders of magnitude to breakeven
What finally claims the ion
Neutron rate vs pressure — the charge-exchange ceiling

Past a certain pressure every added gas molecule is a fusion target and an ion thief in equal measure, and the neutron rate stops rising: that plateau is the ceiling of any gridless trap as a neutron source at fixed voltage and current. The thin curves re-run the same orbit at half and double the well voltage — raising the voltage, not the pressure, is the only way up.

Why temperature is the whole game

Maxwellian reactivity ⟨σv⟩, Bosch–Hale 1992

Fusion reactivity climbs three to four orders of magnitude between 5 and 50 keV. D–T is ~100× easier than D–D at fusor-relevant energies — which is why any marginal concept gets evaluated on D–T first — while D–³He only wakes up above ~40 keV but pays out almost entirely in charged particles that direct recovery can capture.

Model notes & honesty

assumptions you should know about