OpenCretin — the LightCell, playable
a recovered, source-complete 2-D sodium radiative-transfer converter · drag the knobs, the field re-solves live · every number references its source
This model is the real thing, recovered. The 2-D field below is computed by na_grid2d.zig → wasm (44×60 grid). Its source was reconstructed from the shipped binary by disassembly + behavioral-diff, and verified bit-for-bit to the original: temperature field max relative error 1.16×10⁻⁶, all 11 readout scalars 4.5×10⁻⁶, across four drive cases. Every physical constant is traced to primary literature (see the coefficient ledger). Source: engine/src/na_grid2d.zig · disassembly + oracle in web/recovered/.
—
—gas temperature (K) · blue band = cooled silicon across vacuum gap
Drive & geometry
fuel power core deposition2.0 kW
secondary pressure annulus density40 kPa
emitter radius R₂ envelope15 mm
length Lz50 mm
core R₁ = 5 mm · primary 1 kPa · silicon 400 K across the gap
model coefficients — assumptions, not derived
envelope emissivity ε transparent → low loss0.10
conductivity ×k_eff sets core ΔT1.0×
Steady state
useful fraction
—
matched / total crossing the gap
peak core T
—
on axis, mid-length
envelope T
—
gas-edge, sets thermal loss
819 / NIR share
—
of matched output
power crossing the gap — fuel ——
589 (matched) 819/NIR (matched) broadband thermal (parasitic)
Escaping-photon spectrum — what the PV sees whole device, across the gap
Photon-density spectrum — the trapped field click the map to probe a point

What this IS

  • A real 2-D (r,z) sodium converter: 44×60 grid, fuel core + secondary annulus across a radiative vacuum gap to cooled silicon.
  • Energy-balanced — temperature is solved from the per-cell balance, not assumed; closure reads 1.00.
  • Real emissivity ε on the envelope (Stefan-Boltzmann gap loss); the secondary pressure sets the annulus density.
  • Occupancy is a NESS rate balance (S = j/κ, never B_λ): per-cell Na populations → Voigt opacity → Holstein trapping → 589/819 emission.
  • Source-recovered + verified to 1.16×10⁻⁶ against the shipped binary. Every constant traced to primary lit.

What this is NOT

  • NOT the fully-coupled ideal — axial chemistry + flow along the tube and a real ε(λ,T) spectral wall are specified in DESIGN.md, not yet in this kernel.
  • NOT a fixed-ratio model — a pinned 819/589 keystone that rides 60 orders of magnitude is laundered; R must emerge per condition.
  • NOT the 1-D deck viz (physics.js), whose calibration anchor gives a flat R across 12 OOM — use it as an explainer, not ground truth.
  • The secondary chamber + recuperator are boundary-modeled, not co-solved; transport is escape-factor + radiative-diffusion, not short-characteristics.
  • A single converged fixed point — not time-resolved. The Na₂ NIR bands (∝ n², cool skin) are the next physics brick.

References — start here

The recovered engine: na_grid2d.zig · disassembly (WAT) · verification notes · behavioral oracle
The physics, ledgered: coefficient ledger (every number → primary source + DOI) · literature canon · no-equilibrium-smuggling doctrine
The deeper transport: the Λ-iteration CRETIN-class solver — GOAL.md, engine/, validation/ (run bun validation/test_r8_saha.mjs for the anti-Saha gate)
For an agent: read AGENTS.md first — orientation, build/run, the binding doctrine, the capability map.
what this is — and what's next

Now the silicon sits across a vacuum gap, so nothing conducts to it — the gas sheds its fuel power entirely as radiation, and closure still reads 1.00. Two channels cross the gap: the gas line/band emission (matched, what the PV converts) and the envelope's broadband thermal (parasitic, what heats the PV). The useful fraction is the spectral-efficiency knob — it climbs with drive here because line emission outpaces the T⁴ thermal as the gas heats.

Two coefficients are mine, not derived, and now exposed: the broadband ε (envelope + continuum greybody — lower = transparent = the vacuum-gap payoff, but the envelope runs hotter to compensate, and at ε≈0.05 it hits the sapphire ceiling) and the ×k_eff scale (which alone sets the peak core T — raise it and the core flattens). Transport is still escape-factor + radiative-diffusion, not short-characteristics; and the Na₂ bands (∝ n², favoring the cool skin) are the missing matched-NIR — still the next physics brick.