We frame the tungsten filament as a heat-bit-photon exchanger: a finite-mode device that translates thermal-bath input into a stream of irreversible bit erasures (Landauer heat) plus an output photon current. Three identifications from the mode structure of the atom are made. (1) Channel-counting efficiency (Section 6): the visible radiant efficiency obeys eta = nᵥal / (nᵥal + alpha^-1 ln 2), giving etaW = 5. 94% for tungsten (nᵥal = 6) and matching the observed radiometric scale. The alpha^-1·ln 2 loss load reflects ~137 irreversible bit erasures per emission cycle. (2) Universal vacuum-coupling rate (Section 4. 1): the fractional residue delta = 0. 036 per sector acts as the trigger cost of the exchange — the rate at which each sector talks to the vacuum bus, common across CPⁿ levels and tied to the same residue that surfaces in Paper XIV's CMB / H0 analysis. (3) Dynamical reading of ln 2 (Section 5b, new in v3): a complementary, hypothesis-level identification of ln 2 with the Lyapunov exponent of an iterated Born projection at the Bernoulli limit r=4, via the logistic map p₍+₁ = r pₙ (1-pₙ). Stated with three explicit postulates (real phase fixing, recycling step, interval-invariance bound). The reading yields a parameter-free prediction: a period-doubling cascade on heat-up with universal Feigenbaum scale deltaF = 4. 6692, testable on a tabletop tungsten filament under DC bias plus AC ripple at f0. Uniqueness caveat: deltaF is a class signature (unimodal maps with quadratic maximum), not a Born-specific marker; observation establishes class membership, not the Born interpretation alone. A three-stage emission model (quark / proton / electron sectors) predicts documented inflection points in the emissivity curve (Section 7-8). Earlier v2 changes retained: Allen-Dynes lattice-buffer term removed (Section 6, no empirical preference across 15 metals) ; equal-valence Mo-vs-W falsification test (Section 10) ; companion concept-DOI 10. 5281/zenodo. 19446942 (246-bit vacuum-ADC cosmology, Paper XIV). Not peer reviewed. Prepared with the assistance of artificial intelligence.
Alexander Nachtigall (Thu,) studied this question.