We mapped every combination of density and temperature around the frozen point — 667 coupled evaluations per posture through the full eight-constraint systems code (S80/S76 coupled chain). The answer: at the ratified near-thermal Mode-D envelope, 21% of the scanned plane runs the plant at positive net power (the retired hot-ion-era postures had spanned 46–52%).
Below the design point, the window is wide: density can fall 37% and ion temperature 38% before the plant stops closing — it degrades gracefully, shedding output rather than tripping a cliff. Above the point, the story is one-sided: Mode D rides the wall-stabilised βN = 5.0 envelope (S79, with the MHD decks S22/S25) at βN = 4.96 — the binding margin above the operating point is about 1%, and wall stabilization is required (no-wall operation gives no headroom at all).
The closure is profile-resolved, not asserted: the break-even H98 requirement (2.28 at the 48 keV operating reference; 2.38 on the 85:15 banking leg) is computed and emitted by the deposited chain (S76/KX-1D), and the 21% window is what survives all eight constraints simultaneously.
Operators buy windows, not points. A −37%/−38% graceful envelope means startup, ramp, and off-normal operation have somewhere to live; a mapped edge means control systems know exactly which boundary they defend. Publishing the one-sided upside is the marketing: it tells sophisticated readers we measured our room instead of imagining it.
| Feasible fraction of scanned plane | 21% (P_net > 0, βN ≤ 5.0) |
| Graceful room below point | −37% density / −38% temperature |
| Upside at no-wall limit | none (wall stabilization required, S79) |
| Margin to the βN 5.0 envelope | ≈ 1% (binding) |
| Grid | 29 × 23 = 667 evaluations / posture |