FCBP-03: Signed Pressure–Flux Work, Variable-Radius Telescoping, Slow-Scale Critical Lift, Filter-Switch Defects, and Model-Cone Recurrence

Signed Pressure–Flux Work, Variable-Radius Telescoping, Slow-Scale Critical Lift, Filter-Switch Defects, and Model-Cone Recurrence

FCBP-02 proves the filtering route has a complete conditional compiler, but still lacks a global critical stack. The pressure–flux work architecture is attractive because it is sign-preserving by construction and gives an exact endpoint telescoping. The published theorem uses geometric radii r_k=θ^k r_0; the first question is whether the resulting summable weight is intrinsic to the telescoping itself or merely an artifact of the chosen geometric schedule. The answer: it is not intrinsic to the telescoping — this is the first genuinely non-summable schedule found within the FCBP architecture. This round generalizes the exact pressure–flux endpoint telescoping algebra from geometric radii to arbitrary strictly decreasing radii, proving an interior "variable-radius pressure–flux telescoping"; choosing r_k=r_0(k+1)^(-β) (1/2<β≤1) yields a finite total parabolic time but a divergent sum of telescoping weights, producing a genuinely non-summable "slow-scale critical lift window" — this cycle's first breakthrough. It also identifies the main compatibility problem: a single fixed physical filter preserves exact telescoping but loses scale-relative resolution as r_k→0; a scale-relative moving filter preserves critical resolution but breaks exact endpoint telescoping. Proves an exact moving-filter telescoping theorem, with an explicit positive filter-switch defect.