WP38: the charge is CKM-composite - q_u = kappa_u * V_ub exactly per sheet
WP37 established q_anchor = H_12.sin(phi) as a class constant (exact
for u/t, ~1% approximate for c). This package asks whether the
constants are CKM-composite. Checker checkers/wp38_charge_ckm.py,
results results/wp38_charge_ckm.json, 5/5 gates.
The exact u-class law (G1, G2)
Across all 232 viable mixed u-class sheets:
q_u = H_12.sin(phi) = kappa_u * V_ub,
kappa_u = 9.678574167082195
231/232 sheets sit within 9.2e-9 of the median kappa; exactly one outlier at 3.4e-6 (class 394_348_d02). The deviation does not correlate with chi2 (corr 0.14; chi2 is nearly constant 3.36-3.90), so fit noise is excluded: this is an exact per-sheet identity of the fitted ensemble, 17x sharper than WP37's class constancy (2.9e-5). The WP37 picture 'q_u class-constant' was the shadow of the sharper law 'q_u tracks V_ub per sheet'.
t- and c-class status (G3, G4)
- t-class: q_t = 1.125347e-6 constant at 8.4e-9; every CKM element is equally constant within the class, so CKM tracking is unresolvable - the const law stands as the best description.
- c-class: every candidate ratio (const, q/Vub, q/Vcb, q/Vtd) scatters at ~1% with >38% of sheets beyond 1e-4 - genuinely approximate, no hidden exact bulk law. WP37's typing confirmed.
Negative results (G5 + survey)
- The WP36 magnitude factor M in J.Du.Dd = c.sin(phi).M is NOT an edge-magnitude monomial: log-log regression over 232 u-class sheets returns all exponents ~0 (< 5e-4), consistent with M ~ Du.Dd.
- kappa_u is dimensionally yukawa^2 in scan units, but 9.68 > y_t^2 = 0.935, so no pure Yukawa monomial can express it. A bounded exponent search over Yukawa/CKM factors produced only multiple-comparison-grade hits (best: yt^-1.Vus^-1.5 at 8.4e-6) - recorded as numerology, not identification. The structural derivation of kappa_u is OPEN; the natural route is the LO expression of V_ub for u-anchor textures (Vub_LO vs H_12.sin(phi)).
Typing
All statements are chart-space laws of the fitted nine-link ensemble; WP11 stands (phi is chart data). But the u-class law now ties chart data (H_12.sin phi) to a PHYSICAL observable (V_ub) by an exact constant - the first exact chart-to-readout bridge found in this program. If kappa_u proves derivable from the texture LO structure, this becomes a candidate H2LR pairing component.