[cmccomb]

Purpose

Generate fresh corrected active-axis source histories for the 0.005, 0.0025, and 0.00125 s physical steps of pinned Nonlinear_Hydro/ode45/Regular and RegularCIC. An earlier artifact logged an inactive PTO axis, so its heave/speed continuity is useful but its PTO force/power cannot establish convergence. The #659 output-grid A/B already measured the corrected 0.00125 s arm; this run supplies the entire first-three sequence through the same #655 extractor.

Each source step starts a fresh R2025b MATLAB process, avoiding accumulated simulation workspaces. All five clocks equal the chosen physical step for these three arms; the application wave, 150 s duration, ramp, nonlinear-hydro mode, and damping are runtime asserted. The workflow uploads raw histories/settings plus generated BEM HDF5 and STL for independent gate analysis.

This is a source-data regeneration diagnostic. Its success alone does not assert four-step convergence or independent Python parity. The mixed-output finest Regular and RegularCIC trials remain separate; #655 retains the predeclared pointwise, signed-work, and direct Python gates. No original published 0.05 s status changes follow from this PR.

Measured source artifacts

At exact head daa282c9, run 40107175117 completed both cases. Every first-three settings file records all five clocks at its specified step. The 0.005 s corrected active-axis heave, speed, PTO force, wave, and restoring histories match the earlier full-step archive exactly on shared timestamps. All six fresh histories obey the pinned damping force/power law to <1.3e-9 in the recorded units.

For Regular, first and second source changes on the 0.005 s common grid are 0.000928731→0.000463590 m heave, 0.001364218→0.000685294 m/s speed, 1637.061→822.353 N PTO force, and 1339.029→664.157 W power. Signed works are 20,154,278.580→20,098,017.152→20,069,978.377 J. For RegularCIC, the corresponding changes are 0.000967940→0.000482786 m, 0.001461709→0.000735055 m/s, 1754.051→882.066 N, and 1445.451→732.445 W; works are 20,186,950.852→20,127,221.682→20,097,460.025 J. Every first-to-second change contracts below factor 0.7, including signed work. All 38 BEM HDF5 datasets match in value across the two cases, and STL SHA-256 is identical. The final 0.000625 s source and independent Python gates remain separate.

[cmccomb]

The corrected active-axis first-three source histories and settings were consolidated and independently reproduced by the exact-head convergence workflow in merged #655. This diagnostic PR is superseded.