# Passive axisymmetric cylinder — Round 2 review package

Status: **published supporting evidence for Study 001**. The media are derived
review aids rather than independent evidence; compact data, audit summaries,
portable source identifiers, and hashes preserve traceability to the strict
simulations without exposing private workstation paths.

Terminology note: the media use “excited branch” as compact shorthand for the
persistent, much higher-energy large-angle angular branch requested in this
round. Its coexistence, damping survival, passive power balance, and basin
capture are established; the stricter fixed-parameter inverse theta-versus-spin
response-sign test from the current working definition remains unproved for
this particular branch. Publication prose should therefore call it a
**large-angle excited candidate** unless that response test is added later.

## Findings represented

1. **Controlled manufacturing-tilt ladder.** The identical magnetic-frame
   launch is tested at 0°, 1°, 2.121°, and 5.476°. All four retain broad
   large-angle motion. At 0°, however, magnetic torque about the physical
   cylinder axis is identically zero and axial spin follows pure exponential
   damping. Manufacturing tilt is therefore necessary for magnetic axial-spin
   exchange in this point-dipole model, but not for this launch's broad theta
   motion. The response is nonlinear: 2.121° produces more negative mean spin
   than 5.476°.

2. **Same parameters, two attractors.** A synchronized comparison shows the
   coexisting large-angle branch and quiet cone with identical physical
   parameters and drive phase but different initial states. The late mean
   mechanical-energy ratio is about 2.76.

3. **Small-gamma detuning.** Gamma=4.47°, 4.570079°, and 4.67° all preserve the
   same broad, persistently winding regime for 200 s. This is a useful negative
   result. The many wrapped events are ordinary transverse winding-boundary
   crossings, not phase slips from a coherent 1:1 lock.

4. **Cone-to-excited basin bracket.** Along one exactly defined transverse
   body-rate direction, k=126.0 rad/s settles to the quiet cone while
   k=126.5 rad/s settles to the large-angle, average-A/P basin. The separation is only
   0.396% of the midpoint. The wider one-dimensional slice is nonmonotone and
   also contains a parallel high-speed basin, so this is a local bracket, not a
   universal scalar excitation threshold.

5. **Why the marker goes forward.** The body +x marker can wind forward in the
   laboratory while the local cylinder-axis spin is negative because these are
   different angular observables in a strongly nutating body.

6. **Energy flow with psidd=0.** The controller port is absent. The rotating
   field supplies work, damping removes it, and the mechanical-energy ledger
   closes over an integer-cycle settled interval. The small axial torque is
   only one component of the full energy transfer.

7. **Gamma step-and-return.** A scheduled 4.570079° → 5.70° → 4.570079° test
   retains the broad branch and returns statistically to baseline. This is a
   controlled parameter-step/memory probe, not spontaneous decay. A direct
   jump from another microstate gives a different outcome, demonstrating basin
   sensitivity near a nonmonotone boundary.

## Media conventions

- Review videos are H.264 MP4, 1920×1080 or 1280×720, 30 FPS, web fast-start,
  no audio, and at most one minute.
- The tilt ladder is intentionally static. At 30 FPS, a 60 Hz drive advances
  exactly two cycles per frame and an ordinary playback would falsely freeze
  the drive. A future animation should use explicitly labeled slow motion or a
  drive-stroboscopic view.
- Full strict trajectories stay in the simulator research workspace. Public
  NPZ files are synchronized compact decimates plus strobes/spectra where
  relevant.

See `review-round-2-manifest.json` for hashes, portable source provenance,
media metadata, and exact public filenames. All 44 manifest-listed files pass
the post-publication existence, hash, format, and media checks with zero
failures.