Separate day fits give a 10.6% larger late beat harmonic ratio for AR Scorpii in TESS Sector 91

We test pulse-shape measures for AR Scorpii in public 20-second TESS Sector 91 data. Separate Fourier fits retain 17 days and 67,797 samples with zero quality flags. The first-to-second beat harmonic amplitude ratio and the relative phase are unchanged by a common flux scale and a constant pulse delay within a day. The early and late geometric mean ratios are 0.4101 and 0.4536. The relative increase is 10.6%, with an individual 95% interval of 2.3% to 19.6% if days are independent. The relative phase changes by 5.14 degrees, with an interval of -1.55 to 11.82 degrees. Earlier SCI work already showed the point amplitude difference. This work adds day contrasts, relative-phase tests, time-group controls, and synthetic checks. Grouping the same day results for correlated-error controls retains a positive ratio contrast. New fits over two-day segments change the selected samples and averaging; they give 7.4%, with an interval of -3.1% to 19.0%. Thus, evidence for variation depends on the fit and error assumptions. The analysis is exploratory and does not identify a physical cause or rule out a common pulse-phase drift. The package contains the source data, provenance, code, figures, results, and offline reproduction steps.

AR Scorpii · TESS Sector 91 · white dwarf · beat pulse · Fourier shape · pulse timing · time dependence