A bounded quadratic sine clock with six-hour baselines leaves a 2.35-minute phase offset in TESS Sector 36 of WD 1054-226

We test whether a linear frequency change can explain a known phase mismatch in WD 1054-226. We use 80,273 accepted samples from five public TESS 120-second light curves in 2019 to 2026. The main model has quadratic baselines in six-hour bins, one shared phase clock, and separate nonnegative amplitudes. We search 62.2 to 62.5 cycles/day at BTJD 3000 and frequency derivatives from minus 1e-6 to plus 1e-6 cycles/day squared. The best tested quadratic clock has frequency 62.3497557155 cycles/day and derivative plus 1.29010e-7 cycles/day squared. It improves the weighted fit by 6.9593 units over the best constant clock. It still leaves a phase offset of minus 0.10173 cycles in Sector 36, equivalent to minus 2.35 minutes of a cycle. Three-hour and twelve-hour baselines and raw aperture flux give offsets from minus 0.1017 to minus 0.1028 cycles. A denser search and independent normal equations reproduce the main score. We test the residual mismatch with 499 fitted-residual sign draws for each of three block lengths. A feasible coherent fit gives a conservative numerical estimate of each control statistic. The counts are 5, 0, and 0 exceedances for 0.25-day, 1-day, and 2-day blocks. These tests assume independent, sign-symmetric blocks and sufficient interpolation padding. They support a residual mismatch under the tested models. Competing cycle counts remain. The fitted derivative does not establish orbital decay. The contribution is a bounded clock test and a method to check its noise sensitivity without a full alias search in every draw. The package preserves source FITS data, archive hashes, code, competing modes, control draws, figures, checks, and reproduction instructions. The result does not identify a new planet or the source mechanism.

WD 1054-226 · TIC 415714190 · TESS · white dwarf · debris transits · phase coherence · quadratic clock · frequency derivative · cycle count · residual sign controls · reproducible astronomy