Two TESS Sector 51 flux series give five WASP-189 b eclipse depths that differ by 23 to 192 ppm with quadratic fits

We test whether separate eclipses of WASP-189 b need changing dayside light in TESS Sector 51. Five of nine predicted eclipses pass fixed coverage cuts in the released 120 second light curve. We use the same samples, weights, eclipse template, and local quadratic baseline for both flux series. Corrected aperture flux gives lower depths. The decrease is 22.8 to 191.7 parts per million (ppm). This range is comparable to the published 203.4 ppm global depth from the same visit. Thirty-two overlapping false eclipse windows per series give depth RMS values of 53.4 ppm for raw flux and 85.6 ppm for corrected flux. Their normalized RMS values are 1.39 and 2.19 with 60 minute residual-block errors. Tests of constant depth change with the flux series, baseline, and block length. These errors do not support a calibrated physical variability limit. Independent geometry and least-squares checks reproduce the measured column differences. Later published timing and geometry changes have much smaller effects. The result is a reproducible audit of this estimator on these released flux products. It does not identify which flux series gives the true planet depth. It does not show that archive processing removed planet light. The package preserves the exact public input, source records, event and control decisions, code, injections, tables, figures, and offline run instructions.

WASP-189 b · TESS Sector 51 · secondary eclipse · photometry · correlated noise · PDC · reproducible data audit