A sine model leaves less scatter than a delay model for 267 TIC 71396049 eclipse times with Gaia and TRES data

We test light travel delay as the cause of eclipse timing changes in TIC 71396049, a source already reported as a triple. We fit six public TESS QLP sectors with seven choices of flux, baseline, and eclipse width. The main fit gives 267 eclipse times. At the nominal 49.45206-day Gaia DR3 orbit, a delay model with a free signed amplitude leaves a residual root mean square of 1.056 minutes. A free sine leaves 0.877 minutes. An orbit fit with a Gaussian Gaia constraint reduces the delay residual to 0.886 minutes at 49.38421 days. This timing-only curve fails two further tests under the stated assumptions. Four published TRES radial velocities leave 16.2 km/s of scatter about it after a free velocity offset. At the nominal Gaia RV amplitude, its timing scale requires an RV-star mass of at least 5.53 solar masses if the RV star is the tertiary and all timing change is delay. The published single-star mass estimate is 1.329 solar masses. Joint fits use the Gaia covariance, the four TRES velocities, a fixed timing scatter of 0.9 minutes, and extra RV scatter of 0.1, 0.5, or 1.0 km/s. Free-amplitude delay models then leave 1.012 to 1.035 minutes of timing scatter. Sine models leave 0.878 minutes. At the fixed published mass, delay models leave 1.035 to 1.046 minutes. Thus, the tested delay fits retain excess scatter when the independent velocities and the mass scale are used. These conditional fits do not give a formal rejection probability. Dynamical timing and changing eclipse shape remain possible causes. We release the event times, frozen data, code, figures, and checks. No new planet or unique timing mechanism is established.

TIC 71396049 · Gaia DR3 · TESS · eclipse timing · hierarchical triple · light travel time · orbit covariance · reproducible data