In five TESS sectors, fixed 5 by 5 pixel sums show wider KELT-17 b transits in 2023 than in 2021 under a fixed profile

We test whether a reported KELT-17 b transit width increase remains in direct sums of public TESS pixels. We use the same 33 events in sectors 44, 45, 46, 71 and 72. A fixed 5 by 5 square gives a change in intrinsic width at half depth of +0.999 minute from 2021 to 2023. The nominal 95 percent interval is +0.518 to +1.460 minutes from paired whole-event resamples within sectors. The change minus the corrected pipeline result is +0.141 minute, with an interval of -0.050 to +0.337 minute. A 3 by 3 square and two baseline extensions also retain positive changes. The source corrected-flux fits are reproduced exactly. A planned test with model background added reaches a width bound for one sector 44 event. It cannot support a decrease. A later exploratory fit with a quadratic baseline also reaches a bound. A separate exploratory fit adds a free coefficient on the supplied background series. It gives +0.929 minute, with an interval of +0.480 to +1.371 minutes. The positive change does not require the tested SPOC aperture or later light-curve correction. All sums still use the same calibrated, background-subtracted pixels. The intervals hold unknown sector and epoch errors fixed. The result does not establish orbit precession or a new planet. The package preserves exact pixel windows, source hashes, light curves, code, fits, resamples, figures, and reproduction steps.

KELT-17 b · TESS · target pixels · aperture photometry · transit width · background subtraction · public astronomical data