Fixed CCD offsets and proper motion predict QLP minus TESS-SPOC times within 3 milliseconds in 16 Sector 84 stars

We test whether catalog proper motion predicts the relative times in public QLP v01 and TESS-SPOC v1 files. A prior fixed CCD and frame model left a 14.779-millisecond error in one Sector 84 star. A fixed geometric term from both proper-motion coordinates reduces that maximum to 0.500 milliseconds. We then save the formula and sample before acquisition of 16 new Sector 84 stars. Selection uses catalog geometry and favors large predicted effects. No measured time difference enters selection. No model parameter is fit to the new times. Across 173,456 finite matched exposure pairs, the RMS error falls from 14.456 to 0.943 milliseconds. The largest absolute error falls from 61.428 to 2.618 milliseconds. Every new star improves in RMS, and every point passes the saved 10-millisecond limit. The 3-millisecond bound is a measured result. A zero-quality subset gives the same conclusion. In the earlier 96-star sample, RMS improves for 56 stars and becomes larger for 40 stars. Separate correction checks support a static QLP direction and a moving SPOC direction in these files. Millisecond residual structure remains. This result tests relative archive conventions; it does not establish absolute clock accuracy, a production software cause, or a new planet. The package preserves source timing extracts, headers and hashes, catalog responses, ephemeris, code, tables, figures, and offline reproduction instructions.

TESS · QLP · TESS-SPOC · proper motion · barycentric correction · time stamps · archive validation · reproducible research