A fixed frame model predicts QLP minus TESS-SPOC times within 0.01 seconds in 16 sector 85 test stars after baseline calibration

We test whether a known class of coordinate-frame error explains a relative time jump reported for one TESS sector 85 star. We compare public QLP v01 and TESS-SPOC v1 files at identical 200-second exposure labels. The first sample has 32 stars in sector 85 and 32 control stars in sector 84. After a shared gap, all 16 original sector 85 stars on cameras 2 and 4 show large negative changes. Their later QLP minus TESS-SPOC times range from -169.075 to -90.512 seconds. All 16 stars on cameras 1 and 3 keep offsets near two seconds. A fixed J2000 equatorial-to-ecliptic rotation of a public spacecraft ephemeris predicts this pattern. Each star supplies one median baseline before the gap. The model fits no angle, scale, or time shift. We freeze the model and a criterion before acquisition of 16 additional sector 85 stars. All 85,648 later matched points pass the 0.01-second criterion. Their RMS error is 0.001648 seconds and their maximum absolute error is 0.003765 seconds. The result persists when both quality flags are zero. The sector 84 zero-change control has a later RMS error of 0.001568 seconds. An independent reader exactly matches all 640 saved columns from 160 original FITS files. The package preserves native time extracts, source headers and hashes, the ephemeris, failed models, figures, and offline reproduction code. The result supports a selective frame mismatch in these relative file times. It does not establish the production code cause, absolute clock accuracy, or a new planet.

TESS · QLP · TESS-SPOC · sector 85 · time stamps · barycentric correction · coordinate frames · camera comparison · data validation · reproducible research