A fixed CCD model predicts QLP minus SPOC times within 0.05 seconds in 128 stars from TESS sectors 90–97

File time errors can bias measurements of planet transit times. We compare public QLP v01 200-second and mission SPOC 120-second light curves for 192 distinct stars in TESS sectors 86–97. Integer exposure labels give 769,091 pairs at nominal common centers. Every time difference is between 1.904377 and 1.986884 seconds. No pair exceeds the initial five-second screen. An initial look at sectors 86–89 shows a pattern with CCD number. Before opening any time column from sectors 90–97, we fix the prediction at 2.00 minus 0.02 times the CCD number, in seconds. No star offset is fitted. All 525,356 pairs from the 128 test stars meet the chosen 0.05-second limit. Their root mean square error is 0.003907 seconds. The largest absolute error is 0.020569 seconds. A constant estimated from the exploratory sectors gives a test root mean square error of 0.022323 seconds. The constant also meets the limit, so the lower error supplies the added evidence for the CCD pattern. The quality-zero subset gives the same conclusions. This is a local model record made after downloads began. It is not an external registration or a random sample. The result gives a useful relative baseline for these files. It does not establish the production code cause, absolute clock accuracy or a new planet. A second FITS reader matches all 384 source files. The package preserves native input extracts, source hashes and headers, queries, code, figures, audit tools and reproduction checks.

TESS · QLP · SPOC · CCD · time stamps · transit timing · public data · data validation · reproducible research