Free clock terms reduce the nominal 2018 DE CVn curve residual to 2.2 s RMS in five 2019–2024 TESS visits

We fitted 253 primary eclipses of DE CVn in five TESS visits from 2019 to 2024. We tested the nominal quadratic plus sine timing curve of Han et al. (2018). The eclipse times use symmetric profiles and exposure averaging. The original curve has an absolute root mean square difference of 225.9 s from the five sector means. A fit of a new epoch and mean period reduces this difference to 2.17 s. The largest remaining difference is 4.19 s. A separate test cancels both clock terms. The 2022 mean is 9.28 s later than a straight interpolation between the 2019 and 2024 means. The estimated fit and day-group error is 1.34 s. The nominal 2018 curve gives 7.20 s. The two cadence products for the 2022 visit have sector centers that differ by 3.79 s. This difference limits the meaning of small fit residuals. Earlier ground data had already shown that the original absolute prediction failed. The present conditional shape agreement does not confirm a planet. The primary series supplies 250 individual time estimates that do not reach the search bound. The package contains all fit outputs with quality flags and the source light curves. It also contains parameter tests, analysis controls, figures, source code, and reproduction instructions.

DE CVn · TESS · eclipse timing · circumbinary planets · timing curvature · public data · exposure integration · white dwarf binary