Evidence for a 29-minute pulse in V844 Herculis depends on the phase and amplitude model in the 2024 TESS data

We search for the known 29-minute optical pulse in 28,815 accepted measurements from 2024 TESS sectors 78 and 79. A model with separate phases and amplitudes gives a maximum weighted fit improvement of 21.72. One-day random signs on residual time blocks give a scan fraction p = 0.0231. This exceeds the selected 1% threshold. A follow-up model with one phase and one target fractional amplitude gives a semi-amplitude of 1.08% and p = 0.0038. This follow-up was selected after the main result. Its p value has no correction for that model choice. The evidence also changes with the noise block length. Pixel maps at the selected frequency and phase are consistent with the target position, but do not give independent confirmation. The main test recovers 1,910 of 2,000 added pulses at a 1.5% target semi-amplitude. This fraction uses the archive estimate of target light. It is a recovery level, not a confidence bound on the true amplitude. These data do not establish a secure pulse detection or a new spin clock. The package includes source data, accepted pixel arrays, code, figures, provenance, and a checked offline reproduction procedure.

V844 Herculis · TESS · white dwarf · intermediate polar · optical pulse · time series · injection recovery · correlated noise