Repositorytess-single-transit-01SEARCH_LOG.md

Document · SEARCH_LOG.md

Search log: toi-3500-02

Exact products, checksums, queries, exclusions and untested checks.

Project-authored

Historical analysis below. Current evidence and dossier: 2026-09-30 follow-up. S64/S90 dips persist, but official UPDATED_2.0 TESS PRFs do not assign the deficit robustly between the0.18-pixel-separated Gaia A/B pair. Source preference reverses with allowed registration and1/3/5-percent model floors; static PRF residuals exceed nominal interpolation accuracy. S101 remains rejected and supplies no period preference. All16 baseline aliases remain open; Gaia RV summaries provide no orbit or companion-mass bound. Unverified lead.

2026-09-27 verification update: Canonical dossier units, local-null claims and difference-image states were corrected. The follow-up report records the live metadata, independent cached-TPF reduction and two-source Gaussian-PSF grid. The grid is target-favoring in most S64/S90 trials but changes sign across plausible PSF/registration perturbations, so the 3.71-arcsec companion remains unresolved. Older interpretations below are historical where they conflict.

Generated draft; see REPORT.md.

ItemValue
ArchiveMAST (TESS SPOC 120-s light curves), discovered with the MAST Observations API
Products screened4
SelectionQUALITY = 0 with finite, nonzero TIME, SAP_FLUX, PDCSAP_FLUX
Veto{'kind': 'single_epoch', 'veto_hours': 12}
Thresholdcalibrated per light curve (k*), SAP and PDCSAP both, ≥ 2 cadences, baselines 1/2/3 d
Entries outside veto185 (44 distinct events, 21 persistent)
Catalogue servicesNASA_Exoplanet_Archive, SIMBAD, TESS_TOI, VSX
Random seed20260927

Per product#

ProductRowsUsablekEntriesOutside veto
tess2023096110322-s0064-0000000443666343-0257-s_lc.fits19385174023.0011714
tess2021091135823-s0037-0000000443666343-0208-s_lc.fits18249143332.502929
tess2025071122000-s0090-0000000443666343-0287-s_lc.fits20105182733.009191
tess2026060005000-s0101-0000000443666343-0303-s_lc.fits18814154062.505151

Not searched / not tested#

  • Sectors beyond those listed above (at most six light curves per target are retrieved).
  • Full-frame-image and non-SPOC light curves; alternative detrending; difference-image centroids; pointing correlation; ADS literature.