Block noise tests support a 9.6-day signal in 2014–2026 BAT and MAXI data for 4U 2206+54

We test the short X-ray component of 4U 2206+54 in public BAT 15–50 keV and MAXI 4–10 keV data. Separate fits cover 2014–2019 and 2020–2026. The model has 180-day offsets, a fixed 19.1182-day sine and cosine, and variance weights. It adds a component at the published period of 9.5591 days. All four tests pass a correction for four comparisons under 40-, 80-, and 160-day random block sign tests. The largest corrected p value is 0.01588. A joint test shares block signs between instruments and searches 1,028 frequencies from 8 to 22 days. None of 10,000 draws at each block length reaches the data score of 64.863. A denser grid gives the same result at 160-day blocks. Fitted short amplitudes are 0.0002873 and 0.0003751 count cm^-2 s^-1 for BAT, and 0.001571 and 0.001994 photon cm^-2 s^-1 for MAXI. The noise tests require approximate independence and symmetry of block scores. Equal weights weaken some separate tests. The analysis follows exploratory data inspection. Each record has a separate phase, so the result does not establish a stable full-record phase or an orbital period. This extends a prior SCI study of the same data with short-component tests and a joint search noise test. It is not a new first detection of the known period. The package preserves the input data, provenance, code, figures, results, controls, and repeat instructions.

4U 2206+54 · X-ray binary · Swift BAT · MAXI · period search · correlated noise · block bootstrap · harmonic modulation · public data