Block noise tests do not show a 9.6-day period change between 2014–2019 and 2020–2026 for 4U 2206+54

We test whether two public X-ray records for 4U 2206+54 require different periods near 9.6 days in 2014–2019 and 2020–2026. The model permits a separate amplitude and phase in each era. We compare a common frequency with separate frequencies over periods from 9.4 to 9.8 days. Each of 4999 noise trials repeats the frequency search. Random signs act on residuals in 40-, 80-, or 160-day blocks. Swift/BAT 15–50 keV gives p values of 0.3246, 0.3358, and 0.4866. MAXI 4–10 keV gives 0.2572, 0.1964, and 0.1412. The tests do not establish a period difference under these noise assumptions. They also do not establish stability. Added trials show limited resolution: an imposed +0.010-day shift in BAT passes the fitted-null threshold in only 19.1–35.3% of trials. Wave strength and period estimates depend on the noise blocks and retained years. A later-boundary MAXI check gives one borderline p value, which does not persist across block lengths. This follow-up uses records that overlap a prior SCI study. It measures brightness recurrence, not an orbital period derivative. The package preserves the source data, provenance, code, trial results, figures, numerical controls, and an offline reproduction check.

4U 2206+54 · Swift BAT · MAXI · X-ray binary · period search · correlated noise · block sign test · frequency change · reproducible research