Evidence for a 15.196-day cycle in 4U 1909+07 at MJD 60671–61333 depends on the BAT noise model

We test the known 15.196-day X-ray cycle of 4U 1909+07 after the published BAT cutoff at MJD 60670. The new interval ends at MJD 61333. A joint model fits the orbit, trend, and two long-cycle harmonics to 359 valid daily BAT rates. The fitted long-cycle root mean square (RMS) is 0.000619 count cm^-2 s^-1. Calendar-block sign tests give p = 0.02380 for 30-day blocks and 0.07422 for 60-day blocks. Equal weights give p = 0.2821 for 30-day blocks. Thus, the primary daily-data evidence depends on the noise and weight assumptions. Signal injections use the measured early waveform, the actual late observation times, and fixed late weights. At the early RMS of 0.000336 count cm^-2 s^-1, conditional recovery ranges from 23.0% to 42.7% across tested phases and block lengths. A weak later test cannot establish continued suppression at that amplitude. Exploratory checks use 3456 BAT pointing rates and 470 MAXI days. MAXI 4-10 keV gives a feature at the fixed period, with 30-day p = 0.00290 and a fitted-curve similarity of 0.877 to BAT. All MAXI bands are reported. These checks make a return plausible but do not establish a joint discovery probability or a physical cause. The package preserves source tables, hashes, the initial plan, code, figures, complete control results, and reproduction instructions.

4U 1909+07 · Swift BAT · MAXI · superorbital modulation · correlated noise · signal injection · reproducible research