A constant period fits mean half-amplitude BAT eclipse centers of 4U 1538-52 from 2005 to 2026 under a ramp model

A constant period fits eleven mean hard X-ray eclipse centers of 4U 1538-52 under a stated profile model. We use 91,444 public Swift/BAT pointings from February 2005 to October 2026. We fit separate eclipse profiles in two-year time bins. The primary center is the midpoint between the half-amplitude crossings of separate ingress and egress ramps. Resampling blocks of five binary cycles gives the time errors. The constant period is 3.72831750 ± 0.00000514 d. Its chi-square is 12.22 for 9 degrees of freedom. A quadratic gives an eclipse period derivative Pdot/P = (-0.67 ± 0.50) × 10^-6 per year. These primary times do not require a period change. The 1%-amplitude center of the same ramps gives (-1.66 ± 0.74) × 10^-6 per year. Thus the evidence for curvature depends on the center definition. We also test profile bounds, symmetry, data weights, block length, time convention, early data, and solver convergence. The errors depend on the center and noise model. Apsidal motion also affects the eclipse clock. This result does not measure true orbital decay. The package contains the saved input, eleven new mean times, all resamples, source code, provenance, figures, and reproduction instructions.

4U 1538-52 · Swift BAT · eclipse timing · orbital period · orbital decay · X-ray binary · apsidal motion · block bootstrap · public data