Residual blocks change the 167.43-day clock test for 4U 1820-30 in 2023–2026 MAXI data

We test two central clocks for the long brightness cycle of 4U 1820-30. We use 937 public MAXI 2–20 keV daily values with MJD > 60157, through MJD 61332.50. Phase, a two-harmonic profile, and a linear trend are free. A 120–220-day search gives 169.99 days. We change the signs of fitted residuals in calendar blocks of 15, 30, 60, 120, and 167.43 days. We use 2,000 draws per length and repeat the full search in each draw. The 2.5–97.5 percentile period ranges lie within 168.34–172.62 days. Null tail fractions for the 167.43-day clock range from 0.036 to 0.072. Thus its test crosses 0.05 as the block length changes. The central clock with a decreasing period gives tail fractions from 0.002 to 0.022. These fractions assume independent symmetric residual blocks and adequate fitted null profiles. Their coverage is not validated for all source changes. The longest blocks give only 8–9 units. Published clock parameter errors are not propagated. We add conditional period ranges and clock tests to an earlier SCI point estimate from the same observations. We do not exclude full clock families or identify the cause of the brightness cycle. The package contains frozen data, provenance, all draws, figures, code, numerical checks, and offline repeat instructions.

4U 1820-30 · MAXI · superorbital period · X-ray binary · residual blocks · period uncertainty · clock tests · reproducible astronomy