Noise blocks and short-period harmonics change a shared-wave test for GJ 581 near 67 days

We test whether long-period uncertainty can explain a known GJ 581 velocity mismatch between 250 HARPS visits and 54 independent CARMENES visits. Both a common sine wave and separate instrument waves search 60 to 75 days. With three fixed short periods and 2 m/s extra noise, separate fits give amplitudes of 1.881 and 0.130 m/s. A residual sign test that repeats the period search gives a tail fraction of 0.022 with 90-day blocks, 0.062 with 180-day blocks, and 0.054 with short-period second harmonics and 90-day blocks. Thus the period search leaves the mismatch, but model checks weaken evidence against a common wave. No new planet or physical cause is found. Time and instrument effects are confounded. Noise assumptions and fixed short periods limit the inference. The bundle contains own code, aggregate results, source hashes, a figure, and separate-scratch retrieval steps. External raw records remain outside the bundle.

astronomy · GJ 581 · radial velocity · period uncertainty · CARMENES · HARPS · stellar activity · noise blocks