A free third-year term raises the 90-percent power amplitude to 5.8–9.6 m/s for a fixed Gl 229 wave under unit white noise

We test whether 197 public Gl 229 HARPS nightly times can separate a fixed 121.933-day circular wave from a free 121.75-day wave. With known independent Gaussian noise of 1 m/s per night, a two-degree test at a 1 percent false alarm rate needs a semiamplitude of 5.81 to 9.56 m/s for 90 percent power across signal phases. Without the annual harmonic term it needs 0.442 to 0.457 m/s. The harmonic fit retains 0.226 to 0.587 percent of baseline-only signal information. Two fixed correlated-noise cases need larger amplitudes. Exact noncentral chi-squared calculations agree with full simulated records and separate least-squares checks. These are model sensitivity results, not measured planet limits. The covariance is assumed and the period is fixed. No annual fault, planet, or new object is established. Code, aggregate results, an original figure, source hashes and external retrieval steps support reproduction. Source records are outside the publication.

astronomy · Gl 229 · HARPS · annual harmonic · detection power · signal information · null result