We report misleading effects of human-origin signals on very distant gravity experiments in solar eclipse days. Despite Newton's well-established gravity, some research has been underway for some years on alleged gravitational anomalies in the days of solar eclipses, as per still considered reports in mainstream peer-reviewed journals. In solar eclipse days some instruments, notably pendulums and gravimeters, had provided anomalies that were interpreted as unknown features of gravity. French physicist Maurice Allais is credited with first reporting pendulum anomalies during a solar eclipse in 1954. NASA researchers also wanted to investigate the matter, also because they were looking for an explanation for the so-called Pioneer anomaly (explained later), and promoted a worldwide test for the solar eclipse of August 11, 1999. We were recruited for that experiment in July 1999 following our proposal for an innovative experiment. We have since conducted two experiments with stationary pendulums and recorded surprisingly structured anomalies which after years of study proved unnatural. We report these tests with extensive control data, and a limited number of similar anomalies observed by others, some already published but still not as unnatural. In summary we analyze: our stationary pendulum test on Aug 11 1999 in South Italy (unpublished); our stationary pendulum test on May 31 2003 in South Italy (unpublished); a gravimeter test by University of Trieste (Italy) on Aug 11 1999 (unpublished); a report from Reading (UK) for the Aug 11 1999 eclipse (published); a report from Manavgat (Turkey) for the March 29 2006 eclipse (published); a report from Kiev (Ukraine) for the Sept 11 2007 eclipse (published). In the above experiments we found abnormal instrumental behaviors (sudden sharp micro-deviations) that obey a precise structure of equally spaced time markers. The timing scheme was always the same in all of the above experiments. Subsequent research revealed that the scheme matched the one used by a very distant source of high-power, man-made, controlled radio signals. We conclude that the measuring instruments tracked by chance a campaign of RF-based eclipse experiments over the years, not gravitational phenomena. We report in more detail our two pendulum experiments and the gravimetric test of University of Trieste, spending fewer words on the other tests already described in their respective articles.