Constant orbital period ephemerides of eclipsing binaries give the computed eclipse epochs (C). These ephemerides based on the old data can not accurately predict the observed future eclipse epochs (O). Predictability can be improved by removing linear or quadratic trends from the O-C data. Additional companions in an eclipsing binary system cause light-time travel effects that are observed as strictly periodic O-C changes. Recently, Hajdu et al. (2019) estimated that the probability for detecting the periods of two new companions from the O-C data is only 0.00005. We apply the new Discrete Chi-square Method (DCM) to 236 years of O-C data of the eclipsing binary Algol ($beta$ Persei). We detect the tentative signals of at least five companion candidates having periods between 1.863 and 219.0 years. The weakest one of these five signals does not reveal a ``new companion candidate, because its $680.4 pm 0.4$ days signal period differs only $1.4 sigma$ from the well-known $679.85 pm 0.04$ days orbital period of Algol~C. We detect these same signals also from the first 226.2 years of data, and they give an excellent prediction for the last 9.2 years of our data. The orbital planes of Algol~C and the new companion candidates are probably co-planar, because no changes have been observed in Algols eclipses. The 2.867 days orbital period has been constant since it was determined by Sir Goodricke (1783).
Download