ABSTRACT HD 62658 (B9p V) is a little-studied chemically peculiar star. Light curves obtained by the Kilodegree Extremely Little Telescope (KELT) and Transiting Exoplanet Survey Satellite (TESS) show clear eclipses with a period of about 4.75 d, as well as out-of-eclipse brightness modulation with the same 4.75 d period, consistent with synchronized rotational modulation of surface chemical spots. High-resolution ESPaDOnS circular spectropolarimetry shows a clear Zeeman signature in the line profile of the primary; there is no indication of a magnetic field in the secondary. PHOEBE modelling of the light curve and radial velocities indicates that the two components have almost identical masses of about 3 M⊙. The primary’s longitudinal magnetic field 〈Bz〉 varies between about +100 and −250 G, suggesting a surface magnetic dipole strength Bd = 850 G. Bayesian analysis of the Stokes V profiles indicates Bd = 650 G for the primary and Bd < 110 G for the secondary. The primary’s line profiles are highly variable, consistent with the hypothesis that the out-of-eclipse brightness modulation is a consequence of rotational modulation of that star’s chemical spots. We also detect a residual signal in the light curve after removal of the orbital and rotational modulations, which might be pulsational in origin; this could be consistent with the weak line profile variability of the secondary. This system represents an excellent opportunity to examine the consequences of magnetic fields for stellar structure via comparison of two stars that are essentially identical with the exception that one is magnetic. The existence of such a system furthermore suggests that purely environmental explanations for the origin of fossil magnetic fields are incomplete.