2010MNRAS.407.2362P

We present high-speed ULTRACAM photometry of the eclipsing post-common-envelope binaries DE CVn, GK Vir, NN Ser, QS Vir, RR Cae, RX J2130.6+4710, SDSS 0110+1326 and SDSS 0303+0054 and use these data to measure precise mid-eclipse times in order to detect any period variations. We detect a large (∼250 s) departure from linearity in the eclipse times of QS Vir which Applegate's mechanism fails to reproduce by an order of magnitude. The only mechanism able to drive this period change is a third body in a highly elliptical orbit. However, the planetary/sub-stellar companion previously suggested to exist in this system is ruled out by our data. Our eclipse times show that the period decrease detected in NN Ser is continuing, with magnetic braking or a third body the only mechanisms able to explain this change. The planetary/sub-stellar companion previously suggested to exist in NN Ser is also ruled out by our data. Our precise eclipse times also lead to improved ephemerides for DE CVn and GK Vir. The width of a primary eclipse is directly related to the size of the secondary star and variations in the size of this star could be an indication of Applegate's mechanism or Wilson (starspot) depressions which can cause jitter in the O-C curves. We measure the width of primary eclipses for the systems NN Ser and GK Vir over several years but find no definitive variations in the radii of the secondary stars. However, our data are precise enough (Δ R_sec/R_sec< 10^–5) to show the effects of Applegate's mechanism in the future. We find no evidence of Wilson depressions in either system. We also find tentative indications that flaring rates of the secondary stars depend on their mass rather than rotation rates.