Though we seemed to have fixed all the edge cases that were allowing for super-Eddington accretion even when eddlim=1 is set, it seems like the first write to the bcm array during the mass transfer phase doesn't obey the Eddington limit. This is probably due to some writebcm call that happens before the mass transfer limit is imposed.
Below are a few example tracks that show this behavior, both for the "old" Hurley+2022 prescription for applying the Eddington limit for BHs as well as the "new" method derived from Bardeen1970 and Marchant+2017, which is implemented in PR #570. Initial conditions for these systems are also attached.
Though we seemed to have fixed all the edge cases that were allowing for super-Eddington accretion even when
eddlim=1is set, it seems like the first write to the bcm array during the mass transfer phase doesn't obey the Eddington limit. This is probably due to somewritebcmcall that happens before the mass transfer limit is imposed.Below are a few example tracks that show this behavior, both for the "old" Hurley+2022 prescription for applying the Eddington limit for BHs as well as the "new" method derived from Bardeen1970 and Marchant+2017, which is implemented in PR #570. Initial conditions for these systems are also attached.
super_edd_initCs.csv