Just wondering: Apart from the selection that D* should come from the primary vertex, did you do anything special to treat D* from B decays? I found page 20, but that is a bit unspecific in that respect. Some D° happen to fly nearly in the same direction as the B-meson, and I would assume that the D°/slowpi combination cannot resolve this well enough.

(I worked on charm mixing, too, and had the same issue. A reconstruction of some of these events helped to directly measure their influence.)