I don't have a log for the 3.5.3-session - logging was not turned on, unfortunately. I've updated it now and turned logging on and will see of I can produce more crashes.
Attached is the PHD2 log, starting immediately after I turned off the internal solver.