Dave Looman
Ansys Employee

I don't think the two methods for computing rotational mass participation will necessarily be maximum for the same mode.  I haven't seen your other post, but it sounds like you could just connect a torsional spring to the base nodes.