i think denis was on the right track when he mentioned something about the way it engages.

If one side of the clutch disk gripped before the other (and gripped hard) it could explain how one disk broke off, in theory anyway