Im impressed with the erudition of the above posts, but the purpose of this exercise must be to define a fairly simple limit for the KE of a spinner that is a bit cleverer than Mentorns 1500rpm rule. The limit needs to be crude enough to be checkable. That rules out any consideration of weight distribution. Here is how I would do it for heavyweights:-

Spinner weight(kg) * (Radius(m) * Speed(RPM)) squared 5,000,000

Then we get into the more interesting argument about whether 5 million is too much!

Jeremy