Heh, pretty much the same technique to arrive at the 3.3" approximation, then. I just simplified it to working on a 2D circle rather than a 3D cylinder.