Failed to calibrate cylindrical stylus in unusual position using DCC

Failed to calibrate cylindrical stylus in unusual position using DCC

I need to measure a square slot inside a plastic part so I set up a single cylindrical stylus on a star hub and a tip of A90 B180 as shown in the image.

It calibrates fine in DCC in A0 B0, but when I go to calibrate in A90 B180, it starts ok touching the sphere first on the top (Z axis) and then on the equator on the first two points (say south and east) but then as it goes to the third point (north) it arcs into the sphere touching before reaching the  90° north position to make a proper approach.  The extension holding the stylus hub is clear of the top of the sphere.  Tried changing the number of calibration points both less and more but it can't do it.

I then switched off DCC in calibration and manually (joystick) touched 7 points which gave me a form error of 3 micron (more than good enough for our plastic parts).  If I cut the part up and measure with a vernier caliper the sizes measured with CMM correlate.

Two questions
  1. What am I missing to get a DCC calibration? Or is this even possible?
  2. Is a manual calibration not using DCC a legitimate technique?