Duane,

You are correct going by the book. You establish the center plane -C- subject to the constraint being normal to -A-, then rotate the DRF about -B- to make it parallel to -C- and then translate the origin onto -C-.

The problem is that using a plane or a point for datum -C- produces two completely different alignments. If you look in the Y14.5M Standard, Fig. 4-15 they've got it all wrong, because you cannot force the center plane of the slot to go through -B- by rotating the part (they actually figured this out in the Mathematical Definition of Y14.5M Fig. 4-2). The effect of using a plane is that the center of the ring (-B-) will be shifted and this is seldom the desired behavior. The other problem with this setup is that it is very unstable mathematically. If the distance from the slot to axis -B- is much bigger than the length of the slot, any small error in the slot orientation is throwing off the origin by a lot.

Using a point as -C- is much more stable and generally better. It also preserves the origin of the DRF in the center and you just have to "clock" -C- by rotating about -B-. In order to do this you should either have two target points on the slot at the same distance/height and do iterative alignment OR intersect the central plane of the slot with a theoretical cylinder with fixed size (e.g. 2.5" diam. in your case) and axis in -B-, then use this gauge point (line) as datum -C-. Now the slot can be analyzed as a separate feature to A|B|C. This should all be on the drawing. I have the feeling that the engineers want this 99% of the time, but are not conveying it properly. Why not ask them what's the intention.

George K.
Sitius