How is Datum B constraining the Position tolerance?

Make -a- datum cylinder axis as long as humanly possible.   

This drawing is terrible. 

1: -A- datum is ambiguous as there's five interrupted cylinders.  Is the datum just the middle one, or all five features of size there?

2: -B- datum cannot be a "width" between the same vectors like that.  I agree with Neil, that -B- datum should be defined and tied to the bottom extension line of that 1.00" dimension, clarifying the -B- datum as the one planar flange.

The 1" protrusion/boss to the hub flange has no significance to B datum.  The perpendicularity should be only tied to that -B- datum flange face.

If you are getting parts that are out of spec, using A to B, reverse the datums... making B primary and A secondary, and see how the true position comes out.  give that info to dumb engineer.

Hi Louis, the datum A is just the cylinder with the datum attached to it, not the rest of the interrupted cylinders. The very intelligent engineer did update the drawing now but I still don't think its right. Attached is how he drew it. When I try to add datum B pc-dmis won't let me. Also not quite sure about the dash in between datums on the feature control frame, never seen that before.

I just checked the datums separately, but I think I should tell him that I cant check it how he has it called out. What do you think Louis?

Okay after some research looks like this is a "common datum axis" I am going to try to construct a line using the two cylinders on my cmm and use that as the datum

The datum is A-B not A|B.

You need to define Datum A and B individually, and then create a compound datum.

beautiful. Thankyou. These parts were out before but now they are wayyy inside the tolerance. Thanks be to god.

beautiful. Thankyou. These parts were out before but now they are wayyy inside the tolerance. Thanks be to god.