Composite tolerancing of features on a cylinder
We have a part that is a cylindrical tube with 8x mounts that project out from the outside walls of the tube. They are used to mount the tube inside another tube. There are 4x mounts at the same Z height at top, and 4x mounts at same Z height on the bottom. They are at approximately 90 degrees apart. There are threaded holes at the outer faces of these mounts so the part can be attached to the mating part.
The feature control frame is composite, with the features checked to ABC .060, and as a pattern to itself .020 with no datum references and no bonus. The way I am checking this to itself on the CMM is creating a circle out of the top mount locations, a circle out of the bottom mount locations, a 3D line between the circles as my axis alignment, and creating a symmetry point between the two circles as my XYZ origin. For the tertiary alignment I'm creating a 2D line between the mount locations that are 180 degrees apart. I am consistently getting .025-.030 true_position on at least one of the mounts, and up to 4 of them depending on the part. Most of the variation is in the Z axis. I'm using the basic dimensions from the print to check the spacing and orientation. I've verified the distance in Z axis between mounts is off from the basic dimension by .013 with a cadillac and height gauge. But I'm still getting heat from engineering questioning my method and they are trying to get creative in the interpretation of the tolerance. Does it sound like my method is valid and accurate? In my mind it's not even a question if the Z distance between the mounts is .013 when the tolerance is .020
Thanks.
Kevin