From what I can tell, at very long distances even, the beam direction seems very unaffected by errors (even large ones) whereby the plane on the filter is not perfectly perpendicular to the beam axis.
This seems to be a good simple solution for my application. Overall goal being to minimize...