Found the 6.7mm error this morning. The error was not in the EDM measurement, but in the design of the corner cube mount. As the mount is currently designed, we touch off of a metal plate that is sandwiched between the corner cube and the XY translation stage of the mount with a CMM to accurately measure the distance between this metal plate and the ITM; this gives us the distance we need to add to the EDM measurement to get the total distance from the total station to the ITM (the EDM from the total station only measures to the corner cube). This plate was thought to be at the back plane of the corner cube (most corner cubes have a constant that is the offset distance from the virtual focus of the coner cube to its back plane and this distance must be compensated for when using a corner cube for EDM; our corner cube has a 30mm constant) but in actuality the back plane of the corner cube is 10mm inside the metal plate. This subtracted 10mm from the measured distance, indicating that the structure had to be moved in the -y direction. We will start considering a redesign of the corner cube mount, but in the short term we were able to compensate for this using the total station. The total station has the ability to automatically compensate for a corner cube constant, therefore we simply changed the constant from the 30mm we were using (since it was a 30mm prism constant) to 20mm, which represents the distance from the coner cube virtual focus to the front surface of the metal plate. This also means that the FM, which was thought to have been 12.7mm too far forward, is only 2.7mm off of its ideal longitudinal position, which is within the ±3mm of error set out in E1100690.
As a result of today's work, we determined that no more structure moves are necessary and we can continue with the fine pitch and yaw alignments.