So I have looked at every bidirectional TP thread on - I am programming for a customer with 2011 MR1, so width is out of the question (and has been flaky at best in 2014.1).
With only a few years experience, I was never taught constructing generic features. I think I have a grasp on creating generic features for the two slots, my trouble is creating generic features for the datums called out and applying MMC or LMC with xactmeasure when dimension true position. Program seems to be running fine without the conditions, but obviously want to get the bonus and match print.
Vpt.se & Dan_M's suggestion of leaving the Circle-M off the A datum is also what I recommend.
The problem with MMC-boundary fitting on the primary datum (which, by the way, Xactmeasure can handle perfectly if the datum is a cylinder) is that it allows for 4 degrees of freedom of fitting. This creates results that simply cannot be replicated with mechanical measurements without spending $30k on a complex fixture.
As Vpt.se pointed out, the less fitting means it's harder to pass which may be difficult for manufacturing. I have customers who routinely drop all the Circle-Ms to make it harder on themselves, and the result is their passing-inspection parts always have a safety margin and are never near the true limit of failing.
Vpt.se & Dan_M's suggestion of leaving the Circle-M off the A datum is also what I recommend.
The problem with MMC-boundary fitting on the primary datum (which, by the way, Xactmeasure can handle perfectly if the datum is a cylinder) is that it allows for 4 degrees of freedom of fitting. This creates results that simply cannot be replicated with mechanical measurements without spending $30k on a complex fixture.
As Vpt.se pointed out, the less fitting means it's harder to pass which may be difficult for manufacturing. I have customers who routinely drop all the Circle-Ms to make it harder on themselves, and the result is their passing-inspection parts always have a safety margin and are never near the true limit of failing.