Dimension reporting help

[Jay_Q]

I’m working on a part that has a flange as datum A and a perpendicular cylinder as datum B. My problem is there is a flat on the cylinder which is also perpendicular to the datum A flange. This flat has a hole going through it which is perpendicular to the datum B cylinder. The call out is .001|A|B to this hole going through the flat. What’s the best way to report this? According to this datum scheme there is nothing controlling rotation. So depending on which random feature I rotate to I get different results.

[SingularitY]

Any degrees of freedom that isn't being constrained by a datum in a reference frame are controlled by the feature itself.
There are probably a few ways you can do this and I'm sure it depends on who you ask.
1) Some people would draw a line from the feature to datum B and rotate to it, but that will force the hole to be in line with B and you'll essentially only have a deviation from A and a worst axis cylinder position to B. (Not ideal unless it's a very shallow hole and on the side of the cylinder directly, not on a flat. )
2) If the hole has some depth to it, you can pick it up as a cylinder and rotate only to this cylinder. This will properly tell you how far off center from B the feature really is.
3) You can rotate to the flat inside of the AB alignment, then create a best fit alignment recalling this alignment, select the cylinder, select 2D no translation (rotation only) using least squared, hit compute and create. This should give you similar results to number 2. I prefer this method and number 2 if the hole is long enough for a cylinder.
4.) The easiest way it rotate to the flat the hole is located. For a shallow hole, I would prefer this over number 1.

[Jay_Q]

Thanks for the response. I would normally go with 4) and not think twice but with a position of .001 every .0001 counts. I like option 2) it would be perfect but I can only probe .130 long cylinder. I’m not sure that would be enough. 3) I’m a little confused 2d best fit to only datum B cylinder? If they were both in the same working plane I would normally 2d best fit (rotate only) to the datum B cylinder and the feature. If it was to say .001|B only

[SingularitY]

#3 A 2D best fit rotation only alignment to the cylinder FEATURE you are wanting to dimension. But you would most likely want to recall an AB alignment that rotates to the flat so you get correct nominals for the best fit.
I'm assuming we are talking about inches here. You can always try option 2 and see if a .130in long cylinder will be enough to rotate to. You may be surprised. If your company requires a GR&R, you can always dimension that feature all 3 ways (2, 3, and 4) and see which method is more repeatable. (My bet is on 4.) Number 2 would be most proper and will give the customer exactly what they want, but with GR&R requirements and the need for repeatability, sometimes a programmer must use what will still give the customer what they want but also give the repeatability required by their employer. (#4)
But without testing, it's really up to you.

[Jay_Q]

Ok. Gotcha. I figured you meant the feature. I’m probably going to stick with #4 that was my original thought. I will try #2 for fun on the .130” cylinder. I was just exploring options to see if I was going about this the proper way or if I was missing something. Thanks again for the help!

[SingularitY]

No problem