Runout Question

Measure datum feature A and datum feature B as individual cylinders (or construct a cylinder from all measured hits). Define your datum as both features and report the runout with respect to the common datum A-B. I'd avoid leveling to the constructed line as a line is interpreted as a 2D feature and will inherit the orientation of your workplane.

What version of PC-DMIS are you working with?

I'm working with 2022.1. I've read on here that its good to level to a line as i have done, maybe I misinterpreted it.

Can we measure 2 circles on A & 2 on B build a single cylinder & define that as datum A-B?

As JacobCheverie said, the correct way to measure the datums is to measure them as cylinders, using enough hits and levels to ensure you capture a true representation of the surface. xbigfatgoat, you say you are using 2022.1 which means you have access to the geometric tolerance command. That makes it really easy - measure datum A as a cylinder and then define as datum A, measure datum B as a cylinder and define as datum B then, from the datum definition dialog, click the "common datum" checkbox and select datum A and B from the list to define common datum A-B. Finally, create your runout.

Whenever you are using geometric tolerance commands, you should always use the simplest approach possible, sticking to measured-features, auto-features or scans wherever possible. This is because the geometric tolerance command takes a completely different approach to previous dimensioning methods, acting on the hits of the features and datums rather than taking their pre-resolved solutions. BF/BFRE constructions can be used where necessary, because they store the hit information. However, most other constructions just make life more difficult and lead to problems because they do not contain any hits.

Most of the other approaches discussed on this forum - constructing points from centroids, measuring as circles, building lines - are attempts to work around the limitations of old CMM software and dimensioning tools. With PC-DMIS legacy dimensioning for example, the user is required to interpret the print and then build an alignment representing the datum reference frame - hence the common approach of constructing a 3D line through the centroids of circles.

Thanks for the very informative reply Neil. I have one more question if you don't mind, should I be leveling to just one of the cylinders or both of the cylinders? They are different diameter's so that's what throws me off.

For an alignment, the line you described in your first post would be ok - as long as it is a 3D line.

I measured three separate circles for each journal, constructed cast points at each circle and then used those points to construct a line. I levelled to this line.


​

Just remember that the geometric tolerance command works independent of any alignment so, if you create a geometric tolerance runout following the approach I described earlier and try to compare those numbers to a legacy runout using the alignment, the numbers won't match.

make sure the strategy of hits & spacing for datum A and datum B are equal to each other.
EX: If datum A has 300 hits and 30 circles, and datum b has 100 hits and 10 circles, your datum axis will be biased to the projection of datum A's axis.