Alignment iteration only for features specific to secondary alignment

[[Th...]]

I have a part that's going through a drawing change today. The datum structure is being modified to adjust for some manufacturing-related stuff, and it is shifting the secondary datum to a new feature. Currently, my program uses the old datum structure as the base alignment, so I plan on defining the new DRF using a secondary alignment. My problem is that the last time I set an iteration loop on a secondary alignment, it looped the measurement of all 3 datum features, even though the primary datum had not changed. This is something I would like to avoid for this update, as usually it results in a doubling of the number of stylus system changes, and the XT Gold takes something like 40 seconds each time. Since the primary datum feature is the same for both my base and secondary alignments, is it possible for me to get the secondary's iteration loop to only re-measure the other two features?

[[De...]]

No, I don't believe it is going to be possible.

When it iterates it is recalculating the alignment based off of the new measurements, then altering the location of where the measurement of the next iteration is taken to align the new measurements with the results of the last iteration. The purpose is to make sure that your measurements are being taken 'aligned' to your new alignment.

Each iteration it re-aligns, based on the last set of measurements. Are you sure it is actually necessary to iterate the secondary alignment at all. If your process is somewhat stable and the base alignment repeats there should not be much of a reason to loop the secondary alignment unless the difference between it and the base alignment is varying a lot.

[[Th...]]

You're saying that once the base alignment iteration's done its work, I don't need to worry about applying one to my secondary? I'd made the assumption that measuring a different set of features in order to build a new alignment would need a loop/break condition for the same reasons that the base alignment needs it.

[[De...]]

The goal of an iterative alignment is to make sure that the probe is contacting the part in the correct orientation.

Internally in the software, Calypso needs to calculate where the contact points are on any given feature based on where the center of the probe tip is. It uses the known probe radius as well as the nominal geometry to do this. The more accurately the actual scan path is to the nominal scan path geometry, the more accurate the calculated result will be.

An iterative alignment measures the part where it 'thinks' it should be, it then evaluates the points it has measured and makes its best guess as to where the part actually is. When it re-measures it it alters the location and orientation of the scan path the probe travels when scanning the next iteration to more correctly match the orientation of the points it measured in the first iteration.

If there is no reason to believe the measured result is changing then the iterative alignment is likely unnecessary.

One easy method is to check the default printout and check what the delta value was for the alignment for each iteration. If the value changes a lot while it loops, then the iterative alignment was likely to have been useful, however if the delta value is very small then the first location it 'targeted' was very close to the correct location and therefore the measured result would have most likely been pretty close to the same.

The only way to know for sure is to test it and look at the data. It really depends on what your tolerances are. How repeatable your manufacturing process is. In general however, if your process is stable and your parts are pretty accurate between datum features then a secondary alignment probably won't gain you much.

[[Me...]]

Please sign in to view this quote.

This principle should be 1 of the 10 commandments of using a CMM. I've had some pretty interesting discussions about this topic. When I was on The PC-DMIS board it was stated that you must always put your part into the datum alignment scheme before measuring any features that relate to it. That's utter nonsense.

You should use an alignment that produces the least amount of probing/cosine error when taking points on a feature. In other words, probing normal to the surface. That's it. Everything else can be done after probing features by creating more alignments, evaluations, etc. No need to put your part into 6 different alignments while you're probing the features, unless those alignments provide the least amount of probing error.

[[Th...]]

This reads like what I'm already doing, I think. All my features use the base alignment as their workpiece alignment, and my DRF secondary is being used to define evaluation on the characteristic side.

[[An...]]

Thoughts about iteration.
See attached.

Contribution_05_06_2019.pdf