alignmet whit 4 features or more for plane

You can use as many features as you want for an iterative alignment. Unless the part is perfect, you will see error, Pcdmis will balance the error out. So if you are using 4 for the level, you might see two of them at -0.023 and two of them at +0.023 or whatever it happens to be. It will be balanced, it will not be perfect (because there are no perfect parts).

You can use as many features as you want for an iterative alignment. Unless the part is perfect, you will see error, Pcdmis will balance the error out. So if you are using 4 for the level, you might see two of them at -0.023 and two of them at +0.023 or whatever it happens to be. It will be balanced, it will not be perfect (because there are no perfect parts).

so, due that it will be balanced, no problem with the results?..are fine..isn't wrong?..can I use and shared that data without problem and take desicions?

I recently had a part with a contoured surface where the entire surface was defined as the primary datum, no target points. I set up the iterative alignment in two different ways to see what it would do:

1) 20 pts on the contoured surface all used for "Level" in the iterative alignment.
- reported out the surface profile (form only) w/textual analysis, Zaxis location, and vector of a point set created from all 20 pts after the alignment
2) All 20 of the same points but only used 3 of them for the "level" in the iterative alignment.
- reported out the surface profile (form only) w/ textual analysis, Zaxis location, and vector of a point set created from all 20 pts after the alignment

The results only differed by a few .0001's for the individual surface profile points, no difference in surface profile form, and a few .0001's difference in Z location. That is what I expected to see and verified that either approach would work equally well for these parts, so I went with the 20 pts for level as it better matched the drawing intent. You could do a similar type of comparison for your parts to verify that the results that you get are what you expect and are sufficiently accurate for you to share the data and make decisions.