Features Used in Alignment Don't Zero after Alignment

[[Pe...]]

I am seeing a periodic issue that features used to construct an alignment are not zero in actual column for the appropriate axis.

In the enclosed picture I did not show the results for Clock 3d line since it is based on two holes in a bolt circle and used as a common reference for discussing results with CNC folks.

I have a work-around but I don't know if the issue of these features not being zero for appropriate axis should be ignored.

Zeiss calypso zero alignment issue.jpg

[[Ja...]]

Are they being evaluated the same? For instance is Datum B LSQ in the base and Outer Tangential in the feature window?

[[Ju...]]

Also, wouldn't perfect form be required for the origin of the part to be measured at exactly 0?

[[Ow...]]

Are you doing a manual alignment first?
Why are you recalling feature points on Datum B?
If cylinder A isn't perfectly perpendicular to plane B, the A1, A2 are not going to be zero.
A lot of things to consider but, generally your rotation in space datum is going the be datum A if the print calls it out that way but, you need to make sure the length of the cylinder is long enough.
Posting a picture of the drawing or part would help.

[[Er...]]

I have seen the same issue when using a cylinder in an alignment.
As best I could figure, where the cylinder intersects the plane will be zero.

[[To...]]

I can explain why the Datum A cylinder is not zeroed out. When you're looking at the Y and Z values, they are the values at the feature's trihedron, which is nowhere near the BA trihedron. This is because the cylinder has some angular deviation in the A1 and A2. However, where the axis of the cylinder intersects the Datum B plane, I would bet good money that it was X0/Y0/Z0. To test my statement, create an intersection of the cylinder axis and the plane.

Now, I can't quite explain why the Datum B plane is showing those deviations but maybe because you're recalling features point??? I don't know off hand.

If you used the Datum A cylinder as the spatial, I would expect everything to be zero but would also expect to see some angular issues on the B plane and some deviation the X value because it's origin is at the corner of the plane. However, where the cylinder axis pierced the plane, again, would be X0/Y0/Z0.

[[Ri...]]

As a practice for myself.
Base Alignment features are created as simple as possible.
Planes with minimum points, any other 3D features I do not use.
The Base Alignment should be for describing that there is something within the volumetric envelope of the CMM. It doesn't have to be the actual Datums, use what makes it easy to probe manually.
All features should be LSQ.
(Correct me if I'm wrong) Features used in Base Alignments ignore Filters and Outliers.

[[Ma...]]

Features in the base alignment use whatever evaluation settings are set for the feature including fit, filters and outliers. Of course, all of these are subject to limitations due to number of points, etc. for example none of these settings can possibly apply to a plane or circle established from 3 points.

[[Ri...]]

Please sign in to view this quote.

I kind of recall a class, from Zeiss, a while back where it was stated that Base Alignment features ignore filters/outliers/etc. To be more specific during the calculation of the base alignment.
Now I'm going to spend the rest of my life looking for that information. 🧑‍💻

[[To...]]

Please sign in to view this quote.

I don't believe this is true. I believe that was the whole point of apply filters and outliers at the feature level instead of at the characteristics level. Either location would produce the same output but if applied at the characteristics level, the features used in the BA would not take advantage of the filters and outlier elimination, which would make the BA more stable.

[[Ri...]]

From Zeiss:
If they are enabled then yes the filters and outliers will be applied but they may not calculate until after the entire alignment is measured. Many things relative to the base alignment are not calculated until after the entire process is completed unless you "force" it with a formula.

I'll go back thru my ancient CMM notes, I may have misinterpreted what I was told. 😕

[[Je...]]

Please sign in to view this quote.

.

This is an excellent definition of the purpose of a base alignment. I'm going to borrow it in the future 😃
The way I usually describe it is "base alignments are primarily for navigation and only secondarily for evaluation."

This is also the reason I like to introduce form error to the base alignment by including at least four points on a plane, six points on a cylinder, etc.. It shows the CMM how a particular workpiece is located within the measuring envelope and not just a theoretically perfect version of the workpiece. This is similar to the reason we loop base alignments.

.

[[Ri...]]

Please sign in to view this quote.

I've taught classes, and always said, the Base Alignment should never be used for measurements of features.
Make the Base Alignment as simple as possible and it doesn't matter if its not Datums or a Datum structure from an FCF.
Create the necessary Alignments for when the CMM is in CNC mode.
This also removes complicated Base Alignments.