Screenshot_20200804-042135_Samsung Notes.jpg

I am inspecting a part that looks something like this. Please excuse the scrappy drawing.

I did a rough alignment to other features on the part (not shown here but I can assure you its a "stable" alignment to synchronize the part and model) and then I measured a lot of vector points on datum A. reported the profile and it was good. I then moved onto measuring datums B and C.

This is how I defined my datums:

I made a feature set from the vector points taken on datum A, a cylinder from datum feature B and a width from datum feature C.

The print actually specifies perpendicularity of B with respect to A (which is wrong in my opinion because datum A surface is nominally conical and datum feature B is not nominally perpendicular to datum feature A). So I went ahead and dimensioned datum B using position. Here's my issue, when I dimension datum B with respect to Datum A (the feature set) I get it ~.070 off and that's unbelievable but when I measure datum C with respect to A primary (the feature set) and B secondary (the cylinder) I only get a few tenths of an inch off nominal and its in tolerance.

How can I ensure the axis of datum B is not grossly off with respect to datum feature A? I thought creating a feature set from the vector points on Datum A would work but that only worked for C, atleast I think it did.. PCDMIS reported C with respect to AB as conforming.

I tried an alternative inspection method where I made a plane from 3 arbitrary points on the datum A surface (I chose points that were spaced out the furthest) and mentioned it to engineering that they may consider using target points to establish datum plane A. So I did that, and it worked.. atleast I think it did because PCDMIS reported B and C conforming in that manner. By using 3 target points to construct a plane, I am assuming that PCDMIS is comparing the theoretical vector of the plane constructed from the 3 target points to evaluate the vector of datum B with respect to datum A, right? I am trying my best to think of the most functional outcome.

Back to my sketch though... does anyone have any input?

-Why are the results of B with respect to A (feature set) wonky?

-Why are the results of C with respect to A (feature set), B (cylinder) NOT wonky?

-Would you also recommend datum targets on datum feature A instead? (This seems functionally reasonable per ASME Y14.5 2009)

Much appreciated. Thank you thank you thank you!

I am inspecting a part that looks something like this. Please excuse the scrappy drawing.

I did a rough alignment to other features on the part (not shown here but I can assure you its a "stable" alignment to synchronize the part and model) and then I measured a lot of vector points on datum A. reported the profile and it was good. I then moved onto measuring datums B and C.

This is how I defined my datums:

I made a feature set from the vector points taken on datum A, a cylinder from datum feature B and a width from datum feature C.

The print actually specifies perpendicularity of B with respect to A (which is wrong in my opinion because datum A surface is nominally conical and datum feature B is not nominally perpendicular to datum feature A). So I went ahead and dimensioned datum B using position. Here's my issue, when I dimension datum B with respect to Datum A (the feature set) I get it ~.070 off and that's unbelievable but when I measure datum C with respect to A primary (the feature set) and B secondary (the cylinder) I only get a few tenths of an inch off nominal and its in tolerance.

How can I ensure the axis of datum B is not grossly off with respect to datum feature A? I thought creating a feature set from the vector points on Datum A would work but that only worked for C, atleast I think it did.. PCDMIS reported C with respect to AB as conforming.

I tried an alternative inspection method where I made a plane from 3 arbitrary points on the datum A surface (I chose points that were spaced out the furthest) and mentioned it to engineering that they may consider using target points to establish datum plane A. So I did that, and it worked.. atleast I think it did because PCDMIS reported B and C conforming in that manner. By using 3 target points to construct a plane, I am assuming that PCDMIS is comparing the theoretical vector of the plane constructed from the 3 target points to evaluate the vector of datum B with respect to datum A, right? I am trying my best to think of the most functional outcome.

Back to my sketch though... does anyone have any input?

-Why are the results of B with respect to A (feature set) wonky?

-Why are the results of C with respect to A (feature set), B (cylinder) NOT wonky?

-Would you also recommend datum targets on datum feature A instead? (This seems functionally reasonable per ASME Y14.5 2009)

Much appreciated. Thank you thank you thank you!

## Comment