Orientation of a hole on a conical surface

[Gus101]

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!

[YGOLORTEM]

Have you tried to use an iterative alignment or best fit?

[Gus101]

YGOLORTEM thanks for the response, I tried a 3D (least squares) best fit to ABC and the position of B was ~.004" off which would be non conforming but isnt that reporting the location of B with respect to the best fit? How does this tell me how far off B is to datum A? I guess if location is off that much then the vector of B is probably not where it should be? Hmmm... What do you think? (Thanks for making me think deeper on this).

I guess the fact that I am looking at an orientation control of B with respect to A is what threw me off a bit when I looked at the non conforming result (.004") with respect to the best fit. How do I report to engineering the orientation of B with respect to A?

[YGOLORTEM]

Are the axis of 2 features, Datum -A- (cone) and Datum -B- (cylinder) parallel to one another?

Is there a location dimension from center line of -A- to -B-?

[Gus101]

The axes of Datum a and b are definitely not parallel nor are they perpendicular. Think of this as a hole on an ice cream cone. The hole was drilled normal to the surface. Datum A is nominally conical and its just part of a small segment of a cone. This is a bracket where datum A interfaces with a conical counterpart, datum B and C is where it gets bolted from. Attached is my best attempt of a cross section. YGOLORTEM

[Gus101]

Screenshot_20200804-113917_Samsung Notes.jpg

[YGOLORTEM]

So there is no basic angle from -A- axis to -B- axis on the drawing? (or from -A- surface?)

[Gus101]

It's one of those drawings "all dimension not shown are considered basic per the CAD model". No basic dimensions specified on the drawing and I don't know exactly what it is if I try constructing it on PCDMIS.

[Brandnew]

Just a random question regarding the conical part...is the surface all flat as well or does it also curve? The last image looks like its flat?
Random question for learning purposes: The call out of datum B needing to be positioned at .001 in reference to Datum A? Is this just another way to ask for perpendicularity or how do you position a hole within .001 of a feature that is on the actual surface?
thank you

[Gus101]

The surface curves. Its essentially a hole on an ice cream cone. The hole is normal to the surface. The image above is just my best attempt at a sketch. Yes, positioning B in relation to A is another way of controlling the orientation of datum B. (In this case...)

[YGOLORTEM]

Just a random question regarding the conical part...is the surface all flat as well or does it also curve? The last image looks like its flat?
Random question for learning purposes: The call out of datum B needing to be positioned at .001 in reference to Datum A? Is this just another way to ask for perpendicularity or how do you position a hole within .001 of a feature that is on the actual surface?
thank you

-B- looks perp to me as well but -B- could be at an angle to. I know the OP can't post the drawing but I feel like we are missing some valuable information here.

[Gus101]

B is not nominally perpendicular to A. I hope the image below clarifies things a little better. I cannot zoom in any further.. lol

[Gus101]

YGOLORTEM this is my best attempt at the basic angle from A to B axes. (48.4525°)