Interpretation of Position callout

[[Ja...]]

I am trying to clarify how to evaluate this true position. I have a slot that is positionally toleranced to a planar primary datum, cylindrical secondary datum, and cylindrical tertiary datum. There is a basic angle from the tertiary datum to the center of the slot.

I assume that this position is comprised of (a) how centered the slot is on the secondary datum, and (b) how close to the basic angle the slot is from the tertiary datum.

My question is this: Assume I can measure the error from both (a) and (b). How do I combine that information to obtain my result?

Obviously I will level to A, rotate to C and origin XY to B and Z to A. My coordinate system will be on A, centered on B, and clocked to C. This Position establishes a planar tolerance zone opened up at the basic angle on the print, and I am to check whether or not the midplane of the slot falls in the zone.

This wouldn't be the standard 2D/3D Cartesian 2*sqrt(dx^2 + dy^2 + dz^2) nor is it the 2D polar 2*sqrt(r^2 + r0^2 - 2rr0cos(dA)) true position formula....

[[An...]]

Seems to be True Position in Y-direction.
See attached.

Contribution_11_12_2019.pdf

[[Ja...]]

Andreas,

I do not like the idea of your "second alignment". In other words, I establish my ABC coordinate system, and am asked for the position in that coordinate system of the slot width. To rotate back by the basic angle technically puts me in a coordinate system that is different, no?

But I do understand that this may be the only practical way to make this measurement as it was intended. I think theory and practice do not align on this one.

Thank you.

[[An...]]

Please sign in to view this quote.

It puts you back in the direction of tolerancing.
See attached.

Contribution_12_12_2019.pdf

[[Ja...]]

Please sign in to view this quote.

Thanks Andreas,

This "direction of tolerancing" that you mention, is that from the implied coordinate system of the drawing?

[[Ke...]]

I concur with Andreas' suggestion.
True Positions are referenced back to print the basic dimensions. Rotating the part (unless drawn differently) usually puts you back in-line with the prints basic dimensions.
As a sanity test for yourself, check the positional result comparing the two alignment methods (with and without rotation). The Actual should not change from one method to the other, as rotating the alignment rotation does not change the calculation of the position. But adding the rotation allows you to see the actual deviation from X,Y basics (if you report the Additional printout).

[[Br...]]

You don't need to make a "Secondary Alignment". Just reference your [A|B|C] alignment for the DRF in the Position Characteristic, then click the "Special" button and you can offset it from there.

[[An...]]

The "Direction of Tolerancing" is defined by the arrow,
that comes from the "True Position"-tolerance frame.

Contribution_12_12_2019_a.pdf

[[Ja...]]

Thanks everybody,

I will have to examine prints of this variety a little more carefully going forward. This is not the way that I naturally understand this dimensioning. I have always assumed that the coordinate system from which you take your measurements is exactly the same as the coordinate system defined by your DRF and that no translation/rotation is allowed (without material boundaries). I guess the implied zero basic of the slot should have been key as to how my coordinate system was to be defined, it's just a little backwards to me.

[[Br...]]

Please sign in to view this quote.

Simply moving your coordinate system doesn't change the fact that it is based off of the datum features. You're just moving it along basic dimensions to make it mathematically easier for you to calculate. Think about it. If you fixed a coordinate system based off of datum features, then shifted it 1 inch closer to the feature being toleranced to that datum reference frame. You would change its nominal distance by an Inch. Now the error you get will be the exact same as if you moved it back 1" and changed the nominal 1" again. Same goes for angles. It's effectively how you use a SINE Plate.

[[Ja...]]

Thanks Brett,

That makes sense, I was just confused when it came to the planar tolerance zone located at the basic angle. A planar zone is limited in one dimension, infinite in the other two. To try to interpret a planar zone in two dimensions seemed strange. I haven't seen something like that before and certainly didn't connect the idea that I could just rotate my system to align with the theoretical 0 point. Things are much clearer now.

[[Br...]]

Please sign in to view this quote.

I think you were just confusing it with how a Datum Shift would work. In that case you would shift the Coordinate system but the nominals would stay constant allowing optimization. If the nominals always change relative to where the coordinate system moves, you will not affect your result.

[[Ja...]]

Please sign in to view this quote.

Well, I wasn't thinking about the datum shift explicitly but in a sense you are correct. The key there is your last sentence. I wasn't thinking from that perspective. Thank you.