Is this important?
I would imagine so, considering all the parts have datums, but what is the difference from, say, setting a point as Datum A and dimensioning a feature to Datum A, and setting a point and dimensioning a feature to said point?
As far as I understood it, I can take a point on 1 end of a part, and take a point on the other end. If I want to find out the distance between the two points, it shouldn't matter that 1 is a datum and 1 is not. It should matter that I want the distance between the two points. But if I take a point in the middle of a round flat feature, and another at the top of a father away round flat feature, if I want the distance in Y I would ask for the distance in Y, and it would give me the distance in Y, and setting one or the other or both to a datum wouldn't make a difference because Y is still Y and the points haven't moved.
It sounds like you aren't talking about Datum Definition at all. That's a specific thing where you are using XactMeasure and when making a Position dimension, you define the datum for use in building your DRF to the Feature Control Frame.
Distances can be 2D or 3D, and can be parallel or perpendicular to an axis or a feature. If you have a datum feature to another feature, the datum feature should be picked 2nd. If you have a line or plane, the vector of the distance can be related to the vector of the feature. Point to point, you need to specify a direction, either to an axis or 3D. It all depends on the feature, the specific callout, the print, etc.
A 3D distance is measured from centroid to centroid, regardless of orientation. 2D distance is workplane and feature dependent. You must be in the right workplane and specify the right measurement orientation.
If you are measuring a point to a point, you don't need to set the Datum Definitions in PC-DMIS. As Vinni said, you would need to set them if you are using Xact Measure and dimensioning a GD&T callout.
Make sure your alignment are rock solid as well. You could be dimensioning something incorrectly if it is not.
On a SINGLE point, to another SINGLE point, it shouldn’t matter. But if it’s between two planes, or two cylinders, it certainly could. It all depends on the size, vector and distance between the features in question.
It depends on what you have your dimension reporting set for... if it's set for legacy it will be the window on the left, if it is set for Xact measure it will look like the window on the right.
So the general consensus is I should set features as datums, if anything for posterity's sake, considering I take more than a single point to define a couple features.
Works for me!
So a datum definition question in and of itself:
I have a rotary table that clamps down cylindrical parts with a flat face. This face is usually datum B, my Y origin. Datum A is the OD of the cylindrical part itself, which I level to. Datum C is undefined, so I set up points and lines to generate a rotation vector.
Would I be better off defining these features as Datums? Almost all measurements use at least one of these features in its' construction.
Also along these lines, I'm having an issue with Datum B...
I am uneasy with the rotary table and subsequent calibration; I'm not seasoned in PC-DMIS, rotary tables, machining, or anything related to any of them in any way. So I do not trust it's accuracy.
When it rotates to 45°, I am pretty sure it's actually 45°. But I'm not sure how tilted the part is in the rotary table, if that makes sense.
Say I want to generate a plane as Datum B. To do that I would want to rotate to each of the relevant vectors (every 30° or 45° typically) and take a point on the face, then construct a plane out of these points.
How can I be sure that the Y axis is level? I would imagine if the part is tilted sideways a little in the rotab, a point on the left side would be farther away than a point on the right, for example.
Would I be better off trying it and reporting the flatness of the plane?
Is using a single point for a datum such as a Y axis insufficient?
Is using a singe point at each rotation and constructing a plane going to be insufficient due to part sway/tilt from the chuck?
Is there a way to test this?
Is this better asked in a different forum?
Who knows?
