Alignment with a plane, a hole and a rest block

[MUlissi]

Hey folks,

I'm doing an alignment on a part and I can't get it to lock in correctly. I assume I'm either doing something wrong, or the model is incorrect.

The picture is a rough sketch of the idea of what I'm working with.

Datum B is a hole punched through the metal.
Datum A is 3 locations to take points from to construct a plane for leveling.
Datum C is based off of a surface with a draft that butts up against a spring loaded pin on the gauge to hold the part in place.

I've tried a few different approaches including:

• Measuring 2 edge points along the bottom of the draft surface and rotating to that.
• Measuring a plane for Datum C on that draft surface and rotating between a constructed point and Datum B.
• Using the gauge model to plot the exact center of where the pin sits against a nominal part to use as a rotation point between it and Datum B.

Assuming the CAD I'm working with is correct, is there something else I should be doing or is one, (or are a few,) of those approaches correct, and which?

I would have thought a line constructed between two edge points along the bottom edge of the drafted surface was the best solution.

I'm using Iterative alignment to keep the part in car body position, and it keeps throwing my Datum B circle off to the right and up when I execute the alignment as well as acting like the surface of Datum C is off location.

alignment.jpg

Thanks in advance again!

[louisd]

Datum B does not produce constraint for rotation, as it should.
A levels the plane and origins in Z
B only translates XY origin
C can only translate if it is a tertiary datum.

C COULD be used as a rotation datum, but only if Feature Control Frame (FCF) calls out [A][C][B].

If your FCF calls out [A][B][C] kick the drawing back to your customer or engineer if it's an internal print.

[MUlissi]

Oh, I heavily dislike the layout, but as I've found so far, these have generally been agreed upon before they reach me and I'm expected to just make it work. They don't always supply me with the FCF or even a proper print.

Thus my dilemma.

I doubt it matters, but B is also referred to as L1, and C as L2.

[louisd]

On the drawing for this part, do you have profile or tolerance of position controls?
--A FCF is the box that defines those dimensional control strategies (True positions, profiles, concentricities, runouts, perpendicularities, etc all reference datums within the defined FCF)

if you don't have any dimensional controls that reference datums, what's the use of even defining them?

If you are just looking for a way to control all 6 DOF for measurement purposes on the CMM,
Level to A plane
Construct an intersect line between A and C planes. Rotate to the intersect line.
Origin Z to A plane
Origin Y and X to your B circle.
you should be set from there

[MUlissi]

I'm familiar with FCFs. I don't have a part drawing/print for this and to the best of my knowledge there isn't one they can provide to me.

They've given me pictures of the CAD pointing out where they want the points of alignment to be located.

This is unfortunately fairly common.

[AndersI]

Datum B does not produce constraint for rotation, as it should.
A levels the plane and origins in Z
B only translates XY origin
C can only translate if it is a tertiary datum.

Where in the standard(s) is it stated that the second datum *must* lock rotation? The rule is "can, may, must" - if a datum *can* lock a DOF, and it *may* do so (it's not already locked), then it *must*.

A - level, z origin
B - X, Y origin
C - rotation, for example with the help of an intersect line as louisd already have said, or a line perp to C, through B.

If PC-DMIS refuses to accept [A|B|C], I'd suggest it's a limitation of PC-DMIS...

[MUlissi]

I've seen a few limitations to PC-DMIS already, and a lack of limitations in things that can break PC-DMIS that could reasonably/easily be implemented.

[RanDawgg]

I think it's a perfectly valid drawing at [A][B][C]

[Darrollh]

I could be wrong for this. I would just build an alignment as normal then pick up the datums and use exact measure to build the FCF. That’s how I would solve the issue.

[MUlissi]

Conclusion: For anyone interested in the resolution.

I went with plotting a line at the intersection of Plane A and a plane built on the surface where C sits, then plotting a line perpendicular to this line through B. Then I constructed a point at the intersection of this line, and the line it is perpendicular to.

I performed an iterative alignment as follows:
Level to the 3 points that make Plane A.
Rotate using B and the intersection point, which I labeled C.
Origin to B.

Based on the output this gave me, I was reasonably sure my model was incorrect. Went to talk to someone about that, and they confirmed that the material direction on my model was ISM, not OSM as I had been led to believe, thus the largest part of my issue.

Using this info, and the suggestions from you folks, it works.

And everyone lived happily ever after (if not a little annoyed at not having the last bit of info from the outset.)

The End.
conclusion.png