Alingment point C1

[VarSlovenia]

Hallo,
i have problem creating a alignment. Alignment point C1 has deviation. Why is that happening?
Capture.JPG

[NinjaBadger]

What deviation? In what direction?

If you ask for its location AFTER your alignment, it will probably have a deviation in X but zero in y.


Also without seeing more you're probably doing it incorrectly anyway - PNT_C1 has an x vector - you're rotating from centre of circle to this point, which will simply depend on where you measured the point!

[VarSlovenia]

What deviation? In what direction?

If you ask for its location AFTER your alignment, it will probably have a deviation in X but zero in y.


Also without seeing more you're probably doing it incorrectly anyway - PNT_C1 has an x vector - you're rotating from centre of circle to this point, which will simply depend on where you measured the point!

How it is right? I used:
- Plane A1 - Level Z, Origin Z
- Line from centre of circle to point C1 - Rotate X Plus about Z plus
- Circle B1 - translate X and Y
Drawing.JPG

[vpt.se]

Use a line on the lower edge of the part for aligning in X.
The way you align now will create a rotated alignment and just like NinjaBadger already stated, the C1 point will have a deviation in X and most probably in Z, but none in Y.

PC-DMIS is doing just what you tell it to do, but that is not the way you want it

Capture.JPG
Two points on the edge (green) create a line and align using the line in X. Pay attention to the created lines direction.

[VarSlovenia]

Use a line on the lower edge of the part for aligning in X.
The way you align now will create a rotated alignment and just like NinjaBadger already stated, the C1 point will have a deviation in X and most probably in Z, but none in Y.

PC-DMIS is doing just what you tell it to do, but that is not the way you want it

Capture.JPG
Two points on the edge (green) create a line and align using the line in X. Pay attention to the created lines direction.

But the customer requires a alignment as in the drawing.
So the current procedure is wrong?

[NinjaBadger]

Hi Var,

As I said I needed to see more which is why I said 'probably'. That drawing's pretty horrible.

You're using Dat C as a clocking point, vpt is suggesting the edge itself should be clocked to - but he may have not fully studied the drawing, the line he 'highlighted' would be perpendicular to A (and having a co-linear vector) and therefore not clocked to in the manner he suggested.

Having Dat C on the perpendicular edge, it's actually very close to datum B - unless that edge is very square (which I assume from a stamping it won't be) then depending on where you (as the programmer) decide to take that clocking point will create massive variance in alignment and hence the result of the profile control frame.

Furthermore there's no basic dimension from Dat B to Dat C (which whilst isn't the end of the world makes the design intent more tricky to gauge)

See attached

https://www.dropbox.com/s/sqn2p3kkceqbhor/stamping.jpg?dl=0

It it was me I'd seek clarification on the design intent.

IF I couldn't (for whatever reason) I'd approach it like this.

I assume you have CAD? I'd do an iterative alignment on A (3 several points) B (circle) and C (single point) - this keeps the alignment as per model. I'd then take several hits on (what I perceive to be an 'L' shaped) datum C and I'd best fit rotate to them about B.


But really I'd get clarification of design intend.

[vpt.se]

Clarification...
The line would be created between the green points, from left to right, giving the line a positive X direction, parallel/coplanar(?) to surface A.

I would align like this:

Datum A for primary, level and zero in Z
The line in X
X/Y origin in the hole
Then measure away in DCC, hopefully using CAD...

Datum C is more of a line than a point, so if datum C isn't "perfect", the measured point will vary depending on where it is taken (manually) and subsequently as will the alignment (just like the badger already mentioned).

[NinjaBadger]

But Dat C isn't on the edge you've identified at all? Or am I misreading the drawing?

[vpt.se]

No, you are completely correct. My suggestion is to use that edge as a way to quickly achieve an alignment that will work according to the drawing dimensions and get to the DCC stage to continue measuring away. I believe this method is "safer" and probably more consistent than manually probing a point "somewhere" on C.

The idea is then to continue measuring (datum C for example) in DCC and hopefully in conjunction with a CAD model.