Cylinder true position is correct but the perp. is way out

[ScottP815 Member Find all posts View Profile Close ScottP815]

So I'm using PCDMIS 2020 R1, I have a 2 cylinders that have true position and perp. call outs. This seem redundant because the primary dat is the same on both. But any way. I get a true position of .0015 then the perp is .330 on the same cylinder. Any Ideas

[Schlag PC-DMIS demi-Guru Find all posts View Profile Close Schlag]

How many levels and how many hits you using ? 4 hits on 2 levels means your cylinder could solve 3 different way ( 6 if you want to include vector directions )

[Cris_C Ultra Senior Member Find all posts View Profile Close Cris_C]

Consider the depth of your cylinder. If you are measuring at two levels that are very close together, the best fit algorithm can really get messed up if there is any form issue on the feature. I recently had a large diameter cylindrical feature that was very shallow, like .15” long. The part was out of round .050”. The best fit algorithm saw the shallow oval shape and best fit it by constructing the cylinder at a very severe angle. Seems kinda dumb, but that’s just how the math works out, if you cut a cylinder at an angle you will get an oval. If you measure a very shallow cylinder, the levels are so close together that it is essentially no different than measuring them at the same level, which doesn’t help to construct the true vector of the cylinder.

As Schlag suggests, more hits and more levels can help a lot. I’ll add that separating the levels as much as possible will help too. On the part I mentioned above, I had to measure a number of circles normal to the datum being compared to and constructed a 3D line between them. Then I reported the perpendicularity of the 3D line. There may be more elegant solutions, but it was what I could come up with on a time crunch.

[ScottP815 Member Find all posts View Profile Close ScottP815]

I am doing 5 hits on 3 levels. the hole is .376 diameter x .330 deep

[bfire85 PC-DMIS Guru Find all posts View Profile Close bfire85]

I am doing 5 hits on 3 levels. the hole is .376 diameter x .330 deep

What are your datums? Maybe post some code or a snippet of your drawing?

[NinjaBadger PC-DMIS Guru Find all posts View Profile Close NinjaBadger]

How are you reporting this?

Legacy, Xact, GeoTol?

Perhaps you're reporting position in one location of the Cyl - i.e. Start, Axis Average or End - and it's in Position at that location but still out on the Perp?

It's times like this where Graphical Analysis can really let you know what's going on.

[ScottP815]

can you get a graphical image on PCDMIS . I have used Polyworks and that show you what's going on. I have only seen graphic issues if the aligments are not working on PCDMIS

[ScottP815 Member Find all posts View Profile Close ScottP815]

in the picture datum cylinders E +D are the ones in question Dat A is the surface perp to D+E Dat B is the bottom surface dat C is the far side

[KIRBSTER269 PC-DMIS Guru Find all posts View Profile Close KIRBSTER269]

Looks like to me that Datum A is Parallel to D and E or am I seeing this incorrectly? Your A symbol looks like it's on the bottom surface of that picture.

[ScottP815]

no A is the surface the cyls are on

[ScottP815 Member Find all posts View Profile Close ScottP815]

Kirbster269 you are correct

[KIRBSTER269]

Well then you have to check Parallelism not Perpendicularity. Perpendicularity would be back to B

[KIRBSTER269]

If your creating a Composite then your build would be B|A|C position and B for perpendicular, what are you calling out?

[ScottP815]

i was wrong A is perp to the cylinders

[KIRBSTER269 PC-DMIS Guru Find all posts View Profile Close KIRBSTER269]

Do me a favor Post your code, because what you showed in the picture looks like A is parallel to your holes, B is perpendicular, and your Trihedron isn't even corresponding to what your asking. also what is your callout. A|B|C Position and A Perpendicular?

Capture.JPG

[ScottP815]

you were right i go my datums mixed up working on fixing

[ScottP815 Member Find all posts View Profile Close ScottP815]

Ok Thanks All I am apparently still asleep my datums were mixed up thanks to Kirbster269 for noticing.

[KIRBSTER269]

no problem