hexagon logo
Register Sign in
  • Replies 10 replies
  • Subscribers 221 subscribers
  • Views 112 views
  • Users 0 members are here
Options
Related

Alignment Question. BF Line between Holes vs Autoalign Cylinders

Lukas
Hello. I am programming a part using legacy. The top plane is the primary datum and the seconary/tertiary datums are two thru holes that are square to the top plane. So rotate to the line between the holes of course.

I have profile of a surface called out to these datums.

I tried two alignments:

- One alignment uses cylinders. I autoaligned to the top plane and the 2 cylinders.

- The other alignment I created a 2d best fit line between the cylinders. I autoaligned to the top plane, the best fit line, and one cylinder as my tertiary datum.

My profile of a surface looks crooked when I use the first alignment, and it looks much better using the second alignment. Can someone explain why my results would be different? Should I go with alignment #2? Please see screenshots.
Attached Files
  • sounds to me like the 2 cylinders aren't 'coaxial' in reality. They may be to design, but not in the part.

    In 30+ years, I've never used any of the 'autoalign' functions in the alignment, I pick what I want for what I want.

    As for using the cylinders, the best-fit line is probably using the centroid of each cylinder while the other one may be using the entire cylinder, but I am not sure.

    I would check the coaxaility of the cylinders.
  • sounds to me like the 2 cylinders aren't 'coaxial' in reality. They may be to design, but not in the part.

    In 30+ years, I've never used any of the 'autoalign' functions in the alignment, I pick what I want for what I want.

    As for using the cylinders, the best-fit line is probably using the centroid of each cylinder while the other one may be using the entire cylinder, but I am not sure.

    I would check the coaxaility of the cylinders.

    Thanks for your input! I agree, the parallelism between the 2 cylinders must be the issue. Method 2 (BF line method) correlates better with GeoTol. I think this is my second time seeing this same problem. I am going to veer away from using autoalign.
  • Go with what Matt said. All of these gimmicks to make programming apparently "faster" are just sales pitch to QA managers who have no idea how CMM really works.
  • in reply to MattR
    Thank you, Matt. I will stay away from Autoalign from now on, especially when it comes to Cylinders. I wonder if it would have worked ok if the holes were circles.
  • So weird. I did some quick tests in the software.

    I created pierce points where the cylinders meet the primary datum plane (top plane).

    If I autoalign rotate to those points:again, bad results. If I create a line between the points and rotate to that line: good results.

    At this point I really have to stay away from autoalign...I don't get how it's making the calculation whatsoever.
  • in reply to MattR
    i like the auto align feature. as long as you pick the correct datums then it just builds the alignment a little bit faster.
  • in reply to Lukas
    i don't think you guys are using autoalign correctly. autoalign assumes you are selecting your datums in a correct order. if you don't will get different results becaause the alignments arent the same.
    did you try this exact same test without using auto align? what were your results?
  • BIGWIG7,

    Yes I tried that yesterday on the phone with Hex. They are going to review this further. It seems like they can't figure it out off the bat. I think I made a mistake by mentioning Autoalign in this thread without giving all the details. I don't think autoalign is the problem. Basically, I am selecting two circles for the rotational alignment in "Alignment #1." You can see the code in my screenshot. It seems fine...Level to A and "ROTATE_CIRCLE" to the two cylinders.

    My take on the problem is that "ROTATE_CIRCLE" is what is causing the problem (not Autoalign). Autoalign seems to be working properly.
  • in reply to Lukas
    yea all auto align does is create an alignment faster rather than picking each individual feature and locking down your 6 dof.
    autoalign just does it with less mouse clicks. Slight smile
  • in reply to Lukas
    maybe (MAYBE) the issue is that the circles you are using are not in the same plane as the plane you are leveling?????

    I always take the 'long way around' (might take 3 seconds longer) and do not rely on anything 'automatic' in Pcdmis. Even when I have 2 holes IN the plane I am leveling to, I construct a line from hole to hole and rotate to that.
© 2024 Hexagon AB and/or its subsidiaries. Privacy Policy | Cloud Services Agreement