Iterative vs Best-fit Alignment

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Iterative vs Best-fit Alignment

    I have a free-form sheet metal part I need to compare to a model. I establish an external alignment on a fixture then place the part on the fixture and recall that alignment. I then measure four corner points and do an iterative alignment to roughly locate the part. Next I measure the datum points per the drawing (A1, A2, A3, B1, B2, C) and perform another iterative alignment to fine-tune the part location. Once the alignment is complete I utilize the Linear Open Scan to generate 80+ points that are measured and report the T value of each. I then construct a feature set comprised of all the measured points and report the profile along with min/max values.

    I've recently discovered the Best-fit alignment which brings me to my question. What is the difference between iterative and best-fit? Any recommendations as to which would be better in this application? I haven't experimented with best-fit much but the one time I did, I performed a 3D best-fit alignment using all 80+ vector points that were created and measured then generated the same report. The resulting numbers were better (closer to nominal) than the ones related back to the iterative alignment. Which is the best method to use? I have included a section of the drawing if it helps. Thanks!



    drawing.jpg

    drawing 2.jpg

  • #2
    IF you have datums called out, you use iterative alignment, if you do NOT have datums, you can use best fit.
    sigpic
    Originally posted by AndersI
    I've got one from September 2006 (bug ticket) which has finally been fixed in 2013.

    Comment


    • #3
      You kind of answered your own question, in an odd way. Iterative alignments were first written such that they only accepted 6 inputs, i.e. 3-2-1. At that time, Best Fit alignments were developed to take many more (99 at the time, might be higher now) inputs. Since then, I believe itarative alignments were modifed to accept more inputs as well.

      In your case, you aligned to 6 points and then inspected 80 points, relative to the 6. After that, you re-aligned to the 80 and dimensioned those same 80 points, so the alignment is directly tied to those being inspected, they are one and the same, so their location is averaged to "best-fit" the 80, and any error is spread across the 80 as opposed to being based on the original 6.

      Comment


      • #4
        What may help you is if you dimension the datum points after you have established the best fit alignment. This will help you see what effect the best fit alignment had.
        Lately, it occurs to me
        What a long, strange trip it's been.

        2017 R1 (Offline programming)

        Comment


        • #5
          Thank you guys. I was thinking iterative was correct but wanted more opinions on this. And John, the datum point dimensions were out up to .006 mm after doing a best-fit alignment.

          Comment


          • #6
            Originally posted by jszanto View Post
            Thank you guys. I was thinking iterative was correct but wanted more opinions on this. And John, the datum point dimensions were out up to .006 mm after doing a best-fit alignment.
            Just a thought....some organizations allow deviation of datum point locations.
            Boeing does, on forgings/castings and even on machined parts when the datum points are specifically defined.
            .006mm is only about 2-1/2 tenths (.00023").
            Lately, it occurs to me
            What a long, strange trip it's been.

            2017 R1 (Offline programming)

            Comment


            • #7
              Originally posted by John Riggins View Post
              Just a thought....some organizations allow deviation of datum point locations.
              Boeing does, on forgings/castings and even on machined parts when the datum points are specifically defined.
              .006mm is only about 2-1/2 tenths (.00023").
              Sorry, John. That was a typo. Should have been 0.6mm (0.0236").

              Comment


              • #8
                Originally posted by jszanto View Post
                Sorry, John. That was a typo. Should have been 0.6mm (0.0236").
                Yeah, that's a horse of a different color.....
                Lately, it occurs to me
                What a long, strange trip it's been.

                2017 R1 (Offline programming)

                Comment


                • #9
                  Originally posted by John Riggins View Post
                  Yeah, that's a horse of a different color.....
                  http://www.freakingnews.com/pictures...lor--47600.jpg
                  sigpic
                  Originally posted by AndersI
                  I've got one from September 2006 (bug ticket) which has finally been fixed in 2013.

                  Comment


                  • #10
                    Originally posted by Matthew D. Hoedeman View Post
                    Giddy isn't the word...........
                    Poleaxe

                    Beware the environment you live in for it will shape you
                    Be careful the friends you choose for you will become like them.

                    sigpic

                    Comment

                    Related Topics

                    Collapse

                    Working...
                    X