Iterative Help

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

  • Iterative Help

    OK..in 10 years I have never done an iterative alinment.

    1 part is floating around in CAR coordinates.. there are no flat surfaces on this part, everything has a slight curve to it. Print calls out Datum Targets for leveling ect.

    2. I tried using an iterative alignment to get the machine "snapped" to the part, but I keep getting an "Iterative alinment error" I am at a loss.. Some tricks with this would be helpfull. I used 3 points(z), 2 circles(y), 1 circle(x).
    I have z-plus workplane selected

    3. Does it matter that I changed the part/machine (F5) orientation?
    Code:
            WORKPLANE/YPLUS
    PNT5       =AUTO/VECTOR POINT,SHOWALLPARAMS = NO
                THEO/143.8769,31.8168,83.6762,-0.0380018,0.9462436,0.3212148
                ACTL/-12.1191,-26.2549,7.0008,0,1,0
                TARG/143.8769,31.8168,83.6762,-0.0380018,0.9462436,0.3212148
    PNT6       =AUTO/VECTOR POINT,SHOWALLPARAMS = NO
                THEO/144.3498,33.7112,77.8112,-0.0266004,0.9652142,0.2601038
                ACTL/-11.2067,-26.1707,0.8139,0,1,0
                TARG/144.3498,33.7112,77.8112,-0.0266004,0.9652142,0.2601038
    PNT7       =AUTO/VECTOR POINT,SHOWALLPARAMS = NO
                THEO/139.62,32.682,79.0515,0.0191001,0.9561049,0.2924015
                ACTL/-15.9508,-26.8224,1.9246,0,1,0
                TARG/139.62,32.682,79.0515,0.0191001,0.9561049,0.2924015
    CIR1       =AUTO/CIRCLE,SHOWALLPARAMS = YES,SHOWHITS = NO
                THEO/139.7638,33.0709,77.7559,0,0.9549078,0.2969024,0.3937
                ACTL/-15.7286,49.5856,24.6406,0,0.9549078,0.2969024,0.3937
                TARG/139.7638,33.0709,77.7559,0,0.9549078,0.2969024
                THEO_THICKNESS = 0,RECT,IN,CIRCULAR,LEAST_SQR,ONERROR = NO,$
                AUTO MOVE = NO,DISTANCE = 0,RMEAS = None,None,None,$
                READ POS = NO,FIND HOLE = NO,REMEASURE = NO,$
                NUMHITS = 4,INIT = 0,PERM = 0,SPACER = 0.05,PITCH = 0,$
                START ANG = 0,END ANG = 0,DEPTH = 0.1,$
                ANGLE VEC = 1,0,0
    CIR2       =AUTO/CIRCLE,SHOWALLPARAMS = YES,SHOWHITS = NO
                THEO/142.3401,30.8348,84.5772,0,0.9548795,0.2969936,0.3937
                ACTL/-14.0181,47.4565,31.136,0,0.9548795,0.2969936,0.3942
                TARG/142.3401,30.8348,84.5772,0,0.9548795,0.2969936
                THEO_THICKNESS = 0,RECT,IN,CIRCULAR,LEAST_SQR,ONERROR = NO,$
                AUTO MOVE = NO,DISTANCE = 0,RMEAS = None,None,None,$
                READ POS = NO,FIND HOLE = NO,REMEASURE = NO,$
                NUMHITS = 4,INIT = 0,PERM = 0,SPACER = 0.05,PITCH = 0,$
                START ANG = 0,END ANG = 0,DEPTH = 0.1,$
                ANGLE VEC = 1,0,0
                WORKPLANE/ZPLUS
    A0         =ALIGNMENT/START,RECALL:STARTUP, LIST= YES
                ALIGNMENT/ITERATE
                  PNT TARGET RAD = 0.04, START LABEL = , FIXTURE TOL = 0, ERROR LABEL = 
                  MEAS ALL FEAT = ONCE ,MAX ITERATIONS = 99 ,LEVEL AXIS=ZAXIS ,ROTATE AXIS=YAXIS ,ORIGIN AXIS=XAXIS
                LEVEL = PNT5,PNT6,PNT7,,
                ROTATE = CIR1,CIR2,,
                ORIGIN = CIR1,,
                ALIGNMENT/END
    sigpic
    "You cant vote 'present' as the President of the United States"

  • #2
    use surface hits on your circles!
    sigpicV2014mr1

    Comment


    • #3
      Paul...Where you at?
      ~Josh~ Giles Rescue "Ut ceteri vivant"

      Comment


      • #4
        try this (from my personal 'help' file):

        Alignments:

        First of all, to do ANY alignment, 3-2-1 or iterative, you MUST have XYZ values for whatever you use for the alignment.

        Step #1
        Do this no matter WHAT type of alignment you are going to do.
        1) Place the part on the table the way it will sit for your inspection
        2) Import your model data BEFORE you do any measuring
        3) Using the CTRL-F3 option (3-D rotate), rotate the cad data until it matches the way the part sits on the machine, remembering that as you look at the screen you are looking DOWN on the machine
        4) Once you have the cad data rotate (you are actually ONLY rotating the VIEW of the cad data), use F5 to open the setup options.
        5) Go to the Part/Machine tab and while looking at the 3-D rotate cube, set your CAD to MACHINE axis
        6) Click OK
        7) Set your workplane to the ‘top’ of the cube (if the top, facing YOU face of the cube is X-, use the X- workplane)

        You are now ready to do an alignment.

        Iterative alignment
        1) You need to know WHAT the datums are that you have to use for alignment. Use CTRL-F to open the AUTO FEATURES window
        2) Select the type of feature you are going to use (vector, surface, circle, etc.) and pick that feature from the model
        3) Create the feature, BUT DO NOT MEASURE IT!
        4) Continue until you have all the features you need for your alignment, remember, ALL circular feature types NEED a minimum of 3 surface sample hits (circles, cylinders, cones, slots)
        5) Print out the graphics screen for a reference of where the points need to be measured
        6) Mark all the features and execute them, measure them as close to where they need to be as you can. OPTIONAL, do this here or at STEP 13
        7) Go into the alignment and then into ITERATIVE alignment
        8) Select the (minimum of) 3 ‘level’ features
        9) Select the (minimum of) 2 ‘rotate’ features
        10) Select the (minimum of) 1 ‘origin’ feature
        11) Set the tolerance to high values, I use 10mm radius and 1mm fixture
        12) OK out of the iterative window and the alignment window
        13) Mark all the features an execute the program, and measure the features as close to where they go as you can, this is the second place you can do this, but if you didn't do it as step 6, DO IT NOW!
        14) You are now aligned
        15) Turn on DCC mode and repeat steps 1 to 10, making the program executable, meaning it will run without crashing. HOWEVER, for the tolerances, at this point, I use 0.5 for radius and 0.05 for fixture. ALSO, mark the MEASURE ALL ALWAYS box in the iterative window
        16) Now, program the checks of the part

        3-2-1 alignment (mostly for holding fixtures)
        This is a very easy alignment, however, you will have to think back to 2nd or 3rd grade when they taught you to use a number line (you’ll see)
        1) All the alignment features on the fixture SHOULD have XYZ values stamped on them. For this example, I will use a plane, and 2 lines (top corner of the fixture base)
        2) Measure the plane, making sure that all the touches are EXACTLY where you want them to be
        3) Measure a line (the longer edge of the base) for the ROTATE, making sure of the DIRECTION of the line
        4) Measure the second edge of the base (direction makes no difference)
        5) Construct a point at the intersection of the 2 lines
        6) Now, you need to know exactly WHAT each of those features represent. We will say that the PLANE is for the X+ axis, the rotate line points in the Y+ direction (first to last point taken on the line) and we need to know they XYZ value for that corner of the fixture base
        7) Open the alignment window
        8)Level X+ to the plane
        9) Rotate Y+ ABOUT the X+ axis to the rotate line
        10) Set the X origin to the plane
        11) Set the Y axis to the corner point
        12) Set the Z axis to the corner point
        13) Now, it is time to offset the origins. This is where the number line comes into play
        IF the X value for the top of the base is equal to 1250mm, then you need to offset the X origin –1250mm. What you are doing is telling Pcdmis the direction (in this case -) and the amount to move the ZERO point, which right now is located at the corner of the base. This will make the base equal to X1250.

        Do this for each of the axis, remembering that you are telling it the DIRECTION and AMOUNT to move the origin AWAY from where you have the origin set to.

        So, for the corner values of the fixture at X1250, Y-250 and Z2321, your offsets will be X-1250, Y250 and Z-2321

        14) Click on CAD=PART
        15) Close the alignment window.
        You model should now be ‘floating’ in space about the alignment features just like it does on the fixture.
        16) Go into DCC mode and program the alignment as an executable program (one that will run without crashing). The reason for this is simple, the machine will give better, more accurate touches than you, the operator can. The speed of the touch CAN and DOES have a big impact on the final results Pcdmis see for the touch. Faster or slower can change the comp of the point, so use DCC. Not only that, but if you have to re-home the machine or re-calibrate the probes (for whatever reason), you need ONLY run the DCC portion of the alignment, the manual can be left un-done, as long as the fixture has not been moved on the table
        16) Program and check the part.

        I use non-iterative alignments almost exclusively and I never have any problems or issues with them. I also keep the alignment as a separate program, saving the alignment to an external file, then recalling that external file in my check programs. In this way, I can have multiple programs to check various portions/operations of the part WITHOUT the need to align multiple times. Sure, you can cram it all into one program, but that will slow Pcdmis down. Also, by using separate check programs, I can send the various data sets to different Datapage files.

        The biggest advantage to the 3-2-1 CAD=PART alignment is it's ease in adjustment. You can easily adjust your pick-up values to fine-in a holding fixture without trying to figure out what-in-the-**** each individual point in the iterative alignment needs to be to get the desired effect, just change the offset for that axis, find and click on CAD=PART again (do it twice, in V3.7 MR3 it does not 'catch' it the first time) and re-check the features.
        sigpic
        Originally posted by AndersI
        I've got one from September 2006 (bug ticket) which has finally been fixed in 2013.

        Comment


        • #5
          Originally posted by Boomer View Post
          OK..in 10 years I have never done an iterative alinment.

          1 part is floating around in CAR coordinates.. there are no flat surfaces on this part, everything has a slight curve to it. Print calls out Datum Targets for leveling ect.

          2. I tried using an iterative alignment to get the machine "snapped" to the part, but I keep getting an "Iterative alinment error" I am at a loss.. Some tricks with this would be helpfull. I used 3 points(z), 2 circles(y), 1 circle(x).
          I have z-plus workplane selected

          3. Does it matter that I changed the part/machine (F5) orientation?
          Code:
                  WORKPLANE/YPLUS
          PNT5       =AUTO/VECTOR POINT,SHOWALLPARAMS = NO
                      THEO/143.8769,31.8168,83.6762,-0.0380018,0.9462436,0.3212148
                      ACTL/-12.1191,-26.2549,7.0008,0,1,0
                      TARG/143.8769,31.8168,83.6762,-0.0380018,0.9462436,0.3212148
          PNT6       =AUTO/VECTOR POINT,SHOWALLPARAMS = NO
                      THEO/144.3498,33.7112,77.8112,-0.0266004,0.9652142,0.2601038
                      ACTL/-11.2067,-26.1707,0.8139,0,1,0
                      TARG/144.3498,33.7112,77.8112,-0.0266004,0.9652142,0.2601038
          PNT7       =AUTO/VECTOR POINT,SHOWALLPARAMS = NO
                      THEO/139.62,32.682,79.0515,0.0191001,0.9561049,0.2924015
                      ACTL/-15.9508,-26.8224,1.9246,0,1,0
                      TARG/139.62,32.682,79.0515,0.0191001,0.9561049,0.2924015
          CIR1       =AUTO/CIRCLE,SHOWALLPARAMS = YES,SHOWHITS = NO
                      THEO/139.7638,33.0709,77.7559,0,0.9549078,0.2969024,0.3937
                      ACTL/-15.7286,49.5856,24.6406,0,0.9549078,0.2969024,0.3937
                      TARG/139.7638,33.0709,77.7559,0,0.9549078,0.2969024
                      THEO_THICKNESS = 0,RECT,IN,CIRCULAR,LEAST_SQR,ONERROR = NO,$
                      AUTO MOVE = NO,DISTANCE = 0,RMEAS = None,None,None,$
                      READ POS = NO,FIND HOLE = NO,REMEASURE = NO,$
                      NUMHITS = 4,INIT = 0,PERM = 0,SPACER = 0.05,PITCH = 0,$
                      START ANG = 0,END ANG = 0,DEPTH = 0.1,$
                      ANGLE VEC = 1,0,0
          CIR2       =AUTO/CIRCLE,SHOWALLPARAMS = YES,SHOWHITS = NO
                      THEO/142.3401,30.8348,84.5772,0,0.9548795,0.2969936,0.3937
                      ACTL/-14.0181,47.4565,31.136,0,0.9548795,0.2969936,0.3942
                      TARG/142.3401,30.8348,84.5772,0,0.9548795,0.2969936
                      THEO_THICKNESS = 0,RECT,IN,CIRCULAR,LEAST_SQR,ONERROR = NO,$
                      AUTO MOVE = NO,DISTANCE = 0,RMEAS = None,None,None,$
                      READ POS = NO,FIND HOLE = NO,REMEASURE = NO,$
                      NUMHITS = 4,INIT = 0,PERM = 0,SPACER = 0.05,PITCH = 0,$
                      START ANG = 0,END ANG = 0,DEPTH = 0.1,$
                      ANGLE VEC = 1,0,0
                      WORKPLANE/ZPLUS
          A0         =ALIGNMENT/START,RECALL:STARTUP, LIST= YES
                      ALIGNMENT/ITERATE
                        PNT TARGET RAD = 0.04, START LABEL = , FIXTURE TOL = 0, ERROR LABEL = 
                        MEAS ALL FEAT = ONCE ,MAX ITERATIONS = 99 ,LEVEL AXIS=ZAXIS ,ROTATE AXIS=YAXIS ,ORIGIN AXIS=XAXIS
                      LEVEL = PNT5,PNT6,PNT7,,
                      ROTATE = CIR1,CIR2,,
                      ORIGIN = CIR1,,
                      ALIGNMENT/END
          Use surface hits on circles and level to all points and circles. then offset to the two circles and orign to one of them.
          sigpichttp://npmichaels.com/ (My sons book)
          THANKS (CAD IS ALWAYS MASTER)
          4.3 & 4.2MR2/GLOBAL/PH10MQ/WINDOWS XP

          Comment


          • #6
            Originally posted by pniven View Post
            Use surface hits on circles and level to all points and circles. then offset to the two circles and orign to one of them.
            he can't do that unless those circles are also 'level' datums, you can ONLY use what the print tells you that you can use, you can't add datums.....
            sigpic
            Originally posted by AndersI
            I've got one from September 2006 (bug ticket) which has finally been fixed in 2013.

            Comment


            • #7
              Originally posted by Matthew D. Hoedeman View Post
              he can't do that unless those circles are also 'level' datums, you can ONLY use what the print tells you that you can use, you can't add datums.....
              You can use that to get the cad right then type in the right xyzijk values for what the print calls for and realign to the right datum’s for the final adjustment.
              sigpichttp://npmichaels.com/ (My sons book)
              THANKS (CAD IS ALWAYS MASTER)
              4.3 & 4.2MR2/GLOBAL/PH10MQ/WINDOWS XP

              Comment


              • #8
                alignment error

                One thing I've found over the years with iterative alignments. If you try and
                do one and it gives you an error. It's best to start again with a new program, sometimes the error will repeat or come back again. It's usually at the beginning anyway so it's not a big deal.......
                ________________________
                DEA SCIROCCO-15-10-09
                TUTOR P -4.2 MR1-XP

                Comment


                • #9
                  No, but he could "iterate" to it.
                  sigpicV2014mr1

                  Comment


                  • #10
                    +1 Great instructions Matt Boomer what is the tolerance you set the iterative at? The error is telling you the alignment is wrong if it keeps wanting to measure point to point then up your tolerance within the iterative. The best way unless you have a body coordinate fixture is to change your orgin first, if your orgin is not in the datum A plane then you need to level first then orgin then iterative then if you want to get closer do a best fit also. One differance from 321 vrs iterative 321 go off running good or bad iterative tells you that your alignment is skewed. If your alignment is skewed your screwed!!Point blank This is my Patent saying JC


                    In Matts step 4 this is a give me to any setup! JC
                    sigpic
                    Jeremy Cavender
                    CMM Metrology Engineer
                    PC-Dmis 3.2/2010/MR2

                    Comment


                    • #11
                      Designer!!!

                      You want to know what makes an iterative alignment work REALLY WELL????!!!!!!!!!!!!!

                      When the designer gives you the correct model!!!!!!!!!5 hours shot!!!
                      sigpic
                      "You cant vote 'present' as the President of the United States"

                      Comment


                      • #12
                        That sounds normal
                        sigpichttp://npmichaels.com/ (My sons book)
                        THANKS (CAD IS ALWAYS MASTER)
                        4.3 & 4.2MR2/GLOBAL/PH10MQ/WINDOWS XP

                        Comment


                        • #13
                          Ya that would help Why did you Program the wrong Math LOL Revisions and part #'s would be a good sign. Like Bill says here is your sign!! Allways double check JC
                          sigpic
                          Jeremy Cavender
                          CMM Metrology Engineer
                          PC-Dmis 3.2/2010/MR2

                          Comment


                          • #14
                            I think if you change your workplane after you measure your circles to xplus, create a plane in the alignment. Then create a parallel plane with that plane through circle 1. This should sort out your vectors.

                            Comment


                            • #15
                              If you use circles don't you have to use sample hits ? Even if the sample hits are not part of the alignment

                              Comment

                              Related Topics

                              Collapse

                              Working...
                              X