Skewed Plane and Skewed Hit Results

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

  • Skewed Plane and Skewed Hit Results

    Having some issues with a program, I've got a work around, but some of the things I'm seeing has raised some questions. So I'm running an experiment. So far I've ran the same fixtured part without moving it 9 times. Any insight on why the demon is doing what it's doing in these 3 numbered issues? Thanks.


    1) Part is nearly perpendicular to MCS. At this point in manual alignment, I've leveled to X+, Rotated to Z+, and translated to X & Z. The final 2 points establish the Y axis translation.

    If I lock my Z axis travel, 6 out of 6 times I get good results between the 2 points and the program runs.
    But if I don't lock the Z axis travel, 3 out of the 4 times the hit results make no sense at all. The Y & Z values should be within .5" of each other at absolute worst case scenario.

    1st run, Point B4 has Y @ 8.799, Z @ -1.9582. Point B5 has [email protected] .8592, Z @ -7.5833
    2nd run ok, Y values within .15" of each other
    3rd run, Point B4 has Y @ -3.1692, Z @ -7.1652. Point B5 has [email protected] 5.3791, Z @ -2.4503
    4th run, Point B4 has Y @ 7.6684, Z @ -1.3922. Point B5 has [email protected] .4793, Z @ -8.263

    I reviewed the data ahead of this, and every time the vector of level plane is .999... in X and rotate line vector is .999... in Z, so nearly perpendicular to machine. So why would my results be so skewed in Y and Z when they should be reasonably close to each other?


    2) 2 out of 6 times it has skipped the incremental move to clear the part right after going in to DCC mode. Any ideas why?


    3) One time it created the plane below. All 3 of my points are nearly perfect X plus vector, with not much deviation in the X results. Yet the plane constructed from them is this skewed 3D monstrosity that would have crashed the head if I wasn't watching it once going into DCC. Not sure on how the math behind the plane creation works, but in my mind, if all 3 points have the same vector, than that's the vector that the plane is created on. Any insight?

    Thanks

    X PLANE SKEWED.JPG

  • #2
    Can you post your code with actual measured values included?

    Comment


    • #3
      For #1, as mentioned above, we would need to see your code for the alignment and features in question.

      For #2, This happens to me sometimes when I have a MOVE/CLEARPLANE command right after a CLEARP/ command. PC-DMIS can get ahead of itself sometimes, try putting a MOVE/INCREMENT,<0,0,0> right before the move that it sometimes skips. This way, if PC-DMIS gets ahead of itself, it should skip the zero move and execute the actual move.

      For #3, your picture is not showing up, so I cannot see what you're talking about. But the vector and position of a plane is calculated ONLY from the measured XYZ values of the points that make it up. It has nothing to do with the vectors of those points.

      Comment


      • #4
        1. Code added below. It doesn't include the measurement values listed above, but values from a good run. We don't save anything other than the most recent run.

        3. Reattached picture, hopefully it works this time.

        Thanks for taking the time giving me input.

        PNT_X1_MAN =FEAT/CONTACT/VECTOR POINT/DEFAULT,CARTESIAN
        THEO/<0,0.327,0.8112>,<1,0,0>
        ACTL/<14.1465,8.745,-26.0887>,<0.9999955,0.0029955,0>
        TARG/<0,0.327,0.8112>,<1,0,0>
        SHOW FEATURE PARAMETERS=NO
        SHOW CONTACT PARAMETERS=YES
        SHOW HITS=NO
        PNT_X2_MAN =FEAT/CONTACT/VECTOR POINT/DEFAULT,CARTESIAN
        THEO/<0,0.9037,0.6963>,<1,0,0>
        ACTL/<14.1316,9.4387,-25.9086>,<0.999991,0.0042462,0>
        TARG/<0,0.9037,0.6963>,<1,0,0>
        SHOW FEATURE PARAMETERS=NO
        SHOW CONTACT PARAMETERS=YES
        SHOW HITS=NO
        PNT_X3_MAN =FEAT/CONTACT/VECTOR POINT/DEFAULT,CARTESIAN
        THEO/<0,2.5106,1.8651>,<1,0,0>
        ACTL/<14.1274,10.6118,-25.088>,<1,0,0>
        TARG/<0,2.5106,1.8651>,<1,0,0>
        SHOW FEATURE PARAMETERS=NO
        SHOW CONTACT PARAMETERS=YES
        SHOW HITS=NO
        PLN_X_MAN =FEAT/PLANE,CARTESIAN,TRIANGLE,NO,LEAST_SQR
        THEO/<0,1.2471,1.1242>,<1,0,0>
        ACTL/<14.1352,9.5985,-25.6951>,<0.9986558,0.0320459,-0.0407382>
        CONSTR/PLANE,BF,PNT_X1_MAN,PNT_X2_MAN,PNT_X3_MAN,,
        OUTLIER_REMOVAL/OFF,3
        FILTER/OFF,WAVELENGTH=0
        MAN_LEVEL =ALIGNMENT/START,RECALL:STARTUP,LIST=YES
        ALIGNMENT/LEVEL,XPLUS,PLN_X_MAN
        ALIGNMENT/TRANS,XAXIS,PLN_X_MAN
        ALIGNMENT/END
        TIP/T1A90B180, SHANKIJK=0, -1, 0, ANGLE=180
        COMMENT/OPER,NO,FULL SCREEN=YES,AUTO-CONTINUE=NO,

        LOCK (X) AXIS TRAVEL ON CONTROLLER FOR NEXT (2) HITS

        PNT_Z1_MAN =FEAT/CONTACT/VECTOR POINT/DEFAULT,CARTESIAN
        THEO/<-2.33,-0.09,-0.589>,<0,0,-1>
        ACTL/<-2.3723,7.3564,-27.0372>,<0.0407382,-0.0013062,-0.999169>
        TARG/<-2.33,-0.09,-0.589>,<0,0,-1>
        SHOW FEATURE PARAMETERS=NO
        SHOW CONTACT PARAMETERS=YES
        SHOW HITS=NO
        PNT_Z2_MAN =FEAT/CONTACT/VECTOR POINT/DEFAULT,CARTESIAN
        THEO/<-2.33,-0.09,1.811>,<0,0,1>
        ACTL/<-2.4675,7.3899,-24.6711>,<-0.0408175,-0.0011703,0.9991659>
        TARG/<-2.33,-0.09,1.811>,<0,0,1>
        SHOW FEATURE PARAMETERS=NO
        SHOW CONTACT PARAMETERS=YES
        SHOW HITS=NO
        LIN_Z_MAN =FEAT/LINE,CARTESIAN,UNBOUNDED,NO
        THEO/<-2.33,-0.09,-0.589>,<0,0,1>
        ACTL/<-2.4199,7.3564,-27.0372>,<0,0.0141389,0.9999>
        CONSTR/LINE,BF,2D,PNT_Z1_MAN,PNT_Z2_MAN,,
        OUTLIER_REMOVAL/OFF,3
        FILTER/OFF,WAVELENGTH=0
        MAN_ROT =ALIGNMENT/START,RECALL:MAN_LEVEL,LIST=YES
        ALIGNMENT/LEVEL,XPLUS,PLN_X_MAN
        ALIGNMENT/ROTATE,ZPLUS,TO,LIN_Z_MAN,ABOUT,XPLUS
        ALIGNMENT/TRANS,XAXIS,PLN_X_MAN
        ALIGNMENT/TRANS,ZAXIS,PNT_Z2_MAN
        ALIGNMENT/TRANS_OFFSET,ZAXIS,0.519
        ALIGNMENT/END
        WORKPLANE/ZMINUS
        TIP/T1A7.5B0, SHANKIJK=0, 0.1305, 0.9914, ANGLE=0
        COMMENT/OPER,NO,FULL SCREEN=YES,AUTO-CONTINUE=NO,

        LOCK (Z) AXIS TRAVEL ON CONTROLLER FOR NEXT (2) HITS

        PNT_B4_MAN =FEAT/CONTACT/VECTOR POINT/DEFAULT,CARTESIAN
        THEO/<0.42,0,-1.719>,<0,1,0>
        ACTL/<0.3641,7.7629,-1.4061>,<0,1,0>
        TARG/<0.42,0,-1.719>,<0,1,0>
        SNAP=NO
        SHOW FEATURE PARAMETERS=NO
        SHOW CONTACT PARAMETERS=YES
        SHOW HITS=NO
        PNT_B5_MAN =FEAT/CONTACT/VECTOR POINT/DEFAULT,CARTESIAN
        THEO/<-5.08,0,-1.719>,<0,1,0>
        ACTL/<-4.9622,7.9156,-1.6212>,<0,1,0>
        TARG/<-5.08,0,-1.719>,<0,1,0>
        SNAP=NO
        SHOW FEATURE PARAMETERS=NO
        SHOW CONTACT PARAMETERS=YES
        SHOW HITS=NO
        LIN_B_MAN =FEAT/LINE,CARTESIAN,UNBOUNDED,NO
        THEO/<-5.08,0,-1.719>,<1,0,0>
        ACTL/<-4.9622,7.9156,-1.5136>,<0.9995891,-0.0286632,0>
        CONSTR/LINE,BF,2D,PNT_B5_MAN,PNT_B4_MAN,,
        OUTLIER_REMOVAL/OFF,3
        FILTER/OFF,WAVELENGTH=0
        MAN_ALIGN =ALIGNMENT/START,RECALL:MAN_ROT,LIST=YES
        ALIGNMENT/LEVEL,XPLUS,PLN_X_MAN
        ALIGNMENT/ROTATE,ZPLUS,TO,LIN_Z_MAN,ABOUT,XPLUS
        ALIGNMENT/TRANS,XAXIS,PLN_X_MAN
        ALIGNMENT/TRANS,YAXIS,LIN_B_MAN
        ALIGNMENT/TRANS,ZAXIS,PNT_Z2_MAN
        ALIGNMENT/TRANS_OFFSET,ZAXIS,0.519
        ALIGNMENT/END
        Attached Files

        Comment


        • #5
          One thing jumps out the vectors of the 2 hits used for the line are the same as the line vector that is wrong, for a Z vector line the hits should be in the X or Y.

          Comment


          • Woody S
            Woody S commented
            Editing a comment
            Can you elaborate a bit more please? You're comment has piqued my interest. Obviously a line is simply two points connected, but I learn more every day how finicky PC-DMIS can be. So making a line thru the vectors, opposing vectors at that does seem like something that may cause problems in the PC-DMIS world. Just wondering what the impact could be. Thanks

        • #6
          This is not about PC-DMIS being finicky but good metrology practice and understanding, see attached PDF as they say a picture says a thousand words.

          By locking the X Axis as you have the line created in the X workplane may not be representative of how the part could be orientated as the hits are free to
          be places along the Y axis. Locking the Y axis would not help either as this line would be the machine coordinate system Z axis and may not represent the part orientation.
          Attached Files
          Last edited by UKCMM; 01-04-2021, 05:24 AM.

          Comment


          • Woody S
            Woody S commented
            Editing a comment
            Thanks UKCMM. For this part it wouldn't make a difference, flange is only .18" thick, so line can't skew very much. But trying to understand the bigger picture here.

            So since this is a 2D line with an X Workplane, the X values get averaged out and the line can be anywhere in the YZ plane, going from one YZ to the next. So hitting several inches apart in Y would create quite an angled line. But if I'm hitting X vectors I could still travel inches in Y and create that skewed line. So it seems like the way would be to hit only Y Vectors as far apart in Z as the part configuration allows, to control that actual part orientation and not create that skewed line. Is that correct, or am I still missing something?

            Or if I wanted to use X vector hits, change the workplane to Y?

            Thanks

          • UKCMM
            UKCMM commented
            Editing a comment
            To rotate about the X axis the 2D line needs to be created in the X workplane using either a Z vector line ( min 2 off Y+ or Y- hits ) or a Y vector line ( min 2 off Z+ or Z- hits ).
            If you want to use a Z vector 2D line using X vector hits in the Y workplane you cannot use that the rotate about the X axis.

            Can you attach a sketch of the part to better understand the features you are measuring.

          • Woody S
            Woody S commented
            Editing a comment
            Thanks for your feedback UKCMM, when I get a little opening in time I'll mock something up.

        Related Topics

        Collapse

        Working...
        X