hexagon logo
Register Sign in
  • Replies 22 replies
  • Subscribers 222 subscribers
  • Views 198 views
  • Users 0 members are here
Options
Related

CMM Won't Position

BL Metrology
When I set the head to any angle other than A0B0, the CMM will not return to what I set as the X,Y,Z origin during alignment.

For Example:
At an angle of A7.5B-90, if I touch a point on the granite surface, set an alignment as X,Y,Z origin, and then program it to go back to 0,0,0, it will actually go to 0.0145, 0.0094, -0.4356. All of the programming is done in DCC Mode. This problem also occurs when I go through a full alignment on any part.

We do not have this problem when at A0B0, if I do the same thing at A0B0 it will return to 0,0,0 within tenths of a micron. The deviation with the angle is seen with all angles we have tried so far and the deviation does seem to increase with the A angle. All angles used are successfully calibrated prior to use.

The machine was calibrated on 3/17/09 and the technician spent an extra two days trying to troubleshoot the unit (Hexagon). We have also had our local reps in and they have not been able to troubleshoot the problem. The machine seems to be mechanically fine and it seems to be an issue with PC-DMIS.

Has anyone else experienced this and know of a workaround? We are waiting for Hexagon/Wilcox to resolve this, but I am assuming it could take some time.
  • Anyhow, are you moving in the Z direction, where your error is?

    I'm curious about movement for the hit as well. Are the vectors for both points 0,0,1 (or the same really, shouldn't matter what they are as long as they are the same)?
  • JEC_31 - Great thought, thanks for the information. We have a renishaw head. They came in and changed the entire head and still had the same problem.
  • Here is the code, Z is fine, but the X & Y are off way too much. I used an angle of A45B-90 that has been successfully calibrated.

    PART NAME : rotate
    REV NUMBER :
    SER NUMBER :
    STATS COUNT : 1

    STARTUP =ALIGNMENT/START,RECALL:,LIST=YES
    ALIGNMENT/END
    MODE/MANUAL
    TOUCHSPEED/ 1
    FORMAT/TEXT,OPTIONS,ID,HEADINGS,SYMBOLS, ;NOM,TOL,MEAS,DEV,OUTTOL, ,
    LOADPROBE/PS1X20
    TIP/T1A45B-90, SHANKIJK=0.7071, 0, 0.7071, ANGLE=-90
    MODE/DCC
    PNT1 =FEAT/POINT,CARTESIAN
    THEO/<35.3481,152.2747,-587.4724>,<0,0,1>
    ACTL/<35.3186,152.2722,-587.4723>,<0,0,1>
    MEAS/POINT,1
    MOVE/CLEARPLANE
    HIT/BASIC,NORMAL,<35.3481,152.2747,-587.4724>,<0,0,1>,<35.3186,152.2722,-587.4723>,USE THEO = YES
    ENDMEAS/
    A1 =ALIGNMENT/START,RECALL:STARTUP,LIST=YES
    ALIGNMENT/TRANS,XAXIS,PNT1
    ALIGNMENT/TRANS,YAXIS,PNT1
    ALIGNMENT/TRANS,ZAXIS,PNT1
    ALIGNMENT/END
    MOVE/POINT,NORMAL,<64,47,90>
    PNT2 =FEAT/POINT,CARTESIAN
    THEO/<0,0,0>,<0,0,1>
    ACTL/<-0.0287,-0.0025,-0.0004>,<0,0,1>
    MEAS/POINT,1
    MOVE/CLEARPLANE
    HIT/BASIC,NORMAL,<0,0,0>,<0,0,1>,<-0.0287,-0.0025,-0.0004>,USE THEO = YES
    ENDMEAS/
  • ...Metric
     

    PNT2       =FEAT/POINT,CARTESIAN
            THEO/<0,0,0>,<0,0,1>
            ACTL/<[COLOR="Red"]-0.0287[/COLOR],[COLOR="Navy"]-0.0025[/COLOR],-0.0004>,<0,0,1>
            MEAS/POINT,1
            MOVE/CLEARPLANE
            HIT/BASIC,NORMAL,<0,0,0>,<0,0,1>,<-0.0287,-0.0025,-0.0004>,USE THEO = YES
            ENDMEAS/

    Remembering that it is in metric, that converts to .0011 inch. While not to be dismissed, it isn't exactly huge either.
    Again, metric, that's only .0001 inch - you're unlikely to get better than that.

    Can you run us through the procedure you use to calibrate the angles?
  • Remembering that it is in metric, that converts to .0011 inch. While not to be dismissed, it isn't exactly huge either.
    Again, metric, that's only .0001 inch - you're unlikely to get better than that.

    Can you run us through the procedure you use to calibrate the angles?

    +1

    The feature that you are measuring has NO X or Y components. Only Z!
    The error you see is the inablilty of the CMM to maintain the position of the probe in an exact position. Completely within any reasonable expectation of the machine to hold. The Z direction has the surface of the granite to control when the probe is activated. There is no surface to control the X or Y
  • +1

    The feature that you are measuring has NO X or Y components. Only Z!
    The error you see is the inablilty of the CMM to maintain the position of the probe in an exact position. Completely within any reasonable expectation of the machine to hold. The Z direction has the surface of the granite to control when the probe is activated. There is no surface to control the X or Y
    And THUS the reason for SNAP points.

    There is a setting you can try to set, F10 set up, under the OPTIONAL PROBE tab, there is a POSITIONING ACCURACY you can change. Of course, this WILL have to be done in every program since it is only an OPTIONAL feature. If you set it too low, the machine will never "set" and it will sit there hunting for hours for those fine micron's of position.
  • +1

    The feature that you are measuring has NO X or Y components. Only Z!
    The error you see is the inablilty of the CMM to maintain the position of the probe in an exact position. Completely within any reasonable expectation of the machine to hold. The Z direction has the surface of the granite to control when the probe is activated. There is no surface to control the X or Y

    Good point. I missed or didn't think about that. The X and Y values are what the machine is capable of driving itself to. The Z value is what the machine is capable of measuring. A 3D feature (sphere) or 1 feature in each axis would be a better test.
  • The X&Y error is seen on 3D features such as spheres as well. We actually noticed it initially while trying to do a scan across a sphere.

    The angles are calibrated using:
    Operation -> Calibrate/Edit -> Active Probe -> select the angles -> Measure -> select options (DCC+DCC, etc.) -> Then run the calibration
  • The X&Y error is seen on 3D features such as spheres as well. We actually noticed it initially while trying to do a scan across a sphere.

    The angles are calibrated using:
    Operation -> Calibrate/Edit -> Active Probe -> select the angles -> Measure -> select options (DCC+DCC, etc.) -> Then run the calibration

    What about your answer to the "Has the calibration sphere moved?" question?
© 2024 Hexagon AB and/or its subsidiaries. Privacy Policy | Cloud Services Agreement