Autocailbration losing sphere mid calibration

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

  • Autocailbration losing sphere mid calibration

    Where I work we have 6 CMM's. We use the same AutoCalibrate program on every machine. It works fine on 5 of them, but on one it will run for about an hour touching the sphere and calibrating the angles we have set. Then randomly it will go out in space and try to take a hit in the air. My boss (my companies actual CMM Programmer, I'm new and still learning ) said she has never seen anything like this. We talked with Hexagon support and they us to not reset tips to theo before the calibration and that worked until the next time we calibrated and it lost the sphere mid calibration again. They then advised us to delete all the probe files and recreate them, which I did and the same thing happened. Worked that one time and the next time lost the sphere.

    Maybe someone on here has experienced this or something like this because we don't know what else to do with this machine.

    The code for this autocalibrate program is just some autocalibrate commands for each tip and move points to move above the sphere to change probes. I can post the code if needed. We are using version 2014.1.
    Last edited by HxSwartwood; 07-06-2018, 12:15 PM.

  • #2
    Have you cleaned the scales?
    Tried moving the cal sphere to a different location on the table?
    Sounds more like a hardware issue.

    Comment


    • HxSwartwood
      HxSwartwood commented
      Editing a comment
      Yes we have done both. We've had a tech out here for another problem with the machine and i happened to run it by him and he said it sounded more like a software/program issue lol.

  • #3
    I've seen similar thing happen to me when calibrating steep angles. What's happening to me is the probe is hitting the sphere shank when it's doing initial touches and makes the sphere look bigger than what it is. Than the machine attempts to take a hit on that bigger sphere.
    Orienting the sphere differently (with the shank pointing some other direction) fixes the problem form me after some trial and error. I hope this makes sense.

    Comment


    • HxSwartwood
      HxSwartwood commented
      Editing a comment
      I don't believe we've tried this I'll give it a go.
      Thanks!

  • #4
    Someone is not following proper procedure when qualifying tips . If you know the tipname of the one that is going out in left field. Reset just that tip to theo. You may have more to come as you proceed through the program. Might be better to reset them all, and educate everyone on proper procedure.

    Comment


    • HxSwartwood
      HxSwartwood commented
      Editing a comment
      I just reset the tips to theo and i will see how it works out.
      Thanks!

    • HxSwartwood
      HxSwartwood commented
      Editing a comment
      Seems that resetting the tips was the way to go. Hasn't stopped yet and went past the first probe where it usually stops.

    • DJAMS
      DJAMS commented
      Editing a comment
      Just be sure everyone using the machine understands the qualification process so that it doesn't happen again.

  • #5
    Other than this standard calibration, do you see any information at the faulty workstation that someone is incorrectly qualifing a single or 2 angles? All angles are tethered to a single auto-cal proceedure? As in 1 tool, 1 prompt, all the angles you are trying to qualify? It may not be a bad idea to post the code excerpt... Look in your probe directory for the .results for that particular probe and see if it shows a different date set than the other stored data. Good luck! Rebuilding the probe and angle and parameter sets worked for me when similar happened a few years ago.

    Comment


    • R2ah1ze1l
      R2ah1ze1l commented
      Editing a comment
      Have you tried cycling this 'program' in offline mode? You could then go and look at the data for your specific tip and see what stands out different than with any of your angles that run through properly. (Ruby size is off, XYZ position is off,etc...)

  • #6
    Here you go R2ah1ze1l had to remove the summary or it would be way too long. But i ran the program offline and everything looked just fine. The results files are fine as well.
    Code:
    STARTUP =ALIGNMENT/START,RECALL:USE_PART_SETUP,LIST=YES
    ALIGNMENT/END
    MODE/MANUAL
    FORMAT/TEXT,OPTIONS, ,HEADINGS,SYMBOLS, ;NOM,TOL,MEAS,DEV,OUTTOL, ,
    MOVESPEED/ 100
    LOADPROBE/S6-3X70MM
    TIP/T1A0B0, SHANKIJK=0, 0, 1, ANGLE=0
    COMMENT/OPER,NO,FULL SCREEN=NO,AUTO-CONTINUE=NO,
    Use SPHERE 1
    Marked on actual sphere.
    Select Sphere 1 from probe utilities box.
    AUTOCALIBRATE/PROBE, PARAMETER_SET=ALL-TIPS-WITH-DEFAULTS, QUALTOOL_MOVED=NO,
    SHOW_SUMMARY=NO, OVERWRITE_RESULTSFILE=NO
    MODE/DCC
    MOVE/POINT,NORMAL,<231.857,410.598,-271.801>
    LOADPROBE/S5-4X110MM
    TIP/T1A0B0, SHANKIJK=0, 0, 1, ANGLE=0
    AUTOCALIBRATE/PROBE, PARAMETER_SET=ALL-TIPS-WITH-DEFAULTS, QUALTOOL_MOVED=NO,
    SHOW_SUMMARY=NO, OVERWRITE_RESULTSFILE=YES
    MOVE/POINT,NORMAL,<231.857,410.598,-271.801>
    LOADPROBE/S2-2X70MM
    TIP/T1A0B0, SHANKIJK=0, 0, 1, ANGLE=0
    AUTOCALIBRATE/PROBE, PARAMETER_SET=ALL-TIPS-WITH-DEFAULTS, QUALTOOL_MOVED=NO,
    SHOW_SUMMARY=NO, OVERWRITE_RESULTSFILE=YES
    MOVE/POINT,NORMAL,<231.857,410.598,-271.801>

    Comment


    • R2ah1ze1l
      R2ah1ze1l commented
      Editing a comment
      Make a real parameter set, not just 'all-tips-with-defaults' . You can have 'Reset to Theo' selected and enforce which tool (Sphere 1 as seen above) will ALWAYS be used. As acgarcia and others on this forum will suggest, use a 'master probe' to define ALL tool locations in 1 shot if possible, then run the rest with qualtool_moved=no. Good luck, glad your 'reset to theo' worked! (this round)

  • #7
    Do you have a parameter set for your master probe? Do you calibrate/locate the sphere with your master probe first?

    My cal programs calibrate/locate the sphere first with the master probe in a parameter set (qualtool_moved=Yes DCC, then runs the rest of the probes using qualtool_moved=NO.

    Unless your ran the master probe before running the auto cal program.

    Maybe your probes and tip angles are referencing more than 1 master probe?
    Last edited by acgarcia; 07-06-2018, 03:34 PM.

    Comment


    • #8
      Originally posted by acgarcia View Post
      do you have a parameter set for your master probe? Do you calibrate/locate the sphere with your master probe first?

      my cal programs calibrate/locate the sphere first with the master probe in a parameter set (qualtool_moved=yes dcc, then runs the rest of the probes using qualtool_moved=no.

      unless your ran the master probe before running the auto cal program.

      Maybe your probes and tip angles are referencing more than 1 master probe?
      this ^--Master Probe is not in program so this is going to go back to last time someone answered "yes" to probe has moved prompt ; the first probe /angles calibrated is master, is everyone using the same probe. We use something short 20mm and 3 or 4 mm Ruby at A0B0
      Last edited by Rich P; 07-06-2018, 04:27 PM.

      Comment


      • HxSwartwood
        HxSwartwood commented
        Editing a comment
        This is exactly what we use. I accidentally grabbed a weird version of the program (old outdated one that we don't use). We do use a 4x20 probe as our master probe.

    • #9
      +1 on 'Master Probe' parameter missing. I saw similar things happen in my earlier days, a fellow app eng schooled me many moons ago and have not seen any problems since. Here's a snipet of an old autocal program:
      Code:
      $$ NO,
                  ====================
                  SLOT 1
                  ====================
                  LOADPROBE/LSPX1_5X20
                  TIP/T1A0B0, SHANKIJK=0, 0, 1, ANGLE=0
                  AUTOCALIBRATE/PROBE, PARAMETER_SET=MASTER, QUALTOOL_MOVED=YES_DCC,
                                SHOW_SUMMARY=NO, OVERWRITE_RESULTSFILE=NO
      $$ NO,
                  ====================
                  SLOT 2
                  ====================
                  LOADPROBE/LSPX1_3X50
                  TIP/T1A0B0, SHANKIJK=0, 0, 1, ANGLE=0
                  MOVE/POINT,NORMAL,<20.4985,15.4721,>
                  AUTOCALIBRATE/PROBE, PARAMETER_SET=3X50_TACTILE_SET updated, QUALTOOL_MOVED=NO,
                                SHOW_SUMMARY=NO, OVERWRITE_RESULTSFILE=NO
      $$ NO,
                  ====================
                  SLOT 3
                  ====================
                  LOADPROBE/LSPX1_1PNT5X40
                  TIP/T1A0B0, SHANKIJK=0, 0, 1, ANGLE=0
                  MOVE/POINT,NORMAL,<20.4985,15.4721,>
                  AUTOCALIBRATE/PROBE, PARAMETER_SET=1PNT5X40, QUALTOOL_MOVED=NO,
                                SHOW_SUMMARY=NO, OVERWRITE_RESULTSFILE=NO
      All probe tips are selected in order that will rotate B axis in a clockwise direction to avoid crashing into carriage beams.

      "Slot" represents garage port on probe rack.
      Last edited by Kp61dude!; 07-06-2018, 04:59 PM.
      PcDmis 2015.1 SP10 CAD++
      Global 7-10-7 DC800S

      Comment


      • #10
        Yep yep.
        Sounds like it's incorrectly programmed.
        When I created our AutoCal program, I had to make sure I studied up on what the **** I was doing for a few days.
        It took some time, but quite frankly I understand the calibration procedures much better now, to the point I was able to make a program to do it automatically.
        Make sure you understand exactly the process that is needed for calibration, then take a look at your program and make sure it is programmed to do exactly that.

        Comment


        • #11
          you can also add comments and if/end commands for when a probe tip exceeds the standard deviation

          I currently have the autocal repeat itself if the first run didn't pass my requirements. if it fails again, it goes to a operator commnets saying to seek help from myself or foreman to continue. Usually need the putty to clean the probe modules.

          Comment

          widgetinstance 190 (Related Topics) skipped due to lack of content & hide_module_if_empty option.
          Working...
          X