Standard deviations and probe builds

  • Filter
  • Time
  • Show
Clear All
new posts

  • Standard deviations and probe builds

    Hello all,

    I completed the PC-Dmis 201 class a few months ago where I learned about designating a 'master probe', using it to define my calibration sphere, then loading my probe I want to use and saying NO it hasn't moved.

    Currently, for my master, I'm using a 4x10 stylus on an extended force tp-20, and getting about .003-.004 standard deviation. I'm doing 51 hits on 6 levels. Can anyone confirm if that is normal? Am I ok with using an extended force for a short probe?

    Next question:

    I recently received a 90mm carbon probe extension, and have a 2x20 stylus on it on a medium force tp-20. For these tips, I do 25 hits on 4 levels. I know with such a long probe I was going to be sacrificing some accuracy, but when I calibrated it, I ended up with a SD of .008. Can anyone confirm for me if this is normal as well, or should I be trying something different? I tried slowing touch speed down from 5 down to 3 and recalibrated it and saw no change.

  • #2
    depends on how accurate you want to measure
    and thats a lot of points mi friend (and a pretty high STD =)

    try measuring a part with cmm and then measure it manually and compare results =)


    • pcdmisstudent
      pcdmisstudent commented
      Editing a comment
      I think there is a good thread on this subject if you search on the top right =)
      Do you reset ur master probe to theo at start of calibration?

      Prime numbers are a good rule of thumb, me I use 13 points. I think most people who are not using scanning probes use between 5-13 points at 3 levels

      Also you can experiment with different parameter settings, example if your tip is alot bigger than ur shaft you can try measuring from under the equator.
      ( maybe start at -5 degrees) and if ur tip is almost as big as the shaft you can take points over the equator to avoid shaft hits (maybe around 10-15degrees)
      Try slowing down ur moving speed:
      Use DCC+DCC instead of DCC only.

    • bigblue97mustang
      bigblue97mustang commented
      Editing a comment
      Ok, I'll give cutting the points back a try and continue trying different touch speeds. Do you think there is any problem with using an extended force tp-20 for such a short probe? And then using it as a master? I'm unsure if any of that makes any difference.

    • pcdmisstudent
      pcdmisstudent commented
      Editing a comment
      I dont have an extended force module, but I wouldnt think so. (not sure tho)
      Its still within specification for the stylus length from just giving it a quick look on renishaw's page and on the picture of the module theres even a 2x10mm tip on it!
      (cant link).

  • #3
    I use Tesa stuff (same as TP20).

    My hardware:

    My MASTER Probe is:
    TIP 4X20 (standard ruby ball & stainless steel)

    While calibrating, I have my DISPLAYPRECISION jacked up to "7".

    For my MASTER, my STDDEV will be between 0.00005"-0.00009"
    Qualifying at 5mm/sec

    All regular probes will have STDDEV between 0.00005"-0.00020"
    Qualifying at various speeds. Speeds determined through previous testing. All between 1.5mm/sec-5mm/sec.

    Longer funkier star probes, STDDEV between 0.00010"-0.00029"
    Qualifying at 1.5mm/sec

    For my MASTER, I use 17 hits.

    For all other probes, I use 11 hits.

    There is nothing wrong with using an extended or medium force for your master probe probe. I did some testing and found that to be the best case scenario for me and my hardware.

    90mm exceeds the spec range for Tp20. If you're going to do something that breaks the rules, do a lot of testing. Be hyper aware of your tolerance range and how accurate this probe is before using it in each tip orientation you plan on using it in.
    Last edited by DAN_M; 12-11-2020, 07:49 AM.
    SF7107(PCD), SF454(PCD), 152614(Quindos), 9159(Quindos), 7107(Quindos), B&S Manual, M&M Gear Checker


    • bigblue97mustang
      bigblue97mustang commented
      Editing a comment
      Thank you so much for your input. Can you think of a reason i would be seeing so much deviation on my master probe then? I'm almost to a stopping point at work where I can do some testing on various things.

      For your hit density on your probes, how many levels are you doing for your master and tethered probes? I'm guessing 3 for each.

      Lastly, I was unaware the TP's had a length tolerance. I had always thought it was about weight, which is why carbon fiber probe extensions exist.

  • #4

    Happy to help!

    Master Probe Issue-->
    Good calibration is all about temp control, cleanliness of hardware, and control of the CMM's speed.

    Assuming your machine is in an adequate environment and you know enough to clean, I'd really hone in on speeds.

    PC DMIS has an option to have the software view your speed commands as an ABSOLUTE VALUE (software sees MOVESPEED/80, it will read as 80mm/sec).. OR... as a PERCENTAGE VALUE (software sees MOVESPEED/80, it will read as 80% of your machine's top speed, whatever that is).

    Open your program. Hit 'F5" for "Setup Options Menu". Click to "PART/MACHINE Tab". Look at "Display Absolute Speeds". If that box IS CHECKED (I keep mine checked) then PCD will look at speed commands as the absolute value.

    A lot of people miss this in lower level training, overlook it, and then chase their tails because of this, while doing calibration testing. They think they're touching at 5mm/sec because they have TOUCHSPEED/5 in their prg, they've overlooked the checkbox, and they're actually touching at 5% of machine's capability (5% of my machines max speed of 300mm/sec would be 15mm/sec, not what they wanted or thought!) Don't want you to fall victim to that.

    Hit Density-->
    I use the software defaults. I have never told the software to use more of less levels nor have I ever done any particular specifying to the software aside from controlling my TOUCHSPEED & hit value (17 for MASTER, 11 for all SLAVES).

    Length Tolerance-->
    Hexagon & Renishaw's literature state that the max recommended length for TP20 probing is 60mm.
    Yeah, they make carbon fiber and yeah, you can build probes longer than 60mm and the software won't yell at you for it.
    However, once you have your speeds nailed down and you feel confident about the testing numbers you're collecting, you'll start to see that the longer you go and the more you swing those long probes, the less accurate your STDDDEV will be.
    If you want to go long, its up to you to determine...A) What is the best STDDEV I can squeak out of those probes after testing? B) Once best STDDEV is found & those conditions are being used, are these probes good enough to repeat-ably measure my customer's product?
    Last edited by DAN_M; 12-11-2020, 10:22 AM.
    SF7107(PCD), SF454(PCD), 152614(Quindos), 9159(Quindos), 7107(Quindos), B&S Manual, M&M Gear Checker


    • DAN_M
      DAN_M commented
      Editing a comment
      Good stuff!

    • bigblue97mustang
      bigblue97mustang commented
      Editing a comment
      Thanks for all your help man! I really hope some other younger programmers stumble upon this and get to learn from this. You did a really great job spelling everything out.

    • DAN_M
      DAN_M commented
      Editing a comment
      Thanks! I'm only 29 lol. ThePudds is the real youngster around these parts!
      All I ask in return is that when presented with your own opportunity to help, please do it. You never know when your comment will make something "click" for someone.
      Have a great day!
      Last edited by DAN_M; 12-11-2020, 12:59 PM.

Related Topics