What is the optimal touch speed on a TP20 for ck'ing close limit toler of .0005"

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

  • What is the optimal touch speed on a TP20 for ck'ing close limit toler of .0005"

    Checking close limit dia's & GD&T of .0001-.001", what is the optimal touch speed for tp 20's? Regardless of length of probes... Currently programs are set at 2(?) I believe it should be .15-.25 TOPS. Please help

  • #2
    mbaero,

    I use tolerances close to 5µm everyday and have used the TP20 for a many applications. Touch Speed 1-2 is repeatable at least on my measured to 1µm if not under 2µm for sure. My machine is environmentally controlled however. I have seen in the past if your Touch Speed is too low then the machine may think it is an error but honestly the lower the Touch Speed I would assume the better the accuracy.
    Anthony Fotion
    Quality/Reliability Engineer Technician
    Fujifilm Dimatix Inc.
    E-mail: afotion@fujifilm.com

    Comment


    • #3
      Hi mbaero,

      There is a lot to consider here.

      TP20 for ck'ing close limit toler of .0005"
      TP20's and other kinematic probes produce a lobing error. This is caused by the varying forces needed to trigger the switch depending on which direction the probe tip is deflected by the part. You can prove this by checking a ring gage with a lot of points and reporting the form. Create a graphical analysis of and you will see it is not round. It will have a shape similar to a rotor in a Mazda rotary engine. Google it. There are articles and diagrams that will do a much better job explaining it.

      what is the optimal touch speed for tp 20's?
      Doesn't exist.

      Regardless of length of probes...
      This has to be taken into account. The shorter the better as a general rule. But the effective stiffness matters as well. Carbon fiber and carbide stems are stiffer than steel for example. Not to mention the different module forces.

      Currently programs are set at 2(?) I believe it should be .15-.25 TOPS.
      In my testing that is too slow. Renishaw tests the TP20 modules for measured uncertainty of repeatability at 8 mm/s. If the measuring speed is to slow when the mechanical switch is triggered / contact broken it creates a noisy signal. This can cause greater variation and less repeatability. For reference most of my tips run between 5 mm/s and 11 mm/s. Again lots of variables here.

      I have been testing all of these phenomena for the last 3 months. I recently started a new job where all the CMM's use TP20's I had to educate myself to their potential pitfalls. All my prior CMM experience was with TP200's and Leitz LSP series analog probes. So what I have done is created a program to perform the ISO 10360-4 scanning test, which involves checking the calibration sphere with 25 points at prescribed levels with a certain number of hits. I repeat this 5 times for a total of 125 hits. I report the total deviation of all the hits. I've used this program as a guide to determine what is the best measuring for any giving setup and log all results in a spreadsheet. First I did pretty exhaustive testing on what number of levels and hits to use for tip calibration, seems like a logical starting point. I read through the forum and wrote down any combination of hits/levels I saw. I asked to the 2 Hexagon techs that came out to calibrate and service the CMM's what they used. I tested them all, 26 combos in all. If you're curious what the winner was in my testing and what I use on all machines now.... 7 hits 2 levels. Let the flaming begin!

      I think you should do some testing, even if minimal, to see what your setup is capable of. You're probably be asking too much from your TP20.

      Mark




      Comment


      • VinniUSMC
        VinniUSMC commented
        Editing a comment
        7 hits, 2 levels. That is intriguing. I had also never heard of the 8mm/s testing speed. Where did you find this info?

        I'm normally running 3mm/s. This could greatly speed up a lot of what I do if this is solid intel.

    • #4
      Vinni, I initially found it on the "Probe Test Certificate" that come with each module, but I have also seen it in other Renishaw literature. If I find a link, I'll post it.

      Comment


      • markchurch
        markchurch commented
        Editing a comment
        I test each combo of module and stylus I use, or try to. Here are a few examples of ones I have tested the most:
        1.5X30MM carbide stem on a SF module -- 5 mm/s
        2X30MM carbide stem on a SF module -- 7 mm/s
        2X40MM carbide stem on a SF module -- 11 mm/s
        6X50MM carbon fiber stem on a SF module -- 8 mm/s
        4X10MM carbide stem on a SF module -- 8 mm/s ( with this combo the difference in total error between 8 mm/s and 15 mm/s was only .000019")

        .

      • VinniUSMC
        VinniUSMC commented
        Editing a comment
        Ok, so you're doing it all based solely on your sphere measuring analysis, not dependent on any other externals (part tolerances, material types, etc)?

        Very interesting info. I'm surprised that the 2x30 does better at a slower speed than the longer 2x40.

      • markchurch
        markchurch commented
        Editing a comment
        I would adjust the speed higher or lower if needed. Like I said earlier, there is a range of speed that the error will be lower, but in most cases it fairly close. For example the 2X30MM tip after running the test 10 times, the mean average error was .000131" for the 7 mm/s setting with a maximum deviation of only .000009" between all runs. So when I need to check tighter tolerances and setup and material allows me to use this speed I do. Now at 17 mm/s the mean average error was .000198" with a maximum deviation between all runs of .000086", still not bad. So if I had a part that had more tolerance and had or lot of features or the sample size was large for final, I would use the higher speed. I think the biggest lesson I learned so far is we go measure a lot faster than we've been told. If you're curious, I did check it a 2 mm/s but only ran the test 3 times. The mean average was .000226" and maximum deviation between the three runs was .000121". Like I stated in my earlier post the slower the switch is opened the 'noisier' the signal is to the controller. I suspect this is why 2 mm/s performs worse.

    • #5
      Make sure you're in absolute values for speed (F5, Use absolutes). Otherwise you're touch speeds are also in percentages. But also as a note, what Renishaw tests their probes at is not necessarily what is best for the CMM hardware. We spec our machines if I remember correctly at 2-6 MM/S for repeatability, but markchurch is definitely doing the best thing by actually testing what works best on his hardware.
      Systems Integrator
      Hexagon Manufacturing Intelligence

      Comment


      • #6
        Another thing to keep in mind with all this is the actual hardware. I have an old Validator, heavy & slow. And no ability to stop "right now" like the new Global we just got. Anything more than 2.5 absolute touch speed is looking to break tips when measuring normal to the probe.
        sigpic
        Originally posted by AndersI
        I've got one from September 2006 (bug ticket) which has finally been fixed in 2013.

        Comment


        • #7
          I didn't see anyone mention it, but one of the main things to consider with the TTP is the touchspeed to prehit ratio, which has made itself evident when I am measuring a hole that requires much less prehit/retract. Such as .01"-.02" prehit, I would need to be somewhere around 1 mm/sec. to get an accurate diameter measurement.

          Comment


          • #8
            Originally posted by David C View Post
            I didn't see anyone mention it, but one of the main things to consider with the TTP is the touchspeed to prehit ratio, which has made itself evident when I am measuring a hole that requires much less prehit/retract. Such as .01"-.02" prehit, I would need to be somewhere around 1 mm/sec. to get an accurate diameter measurement.
            David, all the testing described in earlier posts were performed with prehit and retract set to .100" However, I did do some testing with diminishing settings to find out at what point I needed to reduce speed. A 4X10MM had an optimum setting of 8mm/s with .100" prehit could be run at .040" prehit with the same results. 5mm/s at .020" gave almost equal results. I'm sure this will vary greatly with different machine/controller combos as each one will need a different distance to stabilize it's path to the target.

            Comment


            • #9
              I've worked at Renishaw in the CMM division and all the above is correct but all is so variable. Firstly their is no ideal probe speed for a specific probe / styli setup.
              Also each CMM manufacturer develops their own probe interface within their controller and not always do they adhere to the Renishaw recommendations for electrical switching thresholds.
              Basically the other effect on probe / measurement / repeatability performance is the size of the CMM, meaning the weight of the moving parts which they cause dynamic influences in the measurements.
              Markchurch is doing the correct things in testing and testing his varying probe setups on his own machine.
              The lobbing effect caused by the TP20 is due to the design being of 3 kinematic seats where you have a low and high force depending on the angle of probing in relation to the kinematics where the switching is electro-mechanical. The only way to eliminate this if you are chasing the micron is to use a non-kinematic switching probe, such as TP200, TP7 or SP25 etc.

              Remember the best repeatability you will get also depends on the controller performance as well as the accuracy of the scales fitted to the CMM.
              The CMM manufacturers accuracy statement says it all. where it may say 2.5+L/100
              Where the 2.5 is the repeatability which considers the system repeatability including the probe
              The L/100 is the accuracy over the length being measured.

              Generally if you want to measure in microns then you need to purchase or upgrade your CMM to suit.

              I know some of the above answers are vague but this is a nature of chasing the micron on a CMM, its down to testing and more testing on your own installation, as each CMM even if the same make and model will perform slightly different but should be within the manufacturers spec. as long as it is within a stable environment so measurements taken over a period of time can be compared.

              Comment

              Working...
              X