Runout and Total Runout on a CMM

  • Filter
  • Time
  • Show
Clear All
new posts

  • Runout and Total Runout on a CMM

    Hello all,

    The runout represents for tridimensional metrology a real challenge. The reason is due to the fact that involves object revolution and in 3D measurement there is no such rotational movement.

    I confess that I am not very fluent with this evaluation because it was never an item that I had to use a lot throughout my life.
    It is usually said that runout and total runout are obtained by applying the following formulas:


    This represents the worst possible case since it can only happen if we join the largest deviation of circularity and concentricity in the same section as illustrated below:


    Taking the practical example of a PC-DMIS report we have:


    In this case, the value "RN" refers to the circularity deviation and below is the concentricity deviation resulting from the combination of the XY deviations.
    Applying the formula described above we have:
    0.042 + 0.064 = 0.106

    Like this we assume that the illustrated scenario will for sure occur.
    It is for sure a safeguard and we can say that with this evaluation in the extreme, we only reject good parts and never the opposite”.

    It is true that:


    Where C and T depend on how shape (circularity) and position (concentricity) relates to each other.
    However the modern CMM applications come equipped with modules of direct Runout and Total Runout evaluation. In theory the software can estimate the values of C and T simulating manual measurement.
    In our case, if we do so, this will get the following result:

    This is what makes Runout so difficult to evaluate using a CMM, and it is for me always preferable to evaluate the two components individually.

    My question to you is if you already correlate the results from the direct evaluation to a manual measuring on a dedicated test machine and how accurate the CMM result can be.
    Please give me your thoughts regarding this issue.

    Thank you.

  • #2
    I think runnout = circularity + concentricity is only an approximation.
    In this case, in my opinion, the concentricity should be the radius defect only (not multiplied by 2) + the Chebychev circularity (here, you use the defect RN, which is associated to the used algorithm- maybe least squares).
    Can you try to see if half of the concentricity (sqrt(x^2+y^2=0.32) + circularity = 0.062 (honnestly, I doubt that the difference between LS and Chebychev can be 0.012 !!!!!)


    • Alexandre Santana
      Alexandre Santana commented
      Editing a comment
      Hi Jefman, why use half of concentricity?
      When turning a part 360º you will get the total amount of concentricity.
      Last edited by Alexandre Santana; 11-26-2017, 01:08 PM.

    • JEFMAN
      JEFMAN commented
      Editing a comment
      Alexandre Santana : You're right....
      I think about it because the tol of runnout is a difference of radius, and the tol of concentricity is a diameter...
      I'm not sure that runnout is totally clear in my head : I never measure it !

  • #3
    In addition, another explanation than mine, in a better english !


    • #4
      Runout(cirular element) = FIM (point to point deviations on CMM) with diameter locked in as 0. It is the surfaces departure from the desired location.
      If you measure 360 points on a datum dia and 360 points on reporting feature, dpt1 - fpt1 = # ....dpt100-fpt100 = #, take the highest dev - lowest dev; value should equal what you get with an indicator at the turning center.
      Total-Runout(applies to a surface area) = similar premis, but must move along the feature axis without re-zeroing indicator along entirety of surface. Can still be accomplished with example above however instead of dpt1 - fpt1 it becomes dpt1-f1pt1...dpt1-f4pt1, find the TOTAL range used in all 400 evaluations.

      Last edited by R2ah1ze1l; 11-27-2017, 03:31 PM.


      • #5
        That' clearer like this...
        What about "runnout = circularity + concentricity" on a cmm in this case ?


        Related Topics


        • JoBLOCK
          by JoBLOCK
          Roughly speaking: runoutx2 = concentricity right?

          Example. round part. I am chucking on the OD checking the runout of the ID
          03-30-2009, 02:55 PM
        • John Kingston
          Total Runout
          by John Kingston
          I have a part with /TR/.010/A/B/. -A- is the plane I am leveled to in Z and my Z origin. -B- is a diameter perp. to Z. -B- is also my X,Y origin. The...
          05-29-2007, 03:12 PM
        • JoBLOCK
          Another runout thread
          by JoBLOCK
          Ok, I have a rod with 2 dia.

          The callout is on the 2nd dia. total runout .001 -A-

          When I place the rod in the v-block on -A-,...
          02-23-2010, 10:48 AM
        • AuRules
          by AuRules
          With help from you guys on my earlier thread, we went with runout.

          So I used this example

          The worst circular runout...
          04-23-2008, 04:31 PM
        • c_bearden
          Another Runout Question...
          by c_bearden
          Is it just me or does PCDMIS take the total indicated runout and double it???.....I am checking runout (spec = .0005") and my CMM reports it as .002"......I...
          11-27-2010, 12:42 PM