Large Sphere, Small Part

  • Filter
  • Time
  • Show
Clear All
new posts

  • Large Sphere, Small Part

    So I have a family of parts(pucks) with a convex(out of) or concave(into) top surface.
    The CMM seems to have trouble accurately measuring this sphere,

    The Outer Diameter (OD) of the part 3in and 6in, with the a convex sphere with a 7.5in radius (poorly drawn but cannot post BP)

    Polar radius with an alignment doesn't work well because if the OAL is bad then the Sphere will automatically be bad.

    Tried taking hits along the x and y axis to construct circles instead of a sphere but still only sort of works

    This is also causing issues with the OD-Sphere intersection height since the sphere is not accurate enough

    Gage points seem to be effective, but not all BPs have them

    seems like I have a lot of partial, quick answers but nothing really accurate


  • #2
    Does the sphere directly relate to the rest of the part?

    If you can offset in Z from the bottom, and then offset in XY from the OD (maybe 0) to "set" the sphere center, then you know what the radial values of any hit on the sphere surface should be.


    • Cheetos_The_CMM
      Cheetos_The_CMM commented
      Editing a comment
      So I "can" offset the Z axis to the theoretical sphere centerline but if the sphere was cut higher or lower than it was supposed to be the sphere construction is bad and the distance from "theoretical sphere center" would give bad readings.

      If the part was 100% perfect height that would work, but manufacturing is never perfect

  • #3
    what is bad readings? If the part is the wrong height, wouldn't the sphere be in the wrong location? and wouldn't you want to know this? Otherwise, why are you inspecting it?

    Don't take my writing personally, I am just trying to understand the real question


    • Cheetos_The_CMM
      Cheetos_The_CMM commented
      Editing a comment
      Sorry im relatively new to the pcdmis/cmm world.
      BUT, a bad reading would be if it was either too close to the center or to far
      The sphere doesn't construct very well, and the (x,y,and,z) readouts of a sphere constructed from a series of points (not auto sphere) are not accurate enough

      When I make a cylinder on the OD I construct lines to intersect the sphere so I can get a circle constructed, then take the z location of the circle and that gives me the information needed to tell the machinist to offset

      I need to know if the sphere has:
      A) a bad shape (if its an egg shape)
      B) if the sphere has the proper (Z)
      Last edited by Cheetos_The_CMM; 11-18-2016, 04:25 PM.

    • Cheetos_The_CMM
      Cheetos_The_CMM commented
      Editing a comment
      Im considering writing a loop function that will offset and will add or subtract the Z plane alignment and continuously do the math until the Sphere radius is good, but will time out after (lets say 20 tries) and give a reading for a bad part
      Basically do:
      lets say theoretical sphere center is (-6.5) from the zaxis (cmm plate)

      assign n=0
      start Loop 20times
      N=N+1 (make a counter)
      if N>11 goto Next

      If N=1,N=2,N=3,N=4,N=5,N=6,N=7,N=8,N=9,N=10,
      Assign V1=(N*.001)
      end if
      If N<11 goto END


      If N=11,N=12,N=13,N=14,N=15,N=16,N=17,N=18,N=19,N=20,
      Assign V2=-((N-10)*(.001))

      A1=Alignment Zaxis offset= -6.5+V2+v1
      LINE1: Distance from center to point 1
      LINE2: Distance from center to point 2
      LINE3: Distance from center to point 3
      LINE4: Distance from center to point 4
      Assign PNT=(LINE1+LINE2+LINE3+LINE4)/4

      if PNT,PR= undesired
      goto END
      if Point.PR= desired
      Goto Dimension

      Recall A1
      CONSTRUCT SPHERE (generic with variable PNT)
      (sphere dimesion stuff)


      Loop End
      Last edited by Cheetos_The_CMM; 11-18-2016, 04:23 PM.

  • #4
    After aligning to the part, take a vector point at XY (0,0) and vector (0,0,1). Use this point to offset your Z.

    Offset your sphere measurement by this value (using Auto Relative Measure).

    So, if your sphere is THEO 0,0,-6.5, then use RMEAS to multiple features, selecting the vector point in the Z drop down.

    Just taking points is never going to be very good, because the surface is such a very small portion of a sphere.

    With points, your best bet is always to origin on the theoretical sphere and use polar radial points, just like doing a small arc radius, but in 3D.
    "This is my word... and as such is beyond contestation."


    • Cheetos_The_CMM
      Cheetos_The_CMM commented
      Editing a comment
      I will definitely try/do this

      Thank you!!


Related Topics