Xactmeasure and DOF

  • Filter
  • Time
  • Show
Clear All
new posts

  • Xactmeasure and DOF

    Hi all,
    First, I really appreciate all of your responses to my other posts. I'm trying to learn this software and you all are very helpful. I have a few questions that involve both of the topics I listed. Please see my primitive sketch below. If I had crayons I would have used them.

    Example 1 has a plane on top, a hole, and a slot, marked respectively A, B,C. The FCF callout is A|B|C. Pretty straightforward. My understanding, and correct me if I'm wrong, is that from a functional standpoint you'd level/origin to A (3 DOF), Origin to B (2 DOF) as if you put a pin in the hole, and rotate to C (1 DOF) for the 2nd pin and completely constrain the part. According to what I've read/been told about Xact measure it uses the secondary to rotate to. Even if B and C were created as 3d features or with different depths, their axis are already constrained by A, so they can't be used. So Xact creates a line through them and uses that to rotate to. The line could be angled because of different depths B and C were measured at, but it doesn't matter because the vector relative to A is the same regardless. Is that correct?

    Example 2 has a plane on top, but the hole and slot are on the side of the part, Datum D. The FCF callout is still A|B|C. Based on the previous logic Xact levels to A (3 DOF). After that is where I get confused because of different scenarios.

    If I measure B as a cylinder, does Xact use it's axis and rotate Z/origin Y to it, leaving 1 DOF translational in X? And if so, how is the last translational DOF locked down by C? By the origin/centroid of C? Is the depth it was measured at the location of X?

    Or, if I measure B as a circle and Xact can't rotate to it because it has no axis/vector and uses B and C to create a BF line, where is the line created? If I measured B and C at different depths would it create a line based on those depths and rotate to that?

    Hope this makes sense and sorry for so many questions. I thought this would be a good way to explain what I'm trying to figure out.

    Thanks again for all your help!

    PC-DMIS 2015.1, SP10, CAD++
    Global 7/10/7, 5/5/5
    Renishaw PH10MQ, PH10M, TP200

  • #2
    In my mind, and in both cases you will have to create points at the centre of the cylinder and slot that intersects the surface 'A'. Then you will have to create a line between the two points going from the point in the cylinder and the point in the slot to get the correct Y direction. Call the Line Datum B and then call the point Datum C.

    Im not sure if everyone will agree with me but if you just call the cylinder Datum B and put it into the FCF you are basically allowing the software to pick the direction for you as the cylinder doesn't control any rotational degrees of freedom, which in this case you need. Think of it like your two pins, yes they are two separate pins but they are acting together. If you manually put the part onto one pin with the other not there and then rotate the part so the second hole lines up for you to insert the second pin. You are basically manually aligning the part to an invisible line, you want to do the same in this case. Maybe the software can do that but in my experience you have to tell it what to do or it finds a way to do something unexpected.


    • #3
      Remember the mantra: "Each datum (in order) locks as many DOF (which aren't already locked) it can, no more, no less". And if there are no free DOF left when you get to the tertiary, or the tertiary can't help with the remaining DOF, there's an error in the print.

      Example 1:

      A - Z direction, Z zero
      B - (oriented in Z, no action), X and Y zero
      C - (oriented in Z, no action), coordinate system rotated (around B) until Cx is correct

      Example 2:

      A - Z direction, Z zero
      B - X direction, Y zero
      C - (oriented in X, no action), only DOF left is X zero

      Your example 2 is questionable, as the slot C is supposed to lock the X coordinate, which isn't a measured direction for the slot unless you use sample hits, and then it's not the slot that is the datum, but the plane. I would say "Print error".

      You can't construct a fixture/gage that only touches the surfaces of A, B and C of example 2 and still locks the part fully.
      SW support - Hexagon Metrology Nordic AB


      • #4
        What Anders said basically.

        Example 1 is fine, as you explained well.

        Example 2 does not have anyway of constraining the X translation (and as you stated, the rotation about Z would technically be done to the short axis of the Dat B cylinder.)

        Remember the mantra: "Each datum (in order) locks as many DOF (which aren't already locked) it can, no more, no less". And if there are no free DOF left when you get to the tertiary, or the tertiary can't help with the remaining DOF, there's an error in the print.
        ...also known as Can-May-Must.

        Reading a datum system in a control frame left to right

        If a feature CAN constrain a d.o.f., and it MAY (i.e. it's not already been constrained by a previous feature) then it MUST!

        Automettech - Automated Metrology Technology


        • #5
          Max, sometimes you have to just backup to basics. I don't look at the part in 3D I look at it in 2D. If A and B are perpendicular to each other in the first view a line and a slot. What would that account for Z and X. I flip it on its side, that slot becomes a circle, but I don't know that until I'm in that view. But when the Hole is parallel, then it doesn't help in either view. It can only control X. There is nothing to stop Y from moving up and down. C is simply rotation in both of your views but still nothing to control Y, or in your case X

          (In Memory of my Loving wife, "Ronda" who I lost March 7, 2016. I love you baby.)
          They say "Nobody's Perfect." I must be Nobody.


          • #6
            You can do a lot of testing in PCDMIS offline to figure out how the FCF interacts with the DOFs. Do a simple program just like your example. Create a plane, then a cylinder, then a round slot. Now create a point and declare the first 3 feature as ABC. If you dimension the point to ABC, you won't see the datum shift in your report because all 6 DOF are locked.

            Now dimension the point again but add mmc to B. Now you will see X and Y are unlocked, so this shows B is controlling X and Y translation.
            Now dimension the point again but remove mmc from B and add mmc to C. Z rotation is now unlocked so this shows C is only controlling 1 rotation, about Z.

            Now if you try to do your example 2, PCDMIS will throw up and give you a Tertiary axis is parallel to primary error. Not saying PCDMIS will always be correct, but this kinda throws a red flag and you should question the callout.

            You can use this test for other features to see how the demon controls DOFs.


            • DJAMS
              DJAMS commented
              Editing a comment
              Glad to hear that Xact threw an error on example 2. It is an invalid FCF.

          Related Topics