hexagon logo
Register Sign in
  • Replies 3 replies
  • Subscribers 31 subscribers
  • Views 47 views
  • Users 0 members are here
Options
Related

Hard Probing

freddyk1369
Hey,

I am having a problem understanding how the hard probing works and how it automatically snaps to a certain vector. We have attempted to take hard probed points from different directions on the exact spot (or as close as possible to) and it seems to come up with different vector numbers and about a .06" difference in location in the desired direction of measurement. I have yet to try to work on an angled piece I am going one step at a time because sometimes there are just spots and holes that you can't get the scanner to but I want to make sure that I am getting the exact measurements and not something that is .06" off because I didn't hit it right at the point. What sucks is it seems like it is snapping to a certain vector depending on how I have the probe head pointed and we may be good with aligning the head to the surface but there is no way that point is being taken at exactly 90 degrees from that surface. And then when I am doing a hole it gives me the same vectors as if I were grabbing it from an edge. We are doing this without CAD at the moment just to try and get a better understanding. If anyone has any suggestions on this it would be greatly appreciated.

-Fred
Parents
  • Hey,

    What sucks is it seems like it is snapping to a certain vector depending on how I have the probe head pointed

    -Fred

    That's exactly what's happening with Measured Points.

    When you do (for example) a measured plane, PCDMIS records the XYZ of the probe center, and the IJK of the probe angle. The summation of the IJK info is enough for it to determine what direction the plane actually faces. From there, it recompensates the hits according to the probe size, and generates a Measured Plane. This is what happens 'behind the scenes' whenever we click the Done Button.

    Contrasted to a Measured Point -- there isn't enough IJK information to comp the point in the intended direction. Therefore, the software designers had Measured Points comp in only the six cardinal directions (X+/X-/Y+/Y-/Z+/Z-). Same thing happens with bridge CMMs in PCDMIS.

    In training, I always tell people "Do not use points!" for the above reason. Points are only (semi)worthwhile when you have an active alignment, and the points are on a face that is square to one of the cardinal directions, otherwise you risk what is referred to as cosine error. Not to mention, the form of a surface is usually not perfect...why would we not take more than 1 hit and generate a plane when possible, averaging out the form error? Much more repeatable.

    If you're dead set on using a point for your application (which you haven't posted), and the point is NOT on a surface that is square to the XYZ coordinate system, use an AutoVector Point, and click the CAD model to capture the point's XYZIJK. That will work. However, if you lack a CAD model, you might be in trouble, as finding the IJK of the referenced surface can require a little math work.

    That IJK is what PCDMIS needs to compensate, so don't think you can just ignore it! Good luck.
Reply
  • Hey,

    What sucks is it seems like it is snapping to a certain vector depending on how I have the probe head pointed

    -Fred

    That's exactly what's happening with Measured Points.

    When you do (for example) a measured plane, PCDMIS records the XYZ of the probe center, and the IJK of the probe angle. The summation of the IJK info is enough for it to determine what direction the plane actually faces. From there, it recompensates the hits according to the probe size, and generates a Measured Plane. This is what happens 'behind the scenes' whenever we click the Done Button.

    Contrasted to a Measured Point -- there isn't enough IJK information to comp the point in the intended direction. Therefore, the software designers had Measured Points comp in only the six cardinal directions (X+/X-/Y+/Y-/Z+/Z-). Same thing happens with bridge CMMs in PCDMIS.

    In training, I always tell people "Do not use points!" for the above reason. Points are only (semi)worthwhile when you have an active alignment, and the points are on a face that is square to one of the cardinal directions, otherwise you risk what is referred to as cosine error. Not to mention, the form of a surface is usually not perfect...why would we not take more than 1 hit and generate a plane when possible, averaging out the form error? Much more repeatable.

    If you're dead set on using a point for your application (which you haven't posted), and the point is NOT on a surface that is square to the XYZ coordinate system, use an AutoVector Point, and click the CAD model to capture the point's XYZIJK. That will work. However, if you lack a CAD model, you might be in trouble, as finding the IJK of the referenced surface can require a little math work.

    That IJK is what PCDMIS needs to compensate, so don't think you can just ignore it! Good luck.
Children
No Data
© 2024 Hexagon AB and/or its subsidiaries. Privacy Policy | Cloud Services Agreement