hexagon logo
Register Sign in
  • Replies 10 replies
  • Subscribers 223 subscribers
  • Views 101 views
  • Users 0 members are here
Options
Related

PC DMIS Cone position

Entey
Can someone explain how PCDMIS gets CONE position values.



I have CONE with 3 levels, all 3 levels (circle0-->2), have different positions values to actual CONE. This is understandable, since its probably looking at AXIS of cone.



Where in picture above would position of this particular cone be? Also choosing axis average, worst, etc options does not change position values at all.





We have instances where position of cone varies by quite alot between measured raw part and finished part. The only difference between raw and finished part is that finished part is grounded to have smooth finish and this also reduces part thickness (so the cone is also cca 20%ish smaller).

Cone position is constrained by datums:
A (bottom plane which is grounded down to have smooth finished, raw part has quite uneven surface)
B (outside diameter, does not change at all)
C (one of the cones on part)




I also measured the tool itself. There are coining punches that make that cone and when I measure position of holes or center of coining punches, all positions look OK within microns. Problem comes when I directly measure position of CONE of those punches. Deviation comes to 0.3mm+.



Thanks in advance for any information you have.



  • We have a part very very similar to this. Your picture is truncated, but the through holes below the cone have a concentricity callout to the cones, correct?

    How are you probing the cone? Scanning? Touch Trigger Points? Number of points if TTP?

    We have pre-controls in place in which I have a variable that probe goes down to in order to compensate for the grinding process on both sides.

    What I did on the part was actually find the thru holes first. Then when I probe the cones, I zero on the thru holes. I found out that the amount of error on the cones reduced. Then when I spit out the measurements, I dimensioned them the way the print shows. I worked backwards. Maybe try constructing a 3D line from the circles of the cone and see if that helps. I also added a 4th level to the cone in which I saw better repeatability.

    You also have to remember that the cone's axis to taking perpendicularity into account as well. Maybe the progression of the die is not putting the cones in perpendicular. Run a perpendicularity check to see how much that is eating up the tolerance vs the actual position.
  • I would:

    -Be using touch trigger probing strategy
    -Origin on the ID and plane above the cone to help you find it. THEN probe the cone.
    -Use more then 3 levels. If you're reporting form or position of a cone you'll need more input to the software to make the calculation repeat-able. Would suggest 5 levels.
  • in reply to DAN_M
    I think 31 levels would be a good justification, honestly. Wink
  • Rolling eyesRolling eyesRolling eyes should also construct a POINT at the center of it....not sure what we should call it though.........Rolling eyesRolling eyesRolling eyes
  • in reply to DAN_M
    Definitely not PNT31, but on the other hand, that has a good ring to it. Like it REALLY encompasses a solid alignment.
  • 1: Your true position callout on the print is pointing to a centerline axis.
    --This is not okay per any rev of GD&T. it needs to point to an actual feature of size.
    --This brings up the point that tolerance of position could apply to the bore and not the countersink.
    2: IF your customer/engineer insists "1 above" applies to both the countersink and the bore at one time, which is absolute bullhonkey,...
    --Countersink is only controlled axially to that bore. not controlled by depth. Depth is controlled by a simulated ball with a height of 10.3mm ±0.25mm, so position to three datums (presumably controlling all 6dof) is not necessary or appropriate.

    to directly answer your question i believe a cone's origin from which the dimensional result is derived, is at the yellow theoretical "tip" of the cone, which in this case is projecting your error along an unstable axis line, amplifying your actual variation.

    get clarification on wtf the engineer/customer truly expects with that tolerance of position to 0.3mm to ABC and becoming datum D.
  • I always use touching strategy, as i dont like to calculate correct heights in SCAN strategy.

    I never tried more than 3 levels, tbh.

    Results on assembled tool. Cone position is alot more out than simple circle made on the cone. This isnt the case if I measure just coining punch itself, all 3 are within 0.01mm. The only difference is how Z origin/level are placed. I know and tested it that level does play role in cone position, I would just never assume its this much, especially on tool.




    Thanks for all the information.
© 2024 Hexagon AB and/or its subsidiaries. Privacy Policy | Cloud Services Agreement