hexagon logo
Register Sign in
  • Replies 4 replies
  • Subscribers 221 subscribers
  • Views 108 views
  • Users 0 members are here
Options
Related

Error: Compound or pattern datum must be perpendicular to datum plane.

ninjachachi
I recently was tasked with measuring a part bent at an angle with a pattern of holes on both surfaces of the bend. The bottom plane is Datum A, the bottom pattern of holes is Datum B. There is a profile locating all of the edges on the bottom plane to Datums A|B. The top bent plane is Datum C and this angled bend is controlled with flatness and Profile to Datums A|B. Finally, on the Datum C plane, there is a set of two holes, Datum D. These 2 holes have a position callout to Datums C|B as well as just to Datum C. Everything works fine in PC-DMIS until I try and find the position of these top 2 holes. Finding their position to Datum C works just fine (as it's controlling their relation to each other and perpendicularity to the plane they're on), but when trying to calculate their position to Datums C|B I get the following error: "Compound or pattern datum must be perpendicular to datum plane."

Attached is a rough sketch of the print (everything, but hole positions and Basic dimensions were removed). I certainly get the design intention, and the GD&T kind of makes sense as well. The design intends for the top holes on Datum C to be within XXX position of the Datum B hole pattern (located on Datum A)

Am I doing something wrong? Is the GD&T wrong? If it is wrong, why is it wrong and how can it be corrected? I talked with the design engineer behind this part and they think that they GD&T is correct, so I'd definitely need a strong argument to make any sort of print change.

Sorry for the rambling, I hope it all makes sense!
  • I think I get the design intent, but I also see why the software has an issue with it.

    If Dat C is controlled by profile to A|B then that ensures the bent plane is in the right location.

    If the part assembles on A and B, then the mating part (mounted to the holes in Dat C) needs to be controlled to that location then calling it back to A|B would make sense.

    I'd argue perhaps that the holes in datum C should also be controlled to A|B - perhaps with a separate perp call out to C.

    Alternatively would those holes called to C|A|B also work.
  • If you update to a newer version of PC-DMIS (2020 R2 or more recent), the geometric tolerance command should be able to handle this specification.

    XactMeasure required a secondary datum pattern to be perpendicular to the primary datum feature, a condition that is not required by Y14.5.

    Also, if you have the option to request a drawing change, you should ask what the allowable geometric error on datum pattern B is with respect to datum C in order to properly compute the MMB of datum B.
  • Yet another brilliant engineering exercise that in the real world makes little sense. I got 2016 dmis cad. No fancy stuff. If I see something like this I request assembly drawing & based on that I figure things out. Often I just toss engineering datum schemes out of proverbial window after seeing assembly drawing or Pro/E assembly model & make up my own based on that. The worst problem is engineers forget you need to have 6 degrees of freedom to make sure CMM knows where part is & how it's oriented. Heck you need those in the CNC machining as well. But since their computer models look nice it's all good to them. The exceptions are cylinders like ring gauges or shafts. Anyway, 1st the 2nd hole next to D should have had CBD callout since you can see a basic dim going from D. 2nd, project circles from constructed and/or auto cylinders onto plane C. From those 2 holes have the software check their locations with respect to B. Plug the differences into the T.P equation using a calculator. Screw the datum shift. If the hole locations pass then it's all good regardless. Then report perpendicularity of those cylinders to C plane to take care of the refinement of that composite position. If the bottom part of that composite doesn't have any other datums I don't see any relations between each other.
  • Thank you all for your input! I think NinjaBadger put it best when he said "I get the design intent, but I also see why the software has an issue with it."

    I don't think this GD&T callout is invalid, but it's certainly pretty odd and difficult for my older version of PC DMIS to calculate. I will pass along your comments to those who need to hear them and hopefully we can get something figured out.
© 2024 Hexagon AB and/or its subsidiaries. Privacy Policy | Cloud Services Agreement