hexagon logo
Register Sign in
  • Replies 28 replies
  • Subscribers 223 subscribers
  • Views 377 views
  • Users 0 members are here
Options
Related

Manual vs DCC alignment

acooke
I have taken the basic 1 week PCDMIS course from Hexagon.

Currently I have only modified part programs written by the programmer at our parent company location to run on our cmms.

I work with version 3.7 on cmm 2 and 2014 on cmm 3.

Cmm 1 has Measuremax which I am not sure how to program. This question is regarding cmm 2.

I think I may be a little confused about the purpose of running the dcc alignment after the manual alignment has run. I was under the impression that the manual alignment was to locate the part on the cmm, and the dcc alignment was the "fine tuning" and would adjust for very slight placement changes.

We have a fixture that is bolted to the fixture plate. The part is then mounted on the fixture and screwed into place using 1 screw. The part is secure, but due to different cavities with different die conditions may sit slightly different in the fixture each time. We are measuring the true position of a bore on the x minus side of the part with respect to the plane and bore at the x/y/z 0 position that is used to set the part alignment.

What we have noticed is that when we run the first part with the manual alignment the true position is within tolerance (.04mm). When we run a second part, if I mark all but choose not to mark the manual alignment portion, we see true position values between 0.1mm and 0.2mm.

If I then run the program on that same part again using the manual alignment without moving the part the true position values are again within tolerance. From what I read online yesterday after lots of digging it looks like the dcc alignment serves only to ensure the alignment features are properly probed and defined.

Is there any way around running the manual alignment each time? I don't mind if it's necessary for different cavities or runs, but for a 300pc cpk it's quite time consuming. Thanks in advance for any answers and/or suggestions. Oh, and I'm well aware of the noob hazing here. I have thick skin Slight smile.

Amber
Parents
  • Hello and welcome to the forum, Amber!


    The part is secure, but due to different cavities with different die conditions may sit slightly different in the fixture each time.

    This is the red flag clue.

    What's happening is that when you run the Manual "the part's sitting in this part of the ballpark" alignment on Part #1, the machine knows exactly where #1's datums are and can go into DCC mode and beep them with little to no error, getting a good alignment on them and making a good report.

    And then when you grab part #2 (which is a little different, being from a different cavity made on a different shift etc) and skip the Manual alignment, the machine goes into DCC mode and aims for where it thinks the datums should be - where they were on Part #1. Now they're not too far off, not enough to make it crash but far enough that when it beeps it's not hitting with a DCC approach vector perfectly perpendicular to the datum surface. So each datum point is now being probed with some error, we call it cosine error, getting a shoddy alignment and poor results.

    And then as you figured out simply run the Manual alignment on part #2 and it's fine.

    The solution is to layer a third alignment after 1st two:
    1) Manual alignment - can be skipped as long as the fixture isn't moved.
    2) DCC Rough alignment - never skipped, finds how badly the part is sitting on the fixture.
    3) DCC Fine alignment - never skipped, all motion is based on DCC Rough's findings.

    Also, there is one additional change that must be made: after each Level portion of the alignment and before circles are measures, the WORKPLANE really needs to be set to XPLUS for those circles to get good data. This is probably contributing to some of that positional error.

    ~~~

    Finally, based on your keen observations of the issue and evident problem-solving skills in coming to the forum to sort it out, I predict that you will soon outshine your coworker who made that program. Good luck and keep us posted!
Reply
  • Hello and welcome to the forum, Amber!


    The part is secure, but due to different cavities with different die conditions may sit slightly different in the fixture each time.

    This is the red flag clue.

    What's happening is that when you run the Manual "the part's sitting in this part of the ballpark" alignment on Part #1, the machine knows exactly where #1's datums are and can go into DCC mode and beep them with little to no error, getting a good alignment on them and making a good report.

    And then when you grab part #2 (which is a little different, being from a different cavity made on a different shift etc) and skip the Manual alignment, the machine goes into DCC mode and aims for where it thinks the datums should be - where they were on Part #1. Now they're not too far off, not enough to make it crash but far enough that when it beeps it's not hitting with a DCC approach vector perfectly perpendicular to the datum surface. So each datum point is now being probed with some error, we call it cosine error, getting a shoddy alignment and poor results.

    And then as you figured out simply run the Manual alignment on part #2 and it's fine.

    The solution is to layer a third alignment after 1st two:
    1) Manual alignment - can be skipped as long as the fixture isn't moved.
    2) DCC Rough alignment - never skipped, finds how badly the part is sitting on the fixture.
    3) DCC Fine alignment - never skipped, all motion is based on DCC Rough's findings.

    Also, there is one additional change that must be made: after each Level portion of the alignment and before circles are measures, the WORKPLANE really needs to be set to XPLUS for those circles to get good data. This is probably contributing to some of that positional error.

    ~~~

    Finally, based on your keen observations of the issue and evident problem-solving skills in coming to the forum to sort it out, I predict that you will soon outshine your coworker who made that program. Good luck and keep us posted!
Children
No Data
© 2024 Hexagon AB and/or its subsidiaries. Privacy Policy | Cloud Services Agreement