6-point nest iterative alignments without CAD

John Dee
Regular User
Regular User
Posts: 20
Joined: Wed Nov 14, 2018 10:07 pm

6-point nest iterative alignments without CAD

Post by John Dee » Thu Dec 06, 2018 12:44 am

Hello,

No CAD capabilities here (unfortunately). I'm trying to close in on the nest within .0005 without using CAD. I have been repeating the alignments but with 6 iterations, I'm still seeing deviations of .003".

Does CMM Manager have an iterative option without the need for CAD?

Any help would be much appreciated. I'm kind of on my own here.

Thanks!

medupriest
SUDO
SUDO
Posts: 423
Joined: Fri Mar 13, 2015 11:59 am

Re: 6-point nest iterative alignments without CAD

Post by medupriest » Fri Dec 07, 2018 2:58 pm

The datum alignment is iterative and does not require CAD. You can key in or import points with appropriate x,y,z,i,j,k.

John Dee
Regular User
Regular User
Posts: 20
Joined: Wed Nov 14, 2018 10:07 pm

Re: 6-point nest iterative alignments without CAD

Post by John Dee » Sun Dec 09, 2018 2:44 pm

Ok. I tried your suggestion. First I provided a rough origin. then I keyed in the nominals and vectors to a iterative 321 checking the bestfit output report. From that newly established origin (from that datum alignment), I began a second datum alignment but for a different datum reference frame. when the reports came out for the first and second datum reference frame, the first had nominal figures that were slightly off from the nominals that I had keyed in. The second report (for the 2nd datum reference frame) had correct nominals. Why did the first report show different nominals? I reviewed the program and all the keyed in points were exact. Any ideas as to why the nominals changed?

User avatar
CrashN8
Site Admin
Site Admin
Posts: 576
Joined: Thu Feb 09, 2017 4:07 pm

Re: 6-point nest iterative alignments without CAD

Post by CrashN8 » Tue Dec 11, 2018 11:16 am

John Dee wrote:
Sun Dec 09, 2018 2:44 pm
I began a second datum alignment but for a different datum reference frame. when the reports came out for the first and second datum reference frame, the first had nominal figures that were slightly off from the nominals that I had keyed in.
I"ll guess the shift from one Ref Frame to another is the culprit. You will need to carefully enter nominal values for features and report outputs if programming without CAD.

John Dee
Regular User
Regular User
Posts: 20
Joined: Wed Nov 14, 2018 10:07 pm

Re: 6-point nest iterative alignments without CAD

Post by John Dee » Tue Dec 11, 2018 5:49 pm

I'm not certain that solves the problem:

The program was merely a program consisting of two datum alignments--one after another--with the second one starting off from where the first one finished. In other words, the second datum alignment used the final origin and alignment of the first to begin its procession of iterations (they both essentially have the same origin).

The best fit output report produces two datum alignment results (one for each datum alignment conclusion). The second datum alignment showed exact nominals in the datum alignment best fit output report. Yet the first datum alignment best fit output showed slightly different nominals from where they should be. So the second datum reference frame is on target while the first is not probing the nominal target points--that I personally keyed in. All nominal inputs in the program were keyed in by me, and upon run completion, and a review of the program, ALL nominals are correct and exact. Yet the report lists different nominals.

This is a problem because I cannot proceed forward with using the datum alignment function in my measuring environment until I can explain why the nominals of the first datum alignment are not those listed on the engineering drawing.

Any immediate help would be much appreciated. Thanks!

User avatar
CrashN8
Site Admin
Site Admin
Posts: 576
Joined: Thu Feb 09, 2017 4:07 pm

Re: 6-point nest iterative alignments without CAD

Post by CrashN8 » Wed Dec 12, 2018 10:39 am

First of all... Why two alignments? Just use one... you know one works, you can set the iteration control to establish the desired number of iterations or max RMS deviation and be done with it. Not sure what behavior you expect in running the same alignment twice other than unearthing a potential flaw in CMM-Manager.

Secondly... Send that project file to Help Desk and ask them to investigate. If we can replicate this issue here in development office maybe we can get to the bottom of this.

John Dee
Regular User
Regular User
Posts: 20
Joined: Wed Nov 14, 2018 10:07 pm

Re: 6-point nest iterative alignments without CAD

Post by John Dee » Thu Dec 13, 2018 10:11 pm

The datum alignments are not the same. They are two different datum reference frames.

As for sending the project, I cannot do that (unfortunately), due confidentiality.

With that said, is there anything you can do to help explain why the nominals in the first datum reference frame's Datum alignment's best fit output report are not exact as I had input?

User avatar
CrashN8
Site Admin
Site Admin
Posts: 576
Joined: Thu Feb 09, 2017 4:07 pm

Re: 6-point nest iterative alignments without CAD

Post by CrashN8 » Fri Dec 14, 2018 10:21 am

How do you determine the Nominal values have shifted? Are you reporting something after the alignment? Are you reviewing the Nominal Value from Feature Database? I can provide almost zero help without a project file.

Edit - Maybe a better question. Is the first Datum Alignment occuring in Machine Re Frame? If so, then I expect the Nominal Values to update EVERY time you run the program. You need to read this post - viewtopic.php?f=39&t=805 - I think I suggested you read this already... please do.

John Dee
Regular User
Regular User
Posts: 20
Joined: Wed Nov 14, 2018 10:07 pm

Re: 6-point nest iterative alignments without CAD

Post by John Dee » Fri Dec 14, 2018 9:12 pm

Ok. Never mind the second alignment. It doesn't exist. I created a rough origin. Ok? then I programmed a Datum alignment--inputing 6 target points and their vectors. Then I ran the program. At the conclusion of the run, the Best fit report appeared showing nominals that were not identical with those I input into the datum alignment as the 6 target points (although the datum alignment ran through 3-4 iterations in order to meet the .0003 RMS requirement I set.

If the datum alignment met the requirement, then why do the report's nominals deviate from those I input? deviations of up to .010"?

User avatar
CrashN8
Site Admin
Site Admin
Posts: 576
Joined: Thu Feb 09, 2017 4:07 pm

Re: 6-point nest iterative alignments without CAD

Post by CrashN8 » Mon Dec 17, 2018 10:22 am

I'll look into it... It could very well be similar situation to Cloud to CAD Report and Alignment, where a new nominal is calculated each program run due to position capability of CMM. So this could be by design... Can you confirm that only the position of each point is changing or does it also change along the vector of the point? Also are you using "Iterative Bestfit" or "Iterative 3-2-1" option in the Datum Alignment dialog?

Edit - After extensive testing I do see some strange behavior. I think this is new is last few releases and may have an idea already what the problem is. For now, you just need to be sure that you add a manual alignment immediately before the Datum Alignment. Please be sure this manual alignment matches the Datum structure of the following Datum Alignment. If you have CAD model the work-around is simply adding "Sync Part Ref Frame" operation immediately before the Datum Alignment.

Post Reply