Output points related to new alignment

[JohnCTB]

Greetings,
I'm using 3.7 mr3. I create an alignment on the part (call it align1) It matches my CAD exactly so I do a CAD = Part. I then measure all the required points, holes, bosses, etc. So then lets say an error has occured on the CNC machine and everything is shifted .010- .012 thou. in Y direction. I will be asked if we move the origin .006 in Y then what will everything check? I create a new alignment (call it align 2) and just change the Y origin. Then if I reoutput all the features I measured before, will they be related to the new alignment? Or do they have to be remeasured? Sometimes I can't remeasure because the part is out of the CMM room already. It does not always seem to work quite right. Am I missing something?
Any and all help appreciated.
I'm working on plastic injection molds and we reference a tooling pin hole for our origin.

Thank You,
John

[Matthew D. Hoedeman]

Greetings,
I'm using 3.7 mr3. I create an alignment on the part (call it align1) It matches my CAD exactly so I do a CAD = Part. I then measure all the required points, holes, bosses, etc. So then lets say an error has occured on the CNC machine and everything is shifted .010- .012 thou. in Y direction. I will be asked if we move the origin .006 in Y then what will everything check? I create a new alignment (call it align 2) and just change the Y origin. Then if I reoutput all the features I measured before, will they be related to the new alignment? Or do they have to be remeasured? Sometimes I can't remeasure because the part is out of the CMM room already. It does not always seem to work quite right. Am I missing something?
Any and all help appreciated.
I'm working on plastic injection molds and we reference a tooling pin hole for our origin.

Thank You,
John

No, you can not just make a new alignment and expect to see new, correct results. For an example, any vector or surface point on a radius will now no longer be hitting in the 'correct' location and will have error added in.

[JamesMannes]

What you are doing sounds ok to do. I would ask though, you do have a manual and dcc alignment before CAD=PART, correct?

[JamesMannes]

No, you can not just make a new alignment and expect to see new, correct results. For an example, any vector or surface point on a radius will now no longer be hitting in the 'correct' location and will have error added in.

He's just doing an offset and then re-reporting the measured values, why do you say he couldn't do that?

[#2#]

First, make a copy of your program to mess around with. Don't accidentally do something to your original that would make it unuseable.

Second, as long as the alignment is before all of the dimensions and you answer "yes" to the question "update commands with new part alignment" then it should work as you have described.

[T Miller]

I do that alot I put the new alignment at end of program and recall the features I need to see if they will fall in tol.

[Matthew D. Hoedeman]

He's just doing an offset and then re-reporting the measured values, why do you say he couldn't do that?

Just as I said. Lets take an extreme example, he wants to move the Y origin/pickup 2mm. Lets say there is a radius 'wrap' along the X axis. If his pick-up changes 2mm, then if he were to re-run the program instead of just making a new alignment, the actual touch of the probe, moving at a 45 degree angle to touch the radius will touch LOWER (or higher) in the Z axis and give a larger (or smaller) change than just the Y shift would. Here is a picture as seen from the X workplane. As you can see, the actual touch will now be at a different Z height that will NOT show up in the report.

[JamesMannes]

Just as I said. Lets take an extreme example, he wants to move the Y origin/pickup 2mm. Lets say there is a radius 'wrap' along the X axis. If his pick-up changes 2mm, then if he were to re-run the program instead of just making a new alignment, the actual touch of the probe, moving at a 45 degree angle to touch the radius will touch LOWER (or higher) in the Z axis and give a larger (or smaller) change than just the Y shift would. Here is a picture as seen from the X workplane. As you can see, the actual touch will now be at a different Z height that will NOT show up in the report.

[ATTACH]1231[/ATTACH]

That is true. I was thinking of linear stuff.

[Matthew D. Hoedeman]

That is true. I was thinking of linear stuff.

Yeah, it bit me once MANY years ago (more years than I like to think about).

[Hilton Roberts]

Yeah, it bit me once MANY years ago (more years than I like to think about).

The best kind of wisdom eh.....school of hard knocks...some call it experience...

[Matthew D. Hoedeman]

The best kind of wisdom eh.....school of hard knocks...some call it experience...

Those ARE the ones you NEVER forget, it sure seems like, let me tell ya!

[JohnCTB]

Thank you all,
Matt, I know I can't reoutput any hits on angled or radius surfaces due to cosine error. I only use this on points that are taken in a straight direction.

James, No I don't usually do a DCC alignment. I have done experiments and my accuracy does not get much better. A lot of you guys are dealing with a lot closer tolerances than I.

Goodluck, I always make a copy before I start playing! Learned the hard way years ago.


John