hexagon logo

Alignment Problems in v4.3 w/Infinity

I should preface this with: I have used 3.9 and 4.2 for several years, but just started using 4.3 with our Romer Arm. I have no training on 4.3 or on the Arm, I have gotten everything from the help file.

So far, I haven't had trouble (after figuring out how to create a Profile and calibrate probes, which no one had EVER done) with part programming, except that when I go back to run a program I've written, it's like the program doesn't know how to deal with my alignments since the parts aren't fixtured in exactly the same place as when I wrote the program.

For example, I have some 41 inch rings with boltholes in them, and one of these holes is easily identifiable as the Datum hole. During program composition, I had no alignment problems, true-position worked fine, etc. Going back to run the program, I take the datum hole and the alignment seems to work partially, excepting that it reverses the direction that it derives from the constructed hole I rotate Y+ to. However, it's not just the Y+ flip which is the problem... as far as the boltholes created after the alignment is concerned, it completely loses track of the rotational component of the alignment and tries to tell me they're at all sorts of seemingly-arbitrary angles.

On the other CMMs I've used, there was nothing like this... everything was relative to the active coordinate system and the actual orientation didn't matter at all. Is there a setting I'm missing? I already have the "Treat coordinates as if stored in PCS" option checked... Neutral face Is 4.3 different in this regard? I'm pretty stumped.

Here's a simplified example just to illustrate what happens:
Run #1
Take plane
Take circle
Take bolthole
Construct Line from circle to bolthole
ALIGN:
Level to plane
Translate Z to plane
Translate X and Y to circle
Rotate Y+ to Line about Z+
Take second hole:
Theo: PR=20.00, PA=80.00, Z=-0.100
Act: PR=20.05, PA=80.32, Z=-0.114

Between Runs: Turn part +30 degrees.

Run #2
Take plane
Take circle
Take bolthole
Construct Line from circle to bolthole
ALIGN:
Level to plane
Translate Z to plane
Translate X and Y to circle
Rotate Y+ to Line about Z+
Take second hole:
Theo: PR=20.00, PA=80.00, Z=-0.100
Act: PR=20.05, PA=50.32, Z=-0.114

Of course, the hit targets are effected in a similar manner. I'm pulling out hair here, because it means I can't really use any of my programs. Any suggestions?
Parents
  • Here's another thought. If you are constructing the line between circles while you are still in the "startup" reference system, I would take a step back and get at least the first axis (level) before you create the line. Frankly, if I am using multi-point elements that are workplane-dependent (circle, line, etc.) for an alignment I always set at least the first axis before measuring/constructing those elements.

    I am not sure that this is entirely necessary, but it makes me feel better about giving the software a better opportunity to correctly solve the problem. Slight smile It feels more solid to me and less prone to doing weird stuff like you are experiencing...

    Hope that helps.

    Scott


    +1 especially on a PCMM. Levelling before continuing measurements is more important on a portable unit then on a conventional CMM, though it's good practice on any CMM.

    Can you post some code for us to look at as opposed to an explanation of the process? It might help catch the error.
Reply
  • Here's another thought. If you are constructing the line between circles while you are still in the "startup" reference system, I would take a step back and get at least the first axis (level) before you create the line. Frankly, if I am using multi-point elements that are workplane-dependent (circle, line, etc.) for an alignment I always set at least the first axis before measuring/constructing those elements.

    I am not sure that this is entirely necessary, but it makes me feel better about giving the software a better opportunity to correctly solve the problem. Slight smile It feels more solid to me and less prone to doing weird stuff like you are experiencing...

    Hope that helps.

    Scott


    +1 especially on a PCMM. Levelling before continuing measurements is more important on a portable unit then on a conventional CMM, though it's good practice on any CMM.

    Can you post some code for us to look at as opposed to an explanation of the process? It might help catch the error.
Children
No Data