Profile of a Line on Interrupted Segments

[[Ch...]]

I have a problem which seems extremely simple but for some reason is giving me a ton of trouble. I am attempting to take what is essentially profile of a line around the inside of a rectangle. The issue lies in the fixturing and sharp corners requiring it to be scanned in six separate sections. I have split it up into six line scan paths and am attempting to recall them into one curve so it can be evaluated but the tolerance is getting very confused in the corners and also is interpolating my scanned points across the gaps which I do not want. I just want a profile tolerance based on the CAD model, controlled by the drf with the actual points being evaluated to it. The attached picture is a sketch of what I am trying to do, with the green being my actuals vs the closest I can get calypso to manage it. I have tried every combination of features types I can think of, 2d lines, 2d curves, 3d lines, 3d curves. 

[[Ch...]]

I have figured out how to correctly put all the segments into one curve however I am still seeing odd behavior in the corners of the tolerance and can't get it to stop interpolating the gaps in the curve. 

Edited 20 hours ago

[[Ma...]]

I think you would have the best results with freeform if you have a CAD model. Just recall measured points into freeform feature.

Also you can try to mask unused points, which will result in strange tolerance also, but it will be calculated.

Edited 20 hours ago

[[Ch...]]

Please sign in to view this quote.

Unfortunately my license for freeform did not get renewed so I am stuck trying manage with just curves until next cal cycle. It is just a profile of a line however so I feel like it should be simple as I have done similar with just curve no problem in the past. Its just the interruptions in the corners and fixturing that is really messing this one up. I have masked out the unused points in the best fit but they still are being interpolated and the only exclude option I see only allows for one continuous excluded section, not multiple like I need. 

[[Iv...]]

How are you creating the nominals for the recalled curve?

• My preferred method is using the connect nominal curves (screenshot below). This will only pull the nominals; you then have to recall feature points from the curves to bring over the actuals.
• You can also use the "re-calculate nominal geometry" option from the recall feature points pop-up. But make sure there is no actual data in your curve since it will recall the actual for the nominal data
• For both methods, I would recommend using the curve results points option when recalling the feature points.
• Another option would be to scan over those sharp corners and mask out those points. This option can be found in the evaluation window under the limit evaluation. This looks like it would reduce the number of scans to 2.
  • Maybe an even better option is using only one curve. Create the full curve. Expand the point list and delete the points that are blocked by the fixture. Lastly, create two segments with a clearance plane in between. Update the start and end points accordingly. 
    • This would be my preferred method since it does not involve any recall or combinations of curves. Make sure to make things easier by having the curve start adjacent to one of the fixture locations. 

[[Ch...]]

Please sign in to view this quote.

I did end up going with the last option however the corners still are having issues with the tolerance. They are not smoothly following the profile of the nominal curve as expected and are even a bit tangled in some. Not sure why as the actual curve looks to be correctly matching the CAD. 

[[Ma...]]

You can have big probe tip for that corner radius and low number of nominal and/or actual points.

Point distance is determined by center of probe tip ( i think ), so use smaller tip, reduce speed close to corner and use more points ( can be done in segments, then grouped to keep one scan session )

[[Je...]]

Hi

Please sign in to view this username.

.

Nice work on troubleshooting at least half of the problem.

There are a few ways to get those pesky corners omitted from the zone.

1. Under the evaluation tab for the curve, choose the "Limit Evaluation" option in the middle of that window.  Then, you will be able to mask out the points near those corners. 

   -  Heads-up: This UI can be confusing at first.  Make sure you turn on "Show point numbers" before opening the Limit Evaluation window.  Then, scroll down to the points you would like to mask.  While holding down the control key, click the individual points you would like to mask.  If there are several points in a series to mask, you can hold down shift after clicking the first point in that series and click the last point in that series, and it will choose all points between.  However, if you don't hold down either control or shift, Calypso will only choose one point and reset all others.

Tip: If acceptable per your task, mask a few points on the edge or even outside the corners to make sure outliers don't sneak in to the unmasked area.

2. Newer versions of Calypso facilitate more intricate curve constructions, similar to how

Please sign in to view this username.

 used the connect nominal curves optoin.  You can create a curve with nominal point segments for the lines and delete or omit the corners when creating the original curve.  This can sometimes have undesirable effects on the tolerance zones, either visually or numerically.  I prefer option #1.

Tip: Under the evaluation tab, there is an option called "take segment gaps into consideration." You may want to check that option, but I'll be honest that I don't fully understand what's behind the curtain on that function.

Tip: You can scan the features of interest as line segments and then recall them or scan entire curve features all-in-one (or a hybrid strategy to avoid obstacles).  Curves are like their own Open-World video game.  Lots to explore.

If you have a ZeissCare agreement, I strongly recommend submitting a ticket and asking the Zeiss application specialists to review your measurement plan.  I've learned as much about curves and freeform surfaces from those calls as I have from the actual Curve/Freeform training course.