Surface profile

[[Cl...]]

I am recalling eight 2d curves into a Free Form Surface, the result is 6 inches, Impossible.
When I evaluate each curve separately using line profile, my results are much more believable.
Anywhere from .002" - .003" I've rebuild the FFS and no change. Both methods using
the same Datum Ref. Why the huge disparity?

[[SH...]]

Please sign in to view this quote.

What about the maximum and minimum points in each case. Did the feature is twilted or something like that... I have not worked with FFS.But six inches is like a fairy tale story.....

[[Ma...]]

Are faces on CAD model in right vector? Could it be deviation + radius of a tip *2

[[Da...]]

Check the nominal faces in the free form feature.

I have found the best way to recall into a free form is to:
-create a new freeform (usually through cad->creating features). Put a single point or two on each surface you have points from.
-open the feature, and click the "Surfaces Used" button. Double check there are not any unintended surfaces, and that all the intended faces are there.
-Close the free form window
-Reopen the free form, then do your recall. Uncheck "recalculate nominal geometry" before you click any features.

I think this forces calypso to recognize which CAD surfaces the freeform is associated with. It doesn't seem to work reliably any other way.

To be clear, this is a process I came up with myself by experimentation. I have had apps people tell me straight up not to recall into free form, but it seems to agree well with other measuring methods.

[[Ro...]]

- Why would you want to recall points?

- What exactly are you measuring?

If your trying to create 3D geometry with 2D geometry I believe Calypso will not interpret it correctly based off of work I have completed in the past.

[[Cl...]]

Measuring the surface profile of a cylindrical OD. I always us 2d, or 3d curves depending on the surface and have not ever had this issue. The curves appear perfectly normal until the are recalled into the FFS. I even reported the profile using line profile with 2d curves and the operator made an adjustment based on my numbers and the next (adjusted) part correlated to my suggested move.Using CAD evaluation I can see the points laying on the model correctly and are in tolerance based on the color graphic, but the FFS shows a deviation of up to 6.0"

[[Me...]]

What are the Features in Alignment D & A?

[[Cl...]]

Turns out the 2d curves are actually lines. Normally Calypso will warn me that I
can't use lines. Never got that warning. I re-did them as 3d curves and now it works.

I regularly use 2d curves to construct a FFS, without issue.

[[Cl...]]

Using 3d curves I'm right back to were I started.

If I extract a FFS from faces (individually). They each evaluate as expected.
Why can't I recall the points from each individual FFS into one FFS? I can't
pick all the surfaces and make one FFS because the points aren't organized
in a way that allows for continuous scanning.

CAD_EVAL.PNG

[[Er...]]

I don't have FFS. But I have had Profile give weird numbers when the feature was not in the same alignment as the callout in the characteristic.

[[Ja...]]

I dont know much about FFS but did you apply the same filter to FFS as that in your 2d curve eval?
Does your alignment D-A constrain all degrees of freedom? If not, try geometry best fit alignment to constrain the rest of the DOF using FFS feature and use that in your characteristics.
Also, what do you get when you choose base alignment instead of alignment D-A?

[[Cl...]]

Filtering the FFS does not work like a curve. I tried D & A individually and as a secondary alignment, still same problem.
Can't do a best-fit with on a FFS that is used for profile evaluation relative to a Datum(s).

[[Ro...]]

Sorry about the questions but I believe your close to a resolution*

I suspect the alignment being the issue.

Current assumptions:
1. Starting program at Base Alignment PN (CNC), Clear all results when you start the program.
2. Part doesn't move.
3. Machine is clean and calibrated/Styli.

Alignment Questions:
1. Are these just 2 features?
2. Are these features perpendicular to each other?
3. What are these features and how are they constructed?
4. Have you tried to measure distances referencing the datums to this desired characteristic other ways?
5. Are you moving the machine in the base alignment while it takes scans then recalling a constructed alignment in characteristics for profile?
6. Did you try to use line profile?

To troubleshoot your current alignment you can always create a local separate alignment with new features to apply this Profile Characteristic.

Not sure how much information you need but I keep it simple by creating where appropriate to troubleshoot an alignment:
1. Plane, Line, Point
or
2. 3 Perpendicular Planes
or
3. Spatial-(Cylinder) Perpendicular to Second Feature Planar- (Plane/Cylinder/Line)

IF it is the alignment you may have to modify your features strategy appropriately to while within customer print while making sure Calypso interprets the features actual measured points correctly.

Also as a side note I prefer to keep things simple and less prone to error by creating a 3D feature in Calypso for a 3D Characteristic then modifying the strategy if possible which is where my original comment comes from.

Also new to the forum*** 🙂

[[Ma...]]

Did you tried to select faces in FFS element? There is elimination of points when switching faces.
Usually it will take nominals and autoselect faces on model, but it's not considering move arounds ( segments ), so some points can be on other face and Calypso did not registered it.

Or you can create alignment from curves and try it in profile to see result.

And i personally rather use cylinders instead of spheres 😉

[[Cl...]]

Yes, I tried selecting from faces. The model is divided up into 4 separate faces. The top two extract fine with points in a fairly sequential order. I had to make some minor adjustments and add a few clearance positions. The bottom two faces however
extract with points in no discernable order. It would take a full day to re-arrange them in sequence. I don't have the luxury of time.

I have not tried alignment from curves. Not sure that's even possible? I have a profile to datum's? I thought that was ONLY for
profile to it's self.

[[Da...]]

is there a reason you want to recall to a FFS ? Can't you recall to a cylinder then ask for profile ?

[[Cl...]]

Please sign in to view this quote.

It's not a cylinder. The geometry is actually three cylinders and two planes.
One cylinder is angled slightly off from the other two. All this forms the shape.

[[Da...]]

Is the model complete and free of gaps ?

[[Ma...]]

Just to clarify - you have no bestfit enabled in ffs right?

Can you try making alignment from curves? bestfitting in it - if it will give you wrong numbers then i am at end of ideas.

[[Da...]]

Clarke, Did you resolve this issue???

[[Cl...]]

No. Zeiss remoted in, tried the same thing I did and failed. Then he was able to create a FFS from faces an/or curves. The only problem with that is, the points are not in order, so I had to manually re-arrange them so I could scan the surface(s) uninterrupted.

[[Ma...]]

Sorry for waking up this thread, but i found something.

Now I have same problem. After several attempts i found inverted faces - must click on them to correct them. Then SP was OK.
I make curves with curvature option - and there i noticed, that there is something bad. It's not making nominals.