So many curves!

[[Je...]]

.

Does anyone have an idea on how to optimize the following:


I have to measure a lot of curves and freeform surfaces.

1. Features need to be measured at a minimum of three equally spaced depths (sections).
2. Each of the three sections are evaluated as curves using the profile of a line characteristic.
3. Line plots for each curve must be reported along with numeric values for profile deviation.
4. These surfaces must also be evaluated cumulatively as a freeform surface using the profile of a surface characteristic.


My process is:

1. Use "Creating Features" option to create three sections.

2. Select segments for evaluation from the CAD hierarchy tab, which now contains the sections I created in step one.

3. Create a curve from the selected segments.

4. Manually pick points not to be evaluated, such as radii too small for tactile probing, using the "Limit Evaluation" option. This step has proven tedious but necessary to achieve plausible results. The other option would be to create multiple paths for probing and skip over the non-essential segments, but I cringe at all the stop-and-go probing inside of a single feature.

5. Repeat steps one through four for the remaining two sections.

6. Create a blank freeform surface and select the appropriate CAD surfaces using the "Surfaces Used" button.

7. Recall points from the previously created curves making sure to uncheck the "Re-calculate Nominal Geometry" button before selecting the curves.

8. Repeat steps 1 through 7 for the remaining quadrillion features.


Unfortunately, I cannot simply extract freeform features with auto-generated scanning paths because of the requirement to report three equally-spaced sections.

Any ideas, or should I just embrace the grind?


.
.



Jeff Frodermann
Meier Tool & Engineering
Anoka, Minnesota

[[Cl...]]

So, you are sectioning and creating several 3d curves and then recalling
them into FFS? (Trying to fully understand what you are doing)

[[Je...]]

.

Please sign in to view this quote.

Clarke,

I am sectioning and creating several 2d curves and then recalling them into a FFS. I was instructed that 3d curves are only necessary for threads and other objects that need the third vector considered.


Jeff Frodermann
Meier Tool & Engineering
Anoka, Minnesota

.

[[Cl...]]

Interesting. I was told the opposite. That 3d curves needed to be used if recalling into a (3D) freeform surface.

[[Ma...]]

I am using 3d curves where i can not garant right scanning vectors.

Using 2d curves brings more options for bestfitting method.

In authors case i would use 2d curves too and i don't see any way to big improve.

[[Cl...]]

Please sign in to view this quote.

I can see using 2d curves is just fine in this case, as the surfaces are square and prismatic.
If they were all contoured and curved with obtuse vectors, then 3d curves would be better.

Sorry it's a little off subject, but back to your original question "should I just embrace the grind?"
I would to exactly what you are doing, looks like it should work just fine.

[[Jo...]]

Jeff

"4. Manually pick points not to be evaluated, such as radii too small for tactile probing, using the "Limit Evaluation" option. This step has proven tedious but necessary to achieve plausible results. The other option would be to create multiple paths for probing and skip over the non-essential segments, but I cringe at all the stop-and-go probing inside of a single feature."

There is a way ,maybe using multiple segments, to scan continuously. There is no stop start. I can't remember where I stumbled upon that info, but it was researching multiple segment tolerances and strategies. You can change speed and point density per segment, but scan the entire thing in one go. Someone doing airfoils or blades can chime in on that.

[[Ow...]]

Please sign in to view this quote.

You mean like this?

[[Pe...]]

Are the three curves identical, except for the elevation?

[[Ma...]]

You can make some segments and then group them to scan continuously - but remember - when you will want to recall just one of segment ( recall points -> open feature -> select path(1) ) then all data will be in first segment of group.

As author mentioned - i would go same way - exclude corners from eval, they are not needed ( as far as i know from machining )

[[He...]]

If the slots are the same shape for all the three sections you could make just one curve and copy it and change the axial position. If all the slots are like this and all section are to be measured at the same height you could finish all curves for level 1 and then copy them all twice. Then you make two new alignments at the correct heights and change the alignment with "keep position" active for all curves. This will move the nominal curves to the new location. Then change back the alignment to the original one with "keep position" inactive this will not move the curves but transform them to the correct alignment.

You can use the paintbrush to do this for multiple curves. Still quite some work but a bit better than doing them all individually.

[[Je...]]

Please sign in to view this quote.

.
BOOM! That is exactly the type of improvement I was looking for, Henrik. I have implemented the first suggestion before by creating the first section and then translating the z nominals of the curve using the "change nominal vectors," but I have never attempted to move multiple features at once using the "keep position" option. I will try this out. It could potentially save a lot of time. Thank you!!! ❗


Jeff Frodermann
Meier Tool & Engineering
Anoka, Minnesoat
.