Chord Height on Curves

[[Jo...]]

I know there's a whole separate section of the forum for Curve/FF, but I feel that section sees a fraction of the traffic as the "main" Calypso section. I digress....

Can someone please explain to me what a chord height actually represents? We went over it in training, and maybe something just didn't stick, but it seems to be an important factor in creating a well constructed Curve. I understand that it concentrates data density in areas with high curvature while remaining more spread out in straight or low curvature area, presumably to increase accuracy.

Under the "Modify Nominals" menu, where you can manipulate chord height, min and max point distances is where I am unclear as to what I am changing. We were given "recommended defaults" but I like to have an understanding of their function, and if the defaults are really the best settings for all measuring situations, or if it is appropriate to change these values based on things like part size or if a Curve has sections both of very high curvature with small geometry, and long areas with low or no curvature.

[[Ma...]]

If you know how this looks on circle ( circular segment ) then imagine, that it's not circle where is taken height, but nominal spline.

On attachment is highlighted:
L - that is how far away will be points between each other.
h - distance between line created by two points ( "c" ) and actual nominal curve

This is how i see it.
You can test it - make 2 default curves, and apply different chord height with same min/max point distance.

I was playing with it few years ago - no matter how you set up nominals - eval will be always depending on scanning strategy.
Not using it anymore - no matter what effort i made - result were almost same at corners ( mainly inner )

But still waiting for someone, who knows more - i had no ZEISS training on anything in Calypso

[[Jo...]]

Please sign in to view this quote.

Thanks, a visual representation is just what I needed. When I saw this, a bunch of stuff "clicked" back into place.

Please sign in to view this quote.

I did take the Curve training class a few months ago, and much like most things, just because you "took a course" in something doesn't necessarily mean you know more about it than someone who has years of real-world experience. I still have a lot to learn about it to use it to its (supposed) potential. But, if you want a second take on it or are interested in how they instructed during the class, to boil it all down it seems to be a combination of modifying the number of nominal points - via "Modify Nominals" menu - after creating the curve (where they are evenly spaced out) vs. how you want the nominals to be distributed (more concentrated in areas of high curvature), which is where I think the chord height comes into play. Since the number of actual points you gather on the curve --according to Zeiss-- should be at least three times the number of nominal points assigned to the curve, the chord height seems to be a way to get a more accurate measurement due to using chord height on actuals vs. nominals.

I had tested out a few things to tease out what I could on my own, not just for using chord height but learning how to use it in general. I too have noticed little difference in measurement results, until I crossed a part to program that had an outside closed spline around the perimeter of the part. The part, oversimplified, was like a giant almond shape. Two top and bottom "lobes" of it had very large rads in excess of 13 inches or so and then there were parts of the spline that had tangent rads to those that were very small, or large rads with small arc lengths. There was probably 45-50 inches of linear scanning distance for the closed 2D curve on this part, and it had a profile tolerance of .001".... that was really the only time I saw significant changes when I started tinkering with those settings.