Hello -- I am wanting to create a surface for a cam follower to run on. The cam follower is making a complicated motion relative the stationary cam surface -- it is not moving in any one plane. I created 5 points on the cam axis, enough to cover the surface of the cam follower. Then I traced the curve of each point relative the future cam surface material. At that point, I created a "boundary blend" of the 5 traces. Now the problem: I have to offset the boundary blend .625" 'normal' for my 1.25" dia. roller. When I do this, I get only 1/2 the surface. If I drop down to .25" offset, I get a full surface, but it is not close to being accurate (which scares me!)
I have done this same procedure before with success. The traces make an egg-shaped path. The difference is the size of the small-end radius with my new version. My previous instance had large radii. The one I am having issues with has a very small radius on one end. One would think it is quite possible to offset this boundary blend to the outside, but it fails.
Of course, after I offset the proper amount, I "solidify" that quilt, and cut-out the path for the cam roller. I can do this with the boundary blend of the traces, and then offset that cut surface by .625", but that also fails to work. Am I overlooking something?? I am not a 'surfacing' person. Thanks for any help.
IMO, these issues are always a function of the math of the original surface (your particular surface may have some kind of wrinkle in it that is ultimately confusing the systems ability to calculate the offset nicely). Even if you cut off the top a bit and attempted to offset that (to isolate the area of the surface giving you the issue) you'd still ultimately be constructing the result. One idea would be to make the patch math different using an independent geometry and modify. Maybe if you can share the data you'll get someone stepping up to send you an answer. With you stating that the offset at 0.25" being totally inaccurate, that would feel like some kind of bug that needs reporting to tech support (we should all have confidence the numbers we stick in result in those numbers being 'truths' don't we!)
Thanks. Your question as to the surface being complicated, led me to making a boundary blend using only 2 adjacent traces, which would seem to be a simple surface, yet is representing points making traces that are rotating abut a 7" radius. Same problem: won't offset the entire surface.
Since I can "boundary blend" and solidify the 5 traces seemingly OK, I am trying to understand if it is possible to offset that cut surface somehow in my cam block? One is supposed to be able to turn a solid surface into a "quilt" and then offset that quilt, but I get a "greyed-out" offset command when attempting that.
What is the "Import Data Doctor" all about? Is that applicable here to correct the quilt?
My reference to complicated is merely an expression of how many points are in the curves, these translate into more internal math on the surfaces.
Regarding your picture, looks like an aerofoil section to me. i assume the issue you're having is at the sharp (tail end). Is the curve kindof very tightly coming back tangent into itself there? I could imagine issues in there right now for sure unless it was truly sharp.
my guess is related to the tail end of the offset mesh.
i did notice unexpectedly that my hacked up geometry didn't even like offsetting around the leading edge of the shape, i resorted to using independent geometry to produce more compliant geometry (rather than troubleshoot my original curve).
So the previous JPEG are 7 curves joined by a boundary blend, with the curves showing a radius of .0005". Still not able to handle an offset all around. I wonder what determines where the offset stops?
To attach a file use the Advanced Editor and look in the lower right corner.
I'm guessing there is a reversal in curvature or small kink that causes the surface to self-intersect or otherwise degenerate. Look at the curvature analysis (http://help.ptc.com/creo_hc/creo30_pma_hc/usascii/index.html#page/pma/surfacing/About_Curvature_Plot...)
to attach any file, you have to click Use advanced editor link located in the top right corner of Reply window. Then click Attach link located in the bottom right corner of Reply window.
Important note: Add Reply button is located at the bottom of HTML page. You have to scroll down to see it.
I would have gotten there first, but stopped to get a link to the curvature analysis - new software, more confusing interface. Nice mention on scrolling.
Thank you gentlemen. This has been very helpful. I have attached some JPEG's of my latest curve, which shows a negative portion. Does CREO warn you of the inability to offset in this case? (I'm using CREO 2)
Back again -- I sure need to solve this! Attached is an extruded curve with a back bend. Works fine, so the issue with my pointy curves has to be with the curve at the point. I will play with this. Is it possible to manually modify these 3-D curves?