C2 fillet

[cloughna]

I am trying to make a very simple curvature continuous fillet. I literally have a 3x3x3 cube. I then create a boundary blend between two sketched lines to create a fillet. When i set the end conditions to tangent, I am able to merge this surface and get my fillet. However, if I set the end conditions to 'curvature', it blows up. I have tried tweaking the accuracy to no avail.

Any thoughts, or perhaps a better way to do this fillet?

 

[Jacoolo]

without a picture of the situation You are trying to solve it is difficult to give exact answer.

However - taking as the base the conditions You mentioned in Your post I bet the cause of problem are sketched curve.

To obtain Curvature transition in Boundary Blends, the curve which builds this geometry mus have curvature transition either.

Suppose Your curves does not maintain G2 transition?

 

[zpaolo]

I think he is drawing straight lines parallel to the edge that needs to be rounded, but I can't obtain results in this way: it looks like the curvature continuity doesn't allow to "cut" the solid cube.
Just do the cube as a surface (with closed ends), draw the blend as you already did and trim the cube with the blend. This way it works.

Edited by: zpaolo

 

[jbuckl]

Excuse the ignorance, why not use the round tool? (what am I missing here?)

 

[zpaolo]

Excuse the ignorance, why not use the round tool? (what am I missing here?)

The round tool does a round (ehm), which is not curvature continuous on the edges, it is just tangent. This is not an issue, a perfect round has constant curvature, a flat surface has zero curvature, so the iterface will have a curvature discontinuity. In WF 5.0you have G2 rounds, but in earlier versions you have to use surfaces to obtain the same result.

Paolo

 

[design-engine]

Generating rounds this way ... with surfaces is common practice amongs the industrial designers using Alias Studio in the automobile industry.Pro/ENGINEER handles this type of geometry well becase of it's parent child relations and ability to force tangencies on four boundaries and the further functionality of continious curvature along any number of the four boundaries. The correct term is G2 continuity named from "I believe" Guassian the mathamatician who developed the calculus required for magnetic fields responsible for creating the electric motor.... and the same the math for evaluating the positional reference of the comb plot at the joining curves (that maintian curvature).

I beta tested the WF 5.0 round tool and thats a long time coming! Unigraphics has had that tool since the 90's.

I did not really say much but I have left the board alone for sometime and thought I would toss in a comment or two.

 

[mgnt8]

I could be wrong but I think the 'G' in G1, G2, etc was for 'geometric' as in Geometric Continuity

You might be thinking of Gaussian Curvature which is more a measurement of topology

 

[cloughna]

Just wanted to update, I got the round using the trim method suggested above.

Can't wait for my clients to switch to WF5

 

[Jacoolo]

I doubt round with Curvature conditions will solve more demanded cases. It would be reserved for simple geometry and serve as fine quick makro. Just nice trick to proud self horn(anyway each step forward regarding Pro/E is a plus!).

Same with Advanced rounds with their Transition options - they are cute but mostly worthles. Each time I want to solve rounds trasnition, Round tool fails and I am left to do it manualy.

Same with Pieces - try to separate edge form a tangent chain and then stop its end transition to the vertex(end of edge). Mostly it works fine(condition: only for separate quilt). But change then something in the existing geom or try to create BB which should use it. With out a lot of legwork You end with nothing

just my two cents...

 

[jbuckl]

Would I be correct in saying that the requirement is to generate 'rounds that have no tangency horizon lines'?
I saw/used a method for creating surfaces that did this, using composite curves I belive. The tangency horizons were far fewer and number of surface patches lower too.
The gaussean reflection was less interrupted too.
Am I along the right 'lines'

here?

Wheres the huge advantage in this please?

 

[dtate]

I think the main problem here is that you're trying to create a continuous curvature fillet to surfaces that are flat (no curvature) - you can't do it. Flat planar surfaces shouldn't exist in the world of c-2 surfacing - each surface should have at least a slight curvature or dome, to allow adjacent surfaces to be continuous.

 

[zpaolo]

Flat planar surfaces shouldn't exist in the world of c-2 surfacing - each surface should have at least a slight curvature or dome, to allow adjacent surfaces to be continuous.

This might be true in theory, but I can obtain C2 fillets (and rounds) also with flat surfaces. The problem in the previous post was in the procedure, the trimming of a C2 surface on flat walls does not work always well

 

[burre]

I am trying to make a very simple curvature continuous fillet. I literally have a 3x3x3 cube. I then create a boundary blend between two sketched lines to create a fillet. When i set the end conditions to tangent, I am able to merge this surface and get my fillet. However, if I set the end conditions to 'curvature', it blows up. I have tried tweaking the accuracy to no avail.

Any thoughts, or perhaps a better way to do this fillet?

When you sketch your lines do you use the "line-command"? A line is not c2, use a straight spline instead and it will work.

 

[design-engine]

dtate is right. Pro/E will let you do it but adding g2 rounds (wf5.0) is something you would never do to a box/cube.
Edited by: design-engine