Creating a Surface that Intersects

[bigjimslade]

I have extruded 2 arcs and intersected with a revolution.

Then I added tangent curves to extend and and arcs to fill out the shape.

Here I have created an networksrf from those frames (I get the same problem with sweeps).

I am trying to create a bump on the green surface.

If I try to split the green surface with the blue, I get the two shapes do not intersect.

I started out with a blocked box using the blue shape and tried to do a boolean union and I get the same error that they do not intersect.

How can I create such a bump on the revolution?

I need to end up with:

Problem Intersection.3dm (3.1 MB)

[jeff_hammond]

if you make the tolerances smaller in the NetworkSrf command, you’ll get a surface that will intersect (at the expense of a more complex surface)

…that’s .00001 for edge curves… the resulting surface will then fully intersect with the green thing.

[jim]

@bigjimslade

How can I create such a bump on the revolution?

Seems to me that if you want to make a surface that intersects the simplest, easiest and most accurate way to do that is make the surface extend far enough so that it will intersect and then trim it back using the intersection.

NoProblemIntersect.3dm (95.3 KB)

[bigjimslade]

I’m not getting an intersection with the precision decreased to 0.00001. If I go down to 0.000001 I silently get no network srf.

[davidcockey]

@bigjimslade

Then I added tangent curves to extend and and arcs to fill out the shape.

That’s the route cause of your problem. The extended purple curve has discontinuous curvature.

NetworkSrf creates a surface which is curvature (G2) continuous everwhere including along the edges. If an input curve has discontinuous curvature NetworkSrf determines a surface which approximately fits through the discontinuous curve. The closeness of the fit of the result of NetworkSrf to the input curves is determined in part by the tolerance setting in NetworkSrf.

You need to either create a continuous curvature, or build two surfaces which a join at the curvature discontinuity.

[davidcockey]

@bigjimslade What overall “absolute tolerance” are you using?

[bigjimslade]

I generally use:

0.0001

I find that if I make it larger (0.001), I get details that disappear.

Here I continued the circle and ellipse frames to form complete loops so that the single surface clearly passes all the way through and I get the same no intersection error if I try to split.

[bigjimslade]

I have found that i can split the revolved curve surface if I explode. But it makes things very tough to break everything out.

I also have to increase the number of control points on the revolution curve.

[jeff_hammond]

@bigjimslade

I’m not getting an intersection with the precision decreased to 0.00001. If I go down to 0.000001 I silently get no network srf.

hmm… it works ok here at .00001NetworkSrf tolerance. (and no changes to the file settings/tolerances)

here’s the file:
Problem Intersection_j.3dm (2.5 MB)

the surface intersects ok… trims & joins the green surface well.

[emilio]

Hi Jim

If you explode the green object, then you are able to trim its side surface using the frame curves.
The green and the blue surfaces are then joinable without problems.
( I’ve found that sometimes surface edges can trim what the surface itself is unable to trim … )

HTH, regards

[bigjimslade]

Thanks all.