Creating a Surface Coming to a Points

Here is a problem I am encountering frequently. I have surfaces that need to some to a point.
Problem46.3dm (53.8 KB)
I have faired horizontal and vertical curve. The problem is that the diagonal curve meets both horizontal and vertical lines; preventing them from being used with NetworkSrf.

If I exclude the vertical curves I get a noticeable horizontal defect.

If I exclude the vertical curves, I get a noticeable vertical defect.

And patch always seems to create defects at the point:

What is the best way to create a surface for this kind of shape?

Thanks, but let me add to the complexity. The left and right edges are bound to other surfaces as shown in this picture:

Problem46.3dm (209.8 KB)

Problem46.3dm (209.8 KB)I’m not sure how I would extend the curve then get it to trim so that it would mate.

You’re going aboat this wrong. Trying to handle this in small sections is just going to create problems that can’t be fixed. You have to deal with it in large sections that flow together. That minimize the need for n-gons/trims. Do you have a file with all the hull curves? I’ve modeled a few boat hulls of varying complexity.

1 Like

The second file I attached has the rest of the bow. The highlighted yellow in the picture above is the flat, bow half siding (that widens below).

This is how I have broken it up so far:

The red at the right (stern) I have not done other than the frames. It has the same problem that the bow has. I get a great shape using a PATCH (even on the twin keels) but the stern is totally FUed.

The black segment I split separately because all the frames there have straight segments at the top (and looks fine with a loft)

The cyan segment builds fine. Here the frames are all curves (except for the shared aft-most. That extends until the frame where the bottom starts to rise.

I could probably do a single loft and build the black and cyan pieces together.

I broke the bow up into three pieces:

  1. The bulbous section at the bottom (Has the same problem with loft as the piece above but looks fine with patch).
  2. The section above extending to the FP (looks fine with Networksrf).
  3. The stem, extending beyond the FP.

Along the centerline, everything is flat, perpendicular to the axis of the ship (shown in orange below).

The second file just has a couple of additional surfaces. However just from that I can see a big issue. The bow is really pinched in the mid section.

Good models start with good smooth curves. That fully define the shape of the object. Your curves have an unnecessary number of cv’s with wildly varying counts. By using netwerk surf you are rebuilding these surfaces with higher more evenly distributed point counts. The net result is you have surface continuity approximated over a very short distance. The surfaces will never flow smoothly.


Sadly, the number of point counts varies with the height of the curve. The curves are generated from waterlines, so frames running top to bottom, should have the same number of curves. It’s when they vary in height that the number of points changes.

The bow shrinks down to 2" at the waterline.

Any suggestions on how to correct?

Problem47.3dm (259.4 KB)
Here are the forward frames as I have them now.

I have more frames than I need I have not yet figured out which frames I actually need.

If I move the split between the bow back, I can get more waterlines but lose frames.