Michael,

thanks for the comments! Just as usual, old habits never die (easy) So I started wayyy to complicated by modeling like I usually do.

And yes, almost there, but then time ran out :cry Ah well, maybe I´ll revisit it and finish it one day.

What kind of really bugged me was that I could not figure out how to weld the various surfaces at the edges, leading to a bad polygon flow and of course an open mesh on export. Any tips on how to do that properly?

Thanks,
Michael

Hi Michael,

> What kind of really bugged me was that I could not figure out
> how to weld the various surfaces at the edges, leading to a bad
> polygon flow and of course an open mesh on export. Any tips
> on how to do that properly?

You would normally use the Edit > Join command to do that - that can take surfaces that are touching each along an edge to get joined into a larger structure that has a single common edge between the 2 surfaces.

Then when meshing those joined edges will be treated specially so that they get a common vertex structure in those areas making a "watertight" mesh.

But you've got to have surface edges that match up to within 0.005 units of one another in order for them to be joinable.

In some areas you've got surfaces kind of sticking past one another like here:



That piece sticking out there would need to be trimmed off first before those could be joined.

Also if you're using Network in this kind of "patch by patch" construction process you can additionally run into a problem where Network uses a somewhat looser tolerance than some other commands in order to make a surface result that is not overly dense with control points. But it can sometimes just barely sag away by more than the 0.005 join tolerance. In those cases with this particular Network problem, you can work around the problem by scaling the objects you want to join by 1/10 in size (which also scales down the gap between surfaces), then do the Join, and the scale back up.

This is all quite a bit easier if you work more with solids from the start rather than doing things at this kind of individual surface modeling level which is kind of a more advanced (and also finicky) approach.

If you keep things as solids and carve pieces off of the solids with boolean functions, then you don't need to join things together because everything stays as an already joined solid at every step.

So for example if you watched the video tutorials here:
http://moi3d.com/2.0/docs/tutorials.htm , notice that they work by creating some different solid pieces and then combining them with booleans - then there is no joining step in that kind of workflow.

Your particular project being a helmet made of something like hammered and contoured sheet metal is getting kind of somewhat more in the "organic modeling" territory than something that would be machined which is where the NURBS solids tools become much more of a strength.

- Michael

Hi Michael, so in your case it looks like you would need to use the scale down + join + scale back up case to get it to join well.

There are a couple of things like that extending spot I showed earlier that are not ready to join, but other than that these parts here can be joined:



There is a little slivery hole right here though where there seems to be a missing surface:




But anyway, I've attached the results after joining, and also an OBJ file that was created from the joined result, you should be able to see now that there aren't any cracks in the mesh anymore (aside from that sliver above).

- Michael

Hi Michael,

thanks a lot !

I somehow must have messed up on my way. I actually tried always to snap the end points of my curves on the respective other curves to make the surfaces closed and defined. That was where I tried to aim for in my initial question, since once I have a complex network set up, and skinned the thing, but choose to move one curve or curve point, I very likely loose the "watertightness" of the model. I never figured out how to keep all surface curves aligned at once, so it was a constant hopping around and checking if there are holes anywhere.

I´ll just download the no-save trial version and see if i can reproduce your advice.

Thanks again for the OBJ-files, I´ll check them in Cinema4D!

Michael