Hey Guys,
I got a little question:
I´ve got this surface and I want to make a solid with athinkness of 3mm.
I tried everythin I know to get the second rounded surface and add the small one between with offset, scale, offset lines and profiles, sweep and network..
But I don´t know how I can get this solid.
I got the 3dm data for you.

I´m looking forward for your answers!
Have a good day!

Koi

Hi Koi,

Using the initial curves/surfaces did not work.
Delete the initial surface.

First step was a rebuild of one of the long curves.
Offset the new long curve inwards by 3 mm
Revolve the two curves by 90 degrees.
Copy both of the new edges to make two curves there.
Place 3mm line at vertex and sweep each pair of inner and outer curves. (twice)
Join the 4 surfaces to solid.
Boolean yields only a joined surface, during one trial.

(Planar with two 3mm lines, could be used instead of sweep. Boolean does not give solid, but Join does.)

- Brian

Some of the initial geometry must have been off by a bit?


Another permutation is to rebuild one of the long curves, and revolve it 90 degrees.
Shell the resulting surface 3mm inward.
Perform planar twice, and Boolean Union the 4 surfaces to a solid. (Or use join.)
In this case, the Boolean Union does make a solid.

There is some bad geometry about 1.3 cm from each end of your bottom rounded curve.
See the multitude of points as well. Rebuild appears to clean this up?

- Brian

hI

I agree with Brian

You can rebuild with GORDON function ...
and make a new SHELL

here are 2 Gif Files

Chri

Hi Koi, sweep is having some difficulty making a proper shape near the tip where the 2 rails come together, there is a little fold in the surface there which is messing up offsetting the surface.

If you use Construct > Network to make the surface instead of sweep then it can be thickened using Shell.

- Michael

Hi Koi - actually since you wan the cross section to have the shape of a circular arc it would probably be better to form this shape by using just one of the rail curves and doing a Revolve on it. You can set the revolve angle to 90 degrees before exiting the Revolve command.

- Michael

Hey guys,
thank you so much for your answers!!
Today was so much wirk, I will try it again with your tips tomorrow.
I did not try do build it with the revolve-tool :)
I will tell you,
Thanks a lot

Koi

He guys, thanks so much, i did it!
that was quiet easy... I thought I´m not a nooby anymore...
Thanks again and have a nice day

Hey Brian,
thanks for your answer aswell! I tried this version today, works also :) I just think the way of michael is better, because I did not use exactly 90°..., so in other cases I can´t use revole I think?

Cheers
koi

I´m sorry, I got the next question :D

Now I tried a similiar solid: I just changed the profile curve more complexe. When I use shell, the surfaces on the bottom and left side are not flat anymore (2nd object). When I use b.diff. with a line(redline), the solid got a strange surface/ mistake on the side( 3rd object).
Is there any way o get a clean result?
Have a nice weekend,
koi

Trim with line can remove one sliver to make a flatish area.
Boolean Diff with plane can remove the other sliver, to make a flatish area.

But, near the inside corners, there are still tiny "folded over" edges.

- Brian

Hi koi, so for the bottom not being flat that's partly because your profile curve does not have an ending tangent that points straight downwards, it's angled like this:


So therefore the surface normal that a surface offset will go along will be like this:




I'm surprised that shell actually worked here because the surface you've generated is very strained and badly formed in the area where it collapses down to a point:




I'm not sure what your goal is for this protruding feature as it traverses towards the tip:



But you probably can't get a well formed result by just having that protruding feature directly collapse down to a point all on its own. You'll probably need some additional stations set up so that there isn't a complex shape with near undercuts in it as it collapses down. Something like this using Construct > Network (see attached 3dm) :





And it may be better to directly model the inner surface with its own set of curves rather than generating it with a surface offset.

- MIchael

Hey Brian,
thanks for your answer.
Ißm not sure, if I understood completely right.. I tried to trim and to b.diff, but in both cases there remains a thin surface, like you can see in one of the fotos..
Do I understand you right?
-Koi

Hey Michael,
thanks for this big answer!
"So therefore the surface normal that a surface offset will go along will be like this"
Ok, I understand that, but I need the small surfaces straight. This object should be a rain hood for bull's eye. so I need a straight side to be able to glue it down, I´ve got a picture for you. I want to build it with a 3D printer.
Is tehre any possiblity for that?

"And it may be better to directly model the inner surface with its own set of curves rather than generating it with a surface offset."

I´m not sure what you mean.. Actually I built that solid ffrom a surface with ´shell´. Firstly I tried to built it with profiles and networks. When I joined them in the end, I just got a joined surface every time..

On the other pictures, I tried to trick out the normals, but that did´nt work :D

I´m looking forward for your answer:)

-Koi

I think that "surface offset" equals "shell".

One technique is to make the object a little bit longer, and trim off excess. But first the edge fold (surface fold), must be prevented.

Zooming in, it seems that the shell is causing an edge to fold back on itself, so avoiding shell would seem to be in order.

-Brian

ps, I Like your design.

I did move the trim plane by a tenth of a mm, or so.

Hi Koi,

re:
> I´m not sure what you mean.. Actually I built that solid ffrom a surface with ´shell´. Firstly I tried
> to built it with profiles and networks. When I joined them in the end, I just got a joined surface
> every time..

The way Shell works for thickening an open surface is it builds a surface offset and then connects the edges.

Instead of using a surface offset method (including shell) it could be better to draw a second set of curves for the inner surface, creating it in the same way the outer surface was created.

- Michael

Hi Koi, here's a version built using Network twice, once for the outer surface and then a second network for the inner surface.

- Michael

Hey Brian and Michael,
I thank you so much! I love to learn how MoI and Nurbs are going!
I got your points.
I tried to make it solid with network aswell, but this was not possible, because I did not take a second profile on the sides to prevent overlapping?
Thanks so much, thid forum makes it defintely bigger!

thanks a lot,
Koi

Hey guys,
it´s me again with a new question :)
I build a beautiful objetct with the sweep tool. And made this solid.
On the object is like a small seam to see, i made a screenshot aswell. Is it possible to get a nearly same result without that?

Best regards,

Koi

P.S. I am still working with V4 and, as you can see, not a real pro. Do you think I should switch on v5?

Hi Koi, the problem is basically trying to do too much all in one single surfacing operation. If you try to force a sweep to go through too many shape changes it can result in a stressed surface that can have warping in the surface as it sort of tries to simultaneously pivot and stay squished together.

One method that might be good would be to use a one rail sweep like this: (with Twist = Flat option set)



That's going to make a more relaxed surface and then the top can be formed by a separate surface with a fillet or blend to transition between them:




Compare the surface control points between this sweep and your previous one, note how there is a lot heavier points and some strained areas where things are kind of bunching up. When you get the kind of "seam" like you were seeing it's sort of like too many forces applying stretching, shearing and rotation all at once in a small area and making little bumps.





- Michael