Hi Olga,

> I need at least 2 point thickness (millimeters)

Yeh, 2mm thickness seems extreme for shelling considering the size and shapes involved with this piece.
I gave it a go in another app that is very strong in shelling and the best I could get was 0.5mm,
so what you are asking for is like a type of rough shelling/coring out operation, I think that would be a tough one.
In this situation you could try to hollow the back and reduce the weight by boolean difference with a combo of different shaped solids.



Cheers

Thanks Danny
I appreciate your suggestions. It will try to do it by boolean difference
If the shelling works for 0.5 mm for this surface, can’t it make “smart shelling” – instead of creating a thin wall – it would simply solidify the areas where thicker wall (like 1.5 mm) interfere with each other.

Another way I tried it:
I blew up the surface in 4 pieces, then shelled each and tried to join / boolen/union them.
It did not work.
Could you try that approach please?

Thanks
Olga

an another approch will be that
Use plan whith the bottom loop contour ouline for close the solid
then Scale copy as you want (even 0.999)
Make the good alignment and make boolean difference
With that I believe you can have in theory 0.001 unity difference (max of precision of Moi ;)

Hi Olga,
I might not understand what you want to achieve, but thickening up all those surfaces by 1.5-2mm almost comes
back to the boolean suggestion that I recommended.
This is shelling at 2mm but without the sphere and cone detail, did you want 1.5-2mm on all faces ?



Cheers

Danny
It does come back to the boolean suggestion you recommended but it is more precise because the thickness of the wall remains constant

If you do the same with sphere and cone and join/boolean union them together to create one solid - will it work? It does not for me - I might be doing something wrong

-Thanks - Olga -

Hi Olga,

> Is it possible to produce something like "smart shelling" - to give
> some thickness to this or similar surface?
> So that planes that interfere with each other would
> merge and create one solid.

It's possible in theory, but unfortunately it is a very difficult task.

When you start using "smart" to describe an operation it can often be something that is very difficult to translate into a software algorithm.

Software unfortunately is totally not "smart" - it only does exactly what it is told and it is very difficult to translate things like human judgement into the kind of explicit instructions that are required for program code...

- Michael

Hi Olga,

> If you do the same with sphere and cone and join/boolean union
> them together to create one solid - will it work? It does not for me -
> I might be doing something wrong

Do you have the .3dm model file for this that you can post please?

It's pretty difficult to follow along just from the descriptions alone.

This may be something that will confuse the booleans and you may need to switch to a more surface style of modeling - that's where you use Edit/Separate to break solids down into individual surfaces and work on trimming surfaces one at a time to produce what you need.

It's kind of a more "low level" approach than the boolean type operations, but it can let you make progress if you are having problems with the booleans.

Similarly you may be able to make some progress by breaking things into individual surfaces and then using Offset to build some pieces to work off of to do further trimming, etc..., instead of trying to use Shell which will get confused by the sorts of overlapping pieces.

- Michael

Michael,

> Do you have the .3dm model file for this that you can post please?

That's the one in the first post.

Cheers

> That's the one in the first post.

I guess it sounded like Olga had done some other steps to the sphere and cone and was having a problem with booleaning after those steps?

If so then it could help to see that result more specifically...

- Michael

Hi Michael and Olga,

> I guess it sounded like Olga had done some other steps to the sphere and cone and was having a problem with booleaning after those steps?

That's the bit I don't get, if you do a offset/shell at 1.5-2mm with this piece the main walls of the triangle piece would engulf the offset of the
sphere and cone so the offset of those features wouldn't be necessary, or I am missing something here ?

Cheers

This is what I mean, I took a section and offset the curves at 1.5mm (highlighted)
the result would be a small pocket in the center of the piece as shown in between the red lines
because the sphere and cone offset would be trimmed to the bottom of the piece, is this correct
in what I'm saying ?



Cheers

Hi Danny, yeah I think that is correct. I'm just not quite following along so well with some of the later stages being described there, having a model to look at to go along with a description is always helpful to clear things up.

- Michael

Hi Michael hi Danny

I attach the file where I tried to shell the surface of my design for a pendant
I could not shell the whole thing so I blew up the surface and shelled it in parts but could not join them together and remove extras (I tried Boolean union, merge)
I did not shell two small top corners so this has to be fixed also

Please take a look – you might have an idea how to merge them together and remove extras

Here the thickness is 1 mm.

-Olga-

Hi Olga,

It's a bit of work, spent about 20 min on it but it was a combination of offsets, untrims, trims, booleans, separations and a front end labotomy ;)
If you really, really need a step by step explanation (please say no) ;) then I'll need some time.

It's basically what Michael said earlier Here about 'low level' approach.
I was going to give up after 10 min but it got the better of me so I had to finish it.

Cheers

Hi Danny
No I do not a step by step explanation
That looks exactly I want to achieve.
Did you use my latest file or the earlier one with only the basic surface with no thickness?

-Thanks - Olga-

Hi Olga,

> Did you use my latest file or the earlier one with only the basic surface with no thickness?

this one 0473-9.3dm

Cheers