Hello...

I cant shell this surface even 0.01 mm.
Any advice ?

Thanks !

Hi Anis,

I reconstructed your model.
Then applied shell and did fillet afterwards.

Hi Anis,
As Paolo mentions,

Shell in MoI works better on more simple geometry, so it's better to plan your modeling steps when shell comes into the picture.
You should try and apply small details like, fillets and small ledges as shown in the picture, after the shelling operation as these areas tend to be difficult for shell to calculate.



I've also remodeled your part, using solid cylinders and preforming the shelling operation before the small details and it shells nicely, at a range of thicknesses.
You will find the steps in the .3dm file attached.

Cheers

Hi Paolo and Danny,

The problem fillet after shelling is we have to apply fillet in two steps.
First for outside and the second one for inside.

But anyway your tips working well..... :)

Thanks !

Hi Anis,

Do you want the fillet on the inside to be
in proportion of the outside?
I don't think it is necessary.

Hi Anis - was this something that you modeled in MoI, or was it some imported geometry?

Somewhere along the line some of the edges in the bottom area of your model got split into 2 edges rather than one long single edge:



It is those split edges that seem to be difficult for Shell or Offset to handle, which seems to be a bug. But if you had long single edges there instead of ones split into several fragmented edges, it would work.

I'll see about adding in a "Merge" utility function for the next v2 beta for gluing fragmented edges like this into longer single edges, that will be useful for repairing stuff like this.

- Michael

Hi Michael....

See video below :

Hi Anis, thanks for the video attachment!

But that's actually not the area of the model that has the split-up edges though, they are in this area:



Also one note on the technique you are using there in the video - although that works, you may want to try and avoid that method when possible since it creates a surface that has a collapsed down "pole" point in it - you can turn on the surface control points to see that.

Surfaces that collapse down to a pole can sometimes cause problems, they tend to need special processing for things like determining the surface normal at the pole point.

It can also work fine, it is just something that can be a good idea to avoid if possible since they are more complex. Generally a model that contains a smaller amount of "pole" areas like that is better quality.

In the kind of case that you are showing, there is a different way to construct it without any pole, which is to make a solid and do a boolean to cut it with an angled profile line, like this:





If you turn on control points for this method, you can see that the surface is instead a very simple plane, with trim curves on it, rather than one that collapses down to a pole point:



Basically the booleans will make an extrusion of the profile cutting curve and that extruded surface is intersected with the main shape to form that cap with this method - such extruded surfaces are more simple.

- Michael

Hi Michael....

>> But that's actually not the area of the model that has the split-up edges though, they are in this area:
I split the surface because I need to use loft to close the top area.
I use loft because the area is not planar. You can check on my red curve.

Thanks

Hi Anis,

> I use loft because the area is not planar. You can check on my red curve.

I see - but actually the exact same advice I mentioned previously still applies.

Just use the red curve as the cutting object the same way I used the line in my previous example.

Any time you want to make a cut that follows a 2D profile you can use that method - for example here I have exaggerated the curvature just to clarify that the end result does not have to be planar:





Select the solid and run boolean difference, and select the profile curve as the cutting object to get this result:



Just discard the pieces you do not need:



In this case you don't get a plane as the result surface, but the result is similar in that there is no pole. Instead it is an extruded surface (with a line as the cutting curve, it will be a plane since the extrusion of a line is a plane), here are what the control points look like when turned on:




Not only can you avoid getting a surface with a pole with this method, it is generally faster as well - just cap off your base shape so it is a solid, then you can carve off chunks by doing a boolean with a profile curve as shown here.

When you start with a solid, then the boolean cuts will generate solid pieces with the curves leaving extrusions behind to form the caps.

- Michael

Ohhh, my fool Micahel.... :(
Ya, I see know. So to create 3D model we not have to use surface from beginning.

Thanks Sir !!!

Hi Anis, yup that is correct! :)

When your end result is made up of mostly extruded shapes, it can be easier and faster to work more with solids throughout rather than doing things by one individual surface at a time.

That's due to that mechanism of doing a boolean between a solid and a curve which generates solid results...


When your end result is made up of more freeform surfaces (that may swoop around in multiple directions instead of curving only in one direction) and not so many extrusions, that's generally when you may need to use a more "one surface at a time" type method.


- Michael