Maybe it is because I haven't finished my coffee yet but I think there is a problem here.
A rectangle, a smaller rectangle and I try to sweep the small using the big as a rail. It's working like I have 4 separate lines.

Hi JBT,

I'm not sure of the exact technical explanation,
but it seems the sweep doesn't have enough
control points to reference in order to negotiate
the corners. If you add another rail, it will
work as expected.

-Jesse

@jtb
It will better if you link also the 3dm file :)

@PILOU
It's just two rectangles, but anyway, here it is

Hi JTB - that is a limitation of sweep right now. It isn't able to miter the corners of a multi-segment rail curve like that. It is currently handled just as if each line was separate.

Eventually I want to make it work like you expected, but it is quite a bit of work to make it work reliably - I started to work on it when I was doing sweep but it was difficult enough that I decided to postpone it, I don't think I will be able to do it until after V1.

In the meantime, using 2 rails like Jesse shows is probably the best workaround. It's also possible to do it with your existing curves by using Rail Revolve.

- Michael

Yes Michael, the workaround with the two rails is no problem.
The solution with the rail revolve however is not a nice one because you have to think about the correct position for the axis of rotation. If you just pick a Z axis in a random position, you get results like this

I had the same problem with the sweep using a rounded curve as the rail. I tried to model an inhalator with some curvy grooves (see attachment) on the outside. I tried the following approaches:

- Use "trim" with an extruded curve to derive the contour of the groove along the object. Then I hoped to do a sweep along this curve with a circle. I wanted to subtract this shape from the object using the "boolean" function. However, the boolean did not work, probably because the contour was not properly closed. On rectangular objects I had the same problems as described in this thread.

- Then I used another approach, namely create a "shell" of the extruded curve (with the required thickness of the groove). Using booleans I cut a hole into this shell. Then I bevelled the inside of this hole, scaled the shell with the bevelled hole at 95% and subtracted this from the original shape.

The big downside of the second method is that the "scale" will not result in an equal depth of the groove along the object. Is there another method to create such a groove?

Yours sincerely,

Ray

(I added some screenshots that illustrate the methods described above. I hope it helps)

Hi Ray, I would probably go with the first method that you describe, and try to model the shape of the groove with a sweep and then cut that into place.

> However, the boolean did not work, probably because the contour was not properly closed.

This part I don't quite follow - why wasn't the contour closed? Can you post the model with this problem contour in it so I could check it out?

- Michael

Hi Michael,

I attached two files, the original inhalator before making the grooves and a test file that I used to see whether my problem was related to the complexity of the shape. In both cases the boolean between the "tube" and the base object does not result in the result I expected. I am curious if I made a mistake.

Yours sincerely,


Ray Deleu

Very design!
Added to the Special thread Gallery :)

Hi Ray - one thing is the outer surface of your inhaler-2n1.3dm file is duplicated - I mean if you select it and drag it away you can see there is another copy of it underneath.

I've seen duplicated surfaces like these a few times now and I think there may be a bug in something that generates them. If you have any idea of what might have caused this duplicate please let me know.

Other than that, it looks like you may be running into a problem where the boolean operations don't quite work as expected right now when one of the objects involved is an open surface instead of a closed solid.

You may need to use the Edit/Trim command instead of booleans when working with open surfaces.

So in this case, select the outside surface and the "groove tube". Run Edit/Trim, and at the "Select cutting objects prompt", just push "Done" - this signals that you are going to do a "mutual trim" operation where the selected objects are cutting each other. Then at the next prompt just push "Done" to leave all the pieces behind (since in this case some of the pieces are kind of obscured).

Then you can select and delete the half of the tube and the little strip of the outer surface, leaving you with your groove:

You can then use join to glue these pieces together.


> I am curious if I made a mistake.

I don't think so... I think maybe you got hit with the booleans-with-non-solid issue that is present right now.

Please let me know if you need any more details or if I have misunderstood anything.

- Michael

Just quicky try without any precision (just for the concept) but maybe that could be work :)

With more attention to the orientation and size of circles )
In fact that is not a verygood idea :D

Just a little question : When an arc is "extended" to a boundary object : it's an always an arc or a certain tangent line from the existant arc toward the boundary object?

Ps Very sinuous grooves Michael and tricky tip ! :)

Hello Michael,

Thanks for your reply. I think I understand your method but will try later today to see it for myself.

Originally the outer shell of my inhaler was closed. However, after the sinusoidal cut and bevel (rounding) I deleted the side surfaces were the inner part of the inhaler is located. Perhaps the boolean method would have worked if I had applied it earlier in my modelling process. I will try to remember to do booleans preferably with closed surfaces.

I can imagine that the double surfaces were created in the modelling process as sometimes I did not understand exactly what happened during a trim or boolean.

Yours sincerely,

Ray Deleu.

I added two renders of objects created with MoI, rendered with C4D 6+.

> I think I understand your method but will try later today to see it for myself.

Sounds good, please let me know if it doesn't seem to work right.


> I can imagine that the double surfaces were created in the modelling
> process as sometimes I did not understand exactly what happened
> during a trim or boolean.

I think that something may be duplicating things in an unexpected way. Anyway, if you happen to notice this happening during a modeling session, please let me know.


> I added two renders of objects created with MoI, rendered with C4D 6+.

Looks groovy! :)

- Michael

Other than that, it looks like you may be running into a problem where the boolean operations don't quite work as expected right now when one of the objects involved is an open surface instead of a closed solid.

Yes, I can confirm this. I was trying to make a threaded screw. I swept a circle along a spiral curve, then tried to subtract it from a cylinder. But, when I tried to do the boolean, it didnt work. I looked at the ends of my sweep, and they where open. I selected the ends and closed them with 'planar'..then the boolean worked perfectly.