2
$\begingroup$

I am trying to use the array modifier. When I use the object in the first image the result is a closed shape but it creates unwanted faces between the copies.

enter image description here

When I use the object in the second image the final result doesn't have a closed perimeter.

enter image description here

enter image description here

The use of caps is easily applicable in one dimension array. In a 2-dimension array things become complicated. Is there an effective and not over-complicated solution?

Improve this question
$\endgroup$
8
  • $\begingroup$ what exactly is your problem in the first case? because you know they are there? Do you have texturing problems? $\endgroup$
    – Chris
    Commented Feb 8, 2022 at 12:06
  • 1
    $\begingroup$ I don't think there's any easy solution, is it a problem if at the end you apply the Array and fill the sides? $\endgroup$
    – moonboots
    Commented Feb 8, 2022 at 12:11
  • $\begingroup$ It is for 3d printing and the geometry should be manifold without internal faces. $\endgroup$
    – artos
    Commented Feb 8, 2022 at 12:12
  • $\begingroup$ @moonboots I want it to be non destructive. The size of the object will be variable. $\endgroup$
    – artos
    Commented Feb 8, 2022 at 12:17
  • $\begingroup$ Isn’t there an option in the array modifier that lets you add a cap object or something? $\endgroup$
    – TheLabCat
    Commented Feb 8, 2022 at 12:18

3 Answers 3

Reset to default
2
$\begingroup$

For the geometry you're showing, aiming for something non-destructive but simple, I would make the side walls with a rim-only solidify, and then weld the two copies of the sidewall that would be created by that solidify:

enter image description here

Showing 0 vertices non-manifold after convert to mesh.

For a small trade-off of non-destructive vs simple, you can drive the thickness of the solidify modifier from the distance between vertex-parented empties.

For different geometry, with normals that weren't at right angles to the main plane at the margins, this technique would still be possible, but more complex:

enter image description here

We do a miniscule solidify (because normals depend on non-zero-area-faces) with assignment to a VG, then modify our rim normals to point properly so we can do a displace from custom normals, weld (which will set some of our rim verts to weight 0.5) then finally mask away the unmerged rim verts.

The normals are measured from empty to origin, so the empty should be at the origin except for the Z axis, and should be parented to the object.

Improve this answer
$\endgroup$
1
  • $\begingroup$ I want a more universal solution, that can be applied with more complex shapes. $\endgroup$
    – artos
    Commented Feb 9, 2022 at 16:13
1
$\begingroup$

i made just one side (the other can be made accordingly)

enter image description here

result:

enter image description here

Improve this answer
$\endgroup$
0
$\begingroup$
  1. Start with a cube, rescale it if necessary
  2. Subdivide its top face into squares with loopcut
  3. Select the square faces, resulting from step 2, you want to extrude and extrude them down.

Done!
enter image description here

enter image description here
I removed one face to show no internal faces were created. You need to calculate the number of sub-squares to match the size and number of the impressions, but it still is simpler than others presented here.

If the above is still too much fuss for you then use array, your top image, with Merge option checked select all internal (or non-manifold) faces by hand and delete them with Faces Only option.

Regarding your comment: use array as in your second and third image, extrude the edges of the mesh arrayed and close it with the bottom face.

Improve this answer
$\endgroup$
2
  • $\begingroup$ Thank you, for effort. I know how to model a simple shape like this I want a more universal solution, that can be applied with more complex shapes and it has to be scalable. This was a simple object for reference. $\endgroup$
    – artos
    Commented Feb 9, 2022 at 16:16
  • $\begingroup$ @artos: No problemo, some speed modeling did the job. $\endgroup$
    – Lukasz-40sth
    Commented Feb 9, 2022 at 18:05

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .