A bad case of shingles

[Territan]

Or rather, scales. (Thread title too click-baity? I guess I could change it...)

Scales occur in nature. They also occur in armor, and in this case they occur in Cheetah when I create a single scale shape and seek to shingle said simple shape on a surface. And I'm having a few issues. And I'm certain I'm doing it wrong.

For the first attempt, I created a single shingle/scale, put an Array modifier on/in it, duped it, and offset it to create a layer just below.

Then I stuck an array object in the group because I wanted several rows of those scales, not just two. And that array modifier did nothing.

The second attempt used a ball in the approximate shape I wanted, then I set a Particle Mesh to conform to that, copied over a single scale, put the Particle tag on that, and put it inside the Mesh. That worked...

...mostly. But at the bottom, something different happened:

Affixing the shingles to the Edges of the Particle Mesh worked to get the shingle pattern I wanted on the main surface, but at the bottom of the object, the shingles are rotated 90° to where I wanted them, which is kind of to be expected.

What am I missing, and is there a better way to manage what I'm trying to do here?

 

[frank beckmann]

I'm pretty sure I'd use the particle system for it. But you didn't mentioned the shape it will be applied to.

 

[Territan]

I'm pretty sure I'd use the particle system for it. But you didn't mentioned the shape it will be applied to.

The shape I'm applying to is spherical, as I intended it to at least suggest the top of a space station. Fortunately, I have enough for what I need to do; I just need enough of an image to suggest the scales and curvature. But I want to get a handle on "best practices" for next time.

Also, for what it's worth, I was scraping screenshots of the various elements to show what I was doing, but then I got a clue and realized that the Cheetah file itself is smaller. Essentially one material and three objects.

 

[ZooHead]

Frank is probably right, but you can do it with a stack of modifiers.
But it's a real pain.

 

[Helmut]

* I guess that spherical shapes may be a bit tricky.
* In my experiment I started with an icosahedron, subdivided by Stam-Loop. The rest, as suggested by Frank B, is a particle mesh based on this icosahedron.
* Needless to say, you get a differently meshed geometry of tiles on an icosahedron. As an advantage, these tiles are evenly distributed on the surface.

* Of course, you could do a proper job on a sphere, but this would require applying tailor made shingles to subsequent "rings" of the sphere. Using a particle mesh on polys you can define polygon selections, so whilst this is no problem it is a bit time consuming.
* In this case you will start with triangular shingles on the poles which gradually get squarish towards the equator.
The omniscient Wikipedia tells me that applying the silica tiles to the space shuttles took 2 years for each flight, so there is no hurry.

 

[podperson]

Spheres will run into the coconut combing theorem (you can’t comb a coconut).