# Effective modelling of polar array

What would be the best way to achieve a quad mesh model of this kind of surface:

The object has 12 circular arrays, with 6 holes on the innermost ring, then 12, 18, 24 ... plus 6 per each ring. Please note that the hole diameter is constant.

Modeling each ring with an polar array is possible, but this results in a heavy workload of retopo. I also tried the tissue addon, but could not manage to get clean meshes at a glance.

## 2 Answers

If you run this baby script:

import bpy

idx = 1
vtx = 6
rad = 0.25
inc = 0.25

while (idx < 14):
bpy.ops.mesh.primitive_circle_add(vertices=vtx, radius=rad)
rad += inc
vtx += 6
idx += 1


It will create a series of concentric circles with the right number of vertices. (I'm sure a proper Pythonista could do it in one line, and automate some of the steps which follow...)

• CtrlJ Join the circles into one object
• CtrlE > Bridge Edge Loops each pair of loops in turn, and F..AltP, fan-fill the center.
• You wind up with a mesh with one or two diagonals out of place. I'm not sure it would matter, but I cut 1/6 out as shown, corrected the edges, and arrayed it back to a full circle)
• Select all vertices except those on the perimeter, and CtrlShiftB bevel them, and switch to Face mode. The future holes should be selected.
• Loop-Tools or Space-Bar > 'Circle' with a radius setting, on the selected hole-faces.
• Delete the faces
• Extrude or add a Solidify modifier, and throw on a Subdivision modifier.

The holes aren't perfectly circular but maybe good enough?

• Hmm, it's not quad... but it is after one level of subdiv.... Nov 26 '18 at 0:40
• Reminds me of the stretching and Shirley mappings blender.stackexchange.com/a/93180/15543 May 30 '20 at 8:17
• @batFINGER Could that have been a way through, I wonder.. I wasn't very Pythonic in those days, was I? Only a little more now, to tell the truth.. May 30 '20 at 8:33

Thanks to all of you for your support. After some thinking, i came up with this method:

[

Doing this manually is not the fastest method, but i think in terms of topology this is the best result. Maybe someone can do a script to perform the steps.

• Very nice... :) I can't help thinking there's a neater way than both of these... Nov 26 '18 at 19:23