Equilateral convex polyhedra (new class after 400 years?)

The place to talk about Stella4D, Great Stella, and Small Stella. Feel free to ask questions about them here.
User avatar
robertw
Site Admin
Posts: 676
Joined: Thu Jan 10, 2008 6:47 am
Location: Melbourne, Australia
Contact:

Post by robertw » Wed Nov 05, 2014 1:35 pm

Certainly some polyhedra can't be given equal edge lengths unless they lose their convexity, like many of the duals of the Archimedean solids.

Here's some more convex equilaterals though.

From Stella's near miss library, these two don't need any further variation. I even named them "Equal-edged Near Miss", so that's more equilaterals that were known before that paper. These were ones I found, as convex hulls of new Stewart Toroids.
Image Image

Other near misses in Stella's library can be adjusted for equal edge lengths, such as these near misses discovered by Robert Austin:
Image Image

User avatar
guy
Posts: 86
Joined: Mon Feb 11, 2008 10:30 am
Location: England
Contact:

Post by guy » Thu Nov 06, 2014 1:50 pm

robertw wrote:Certainly some polyhedra can't be given equal edge lengths unless they lose their convexity, like many of the duals of the Archimedean solids.
Yes of course, all those funny groupings of triangles round a vertex for a start.

Something like a heptagonal pyramid requires its equilateral morph to be wound up to give it a star base. That can be done with either a {7/2} or {7/3} base, giving two distinct geometric solutions.

I suspect that those Archimedean duals would also yield two solutions, one with augmentations and the other with excavations.

What could Stella's spring algorithm make of all that?

polymorphic
Posts: 26
Joined: Sun Feb 17, 2008 12:50 pm
Contact:

Post by polymorphic » Thu Nov 06, 2014 11:41 pm

I haven't been able to make one with the rhombic enneacontahedron truncated on the 5-valance vertices. It works on the 3 and 6 valance vertices, but on the 5 it collapses into non-convex.

Roger

User avatar
robertw
Site Admin
Posts: 676
Joined: Thu Jan 10, 2008 6:47 am
Location: Melbourne, Australia
Contact:

Post by robertw » Fri Nov 07, 2014 12:21 pm

polymorphic wrote:I haven't been able to make one with the rhombic enneacontahedron truncated on the 5-valance vertices. It works on the 3 and 6 valance vertices, but on the 5 it collapses into non-convex.
I agree, I don't think that one works.

Here's a really simple one I just found:
Image

polymorphic
Posts: 26
Joined: Sun Feb 17, 2008 12:50 pm
Contact:

Post by polymorphic » Fri Nov 07, 2014 1:37 pm

There are many that can't work. For example if you augment a cube with cubes it will be non-convex and is already equilateral.

That models seems familiar. How did you create it?

User avatar
robertw
Site Admin
Posts: 676
Joined: Thu Jan 10, 2008 6:47 am
Location: Melbourne, Australia
Contact:

Post by robertw » Sat Nov 08, 2014 11:25 am

polymorphic wrote:There are many that can't work. For example if you augment a cube with cubes it will be non-convex and is already equilateral.
Or even if you put a pyramid on each side of a cube.
That models seems familiar. How did you create it?
Weirdly, I started with "6J91 (P4)" which is in Stella under "Stella Library->More Stella Toroids->6J91 (P4)" (it's not a toroid but the shape of a hole in one of the toroids), and took the convex hull, then adjusted for equal edge lengths. But I'm sure there's a more straight forward way.

User avatar
robertw
Site Admin
Posts: 676
Joined: Thu Jan 10, 2008 6:47 am
Location: Melbourne, Australia
Contact:

Post by robertw » Sat Nov 08, 2014 1:26 pm

robertw wrote:But I'm sure there's a more straight forward way.
Actually, it's a rectified truncated octahedron I suppose.

User avatar
robertw
Site Admin
Posts: 676
Joined: Thu Jan 10, 2008 6:47 am
Location: Melbourne, Australia
Contact:

Post by robertw » Sat Nov 08, 2014 1:33 pm

Here's a bunch more. These are all convex hulls of new Stewart toroids that were discovered by Alex Doskey and myself, and which are all available in Stella. No adjustment required as the edges are already the same length due to the nature of Stewart toroids.

Image Image Image Image Image Image Image

polymorphic
Posts: 26
Joined: Sun Feb 17, 2008 12:50 pm
Contact:

Post by polymorphic » Sun Nov 09, 2014 4:45 am

robertw wrote:
robertw wrote:But I'm sure there's a more straight forward way.
Actually, it's a rectified truncated octahedron I suppose.
It can be found by truncating the on 3-valance vertices of the truncated octahedron by a factor of 1/2.

It can also be found by beveling only on the 3-valance vertices in conway notation which *would* translate to this operation string. (Beveling itself would not allows the three 'b3').

dk3datO

Conway Notation specification does not specify that beveling can only be done by vertex valances. I think this may have been an oversight! Beveling is actually truncation by 1/2. Therefore since Conway Notation allows for truncation by vertex valance, so should it be with Beveling. I am going to propose this change.

polymorphic
Posts: 26
Joined: Sun Feb 17, 2008 12:50 pm
Contact:

Post by polymorphic » Sun Nov 09, 2014 3:31 pm

polymorphic wrote: Conway Notation specification does not specify that beveling can only be done by vertex valances. I think this may have been an oversight! Beveling is actually truncation by 1/2. Therefore since Conway Notation allows for truncation by vertex valance, so should it be with Beveling. I am going to propose this change.
I experimented with this and the only difference is how much is truncated from the polyhedron before planarization occurs. Normal truncation is only 1/3 of the way. Beveling is 1/2 of the way. So the results of tN and bN are ultimately the same.

The interesting polyhedron can actually already be generate by ambo of the truncated Octahedron.

conway atO | antiview

Roger

User avatar
robertw
Site Admin
Posts: 676
Joined: Thu Jan 10, 2008 6:47 am
Location: Melbourne, Australia
Contact:

Post by robertw » Mon Nov 24, 2014 1:32 am

polymorphic wrote:Conway Notation specification does not specify that beveling can only be done by vertex valances. I think this may have been an oversight! Beveling is actually truncation by 1/2. Therefore since Conway Notation allows for truncation by vertex valance, so should it be with Beveling. I am going to propose this change.
Are you confusing bevelling with rectification (ambo)? Rectification ("a" operator) is truncating half way so that the neighbouring truncations meet. Bevelling is not just a deeper truncation. It is truncating both vertices and edges. If you limited the set of vertices to be truncated, you'd then also have to decide which edges to truncate.

Rob.

Gerry in Quebec
Posts: 2
Joined: Tue Sep 02, 2008 2:52 pm

Goldberg polyhedron

Post by Gerry in Quebec » Sat Aug 29, 2015 8:34 pm

Rob W. wrote:
Finally, here's a couple of Goldberg polyhedra I made using this new feature, to give them equal edge lengths.

Hi Rob & others:
Great discussion about equilateral Goldberg polyhedra. I've experimented with these on and off in recent years. Here's a SketchUp model (by Taff Goch... from 2010 as I recall), showing the I {3,0} polyhedron.

- Gerry in Quebec

Gerry in Quebec
Posts: 2
Joined: Tue Sep 02, 2008 2:52 pm

Re: Goldberg polyhedron

Post by Gerry in Quebec » Sat Aug 29, 2015 8:38 pm

Let me try posting this again, this time with the link to the 3D warehouse:

https://3dwarehouse.sketchup.com/model. ... 32e8ebf2b1


Hi Rob & others:
Great discussion about equilateral Goldberg polyhedra. I've experimented with these on and off in recent years. Here's a SketchUp model (by Taff Goch... from 2010 as I recall), showing the I {3,0} polyhedron.

- Gerry in Quebec[/quote]

Post Reply