The Autonomous House - Pictures and Details


A Note About Dome Geometry and Frequencies

Throughout this page, and elsewhere, I make mention of dome "frequencies". This is merely a convenient way of describing the geometry of the dome, and the type of divisions the triangles make upon its surface. Most domes nowadays are based on an icosahedron, or 20-sided polyhedron. Each of the 20 triangles is considered a "1-frequency" face (the distance between any two pentagon vertexes has only one segment.) The icosahedron has only pentagons surrounding each vertex, but as frequency increases, you begin to see triangles and hexagons filling in around those pentagons. On any full-sphere geodesic, it will always have 12 pentagons, regardless of the size or number of triangles present.

In the picture below, you will see 5 polyhedra, each becoming more and more spherical. I have kept the orientation of each sphere constant, and increased the frequency of each one. The first is the "1-frequency" face, and has one segment joining the two pentagon centers. The "2-frequency" has 2 segments joining those same pentagon centers in the same locations as the 1-frequency, and so on. Study the picture carefully, and you'll see where the pentagons are on each sphere. The pentagons get smaller as the frequency increases, but they're always there, and help the dome to curve back in on itself. (It it was all just hexagons, it would extend out in a flat plane.)

As the frequency increases, a dome becomes stronger, since it has more structural members to distribute the load across, and can thus handle stresses more efficiently.

Picture showing frequencies 1-5 of geodesic spheres.

Pictures and Details

First off, I'm envisioning a full-sphere house, not a partial sphere dome, as most domes are. Full sphere domes are much stronger, and the structural integrity will be needed when we pick these houses up and move them around. Actually, I'm thinking of a flattened ellipsoid, since I want the house to be about 16 meters (~50 ft) wide, but I don't need it to be 50ft tall! Something like this, actually:

Picture of a 4-frequency geodesic sphere, flattened out a bit.

It will have at least one deck around it, and I'm toying with the idea of two, but I'm not sure about that. Since these are designed to be manufactured in vast quantities, I'm sure that could be an option, just as cars come in 2- and 4-door models. (Click this one for a larger view.)

Picture of the same 4-frequency geodesic, now with a deck ringing it. 
Looks a bit like the planet Saturn, actually. :-)

Hub Designs

One of the tricks, of course, is how to get all those pieces to stay together and be easy to assemble. (Remember, think *BIG*, According to my calculations, we could build 10,000,000 of these per year, every year, for forty years, and still not catch up with the global homelessness problem. So think several orders of magnitude larger than you're used to.) A house like this (4-frequency) would have 320 face panels, and hundreds each of hubs and struts. Multiply this by 10,000,000 and you begin to see that we're talking about manufacturing (and assembling!) tens of *billions* of components.

There's a section of discussion we had about this in the Domesteading Archives which goes into more detail about the rather staggering scale of this whole project.

In fact, I suspect that we're going to have to apply an open source model to the production of these things, but in a hardware sense, rather than software. We'll probably need to make sure the designs are standardized, and freely available to all manufacturers, everywhere, so that we can be assured of high quality and interoperability between all the parts. Thus, we'll be able to take hubs made in China, combined with struts made in India, and aerogel panels made in the USA, to combine with leg sections made in Europe, and still get a perfectly working and high-quality construct.

This will also allow for the best economies of scale, as well as lowest possible pricing (if there are 50 different manufacturers of a single type of part, the market forces will assure that there is little possibility of price-fixing. And even if there were price-fixing, since the designs are open source, other companies could come in at any time to compete.

To this end, I've been working on a design for a sort of "universal hub", which could be used all over the dome, taking into account the varying angles at different points. Actually, we need two different hubs, one for 5-pointed intersections, and one for 6-pointed ones. I saw a thread about this on the DomeHome email list, and thus decided to publish this page, so others could see my designs, and discuss/improve upon them.

In fact, the renderings here have generated a lot of discussion on the DomeHome list, as well as on the Domesteading list, and in personal email. There have been some great ideas, and I'm sure there will be more. Here are a few things that have arisen, in case you are wondering some of the same questions:

I'm envisioning a ball-and-socket type of connection, so that the struts have the freedom to adjust to their necessary angles before being locked into place. (When you create a triangle, all pieces mutually lock each other into place, and nothing wiggles anymore. The triangle is the most stable geometric polygon, which is why it's employed so much in geodesics, and in regular architecture for reinforcing beams, trusses, and internal wall supports.) This design arose from discussions with some of my co-workers (and fellow dome-heads ;^) ), Skeeter Murphy, Don Moore, and Mike Fleck. (Thanks, guys! :-) )

Here are exploded and assembled views of the current hub design. There are both 5-way and 6-way hubs represented. A full-sphere geodesic requires 12 of the 5-way hubs, and a varying number of 6-way hubs, depending on the frequency of the dome. (A 4-frequency full sphere dome, such as in the images above, would require 150 6-way hubs.)

Note that the struts have been shortened for these renderings. Their actual length would depend on dome size, geometry parameters, and their actual position in the overall structure. Note also the cut-away portions in the hub itself, which allow the struts to move within certain ranges. (Yes, I realize there's a slight glitch in the sizing, so the openings don't mesh perfectly. As I said before, these are sketches. :-) ) I've calculated and designed the movement ranges to cover all possible angles in 3, 4, and 5-frequency geodesics. (If you aren't sure what is meant by a geodesic "frequency", or would like a refresher, or would like to calculate your own geodesic measurements, I highly recommend The Dome Calculator.) The top and bottom pieces bolt together, and the bolt remains inside the structure, so that it doesn't get exposed to weather.

Picture of an exploded view of the 5-way hub and strut components. Picture of an assembled view of the 5-way hub and strut components.

Picture of an exploded view of the 6-way hub and strut 
components. Picture of an exploded view of the 6-way hub and strut 

(Click on the 4th image to get a 1024x768 version, suitable as a background image, or just to see all the cool reflections the ray-tracer puts in there. :-) )

I have a feeling it should be relatively easy to have these things cast or machined in very large quantities. I tried to design them to be relatively easy to mass produce. If you, or someone you know, is in the machining/die-casting business, or rapid-prototyping, I'd be interested in getting your feedback on how easy/difficult this would be. Send me some email.

New Hub Designs

After talking with some people who actually do rapid-prototyping for a living, I've gotten some valuable feedback which has led to a redesign of the hub to make it more suited to injection molding/casting. The above renderings from early 2000, while looking very nice with all the reflections and such, are really better suited to milling out of solid chunks of material rather than casting from molten material. (The thick pieces not only add weight and waste lots of material, but would lead to contraction as the parts cooled if they were cast with such thick volumes, leading to inaccuracies in the final product. (Think of how molten candle wax contracts as it cools.)

To this end, in 2002 and 2003 I've redesigned the parts with a thin, honeycombed structure to reduce weight while retaining strength, and which will be much more suited for injection molding and casting.

Picture of the redesigned 5-way
Universal Hub showing new honeycomb design. Picture of the redesigned 6-way
Universal Hub showing new honeycomb design.

Current plans are to create plastic hubs for greenhouse kits as a way to both test the design in real-world (but not human-life-critical) situations, and to raise capital resources that will allow the manufacture of metal parts for use in Autonomous Houses and other structures. Please read this if you're interested in getting involved in the early stages of this project.

I have found one example of a machined ball connection which could be (and in fact, is) easily mounted onto the end of a pipe via a screw mechanism. It's manufactured for furniture that's made by USM for their Haller line of modular furniture. These are only about 3/4" in diameter, which is a bit smaller than what I was designing for, but it quite clearly shows the possibilities. They are currently manufactured in both steel and aluminum. If these were made in the appropriate size with only one milled and threaded hole, (which could be screw-mounted onto a pipe shaft), then we'd have the strut part of the design complete.

Picture of USM's Haller milled metal ball.

In line with the plastic version of the hub listed above, I have been evaluating some mass-produced plastic balls that could be either milled-and-glued, or simply glued directly onto the ends of plastic tubing. Some milling might still be necessary, but at this point milling the end of the tubing seems as effective (and less wasteful of material) as milling the ball.

Up to this point, I've been evaluating PVC balls and tubing, since the tubing and PVC cement are available cheaply at almost any hardware store (in America, at least), and would allow for constructing an inexpensive dome similar to our hydroponic greenhouse, using mass-produced parts. However, after reading Cradle to Cradle by William McDonough and Michael Braungart, I've begun to reconsider the use of PVC, since its manufacture and use is extremely harmful to the environment. I can order the balls made in just about any plastic, and the same holds true for the hubs. It may be that people could still use locally-obtained PVC tubing, but realistically, all the measuring/cutting/milling should be done by us in a controlled fashion, to ease assembly and reduce problems for the end-user. Also, using a single type of material (such as PVC or otherwise) makes for easier gluing/cementing of parts. However, such cements often have hazardous solvents and VOCs (Volatile Organic Compounds). Any pointers to other durable yet more environmentally benign plastics would be most appreciated.


Patrick Salsbury

Home | Mailing Lists | List Archives | Projects | Roadmap | FAQs | Wiki | Contributions
This page last modified: Wednesday, 25-Jun-2003 00:54:05 PDT

Mail comments care of 'webmistress' at this domain.
©1994-2008, Reality Sculptors