The first challenging build I did was the tetrahedron. The method differed from the ones I use now, and I got much more familiar with the chaos of sticks and strings that exists before finishing a structure. Although most likely the icosahedron plays a much more important role for biotensegrity, our vertebrae possess a very tetrahedra-like appearance.
I experimented lots with combining tetrahedra into larger structures. Two of them joined at a corner produce an hourglass shape. I build as well a tower of five stacked tetrahedra, which took me ages to balance, and still doesn't satisfy me much. Most of the time, I approached larger structures with tetrahedra in a more complex way than needed. When two 6-strut tetrahedra are joined at a face, the 'center' triangular consists of 6 struts, while only three are needed.
In a tetrahelix, faces of tetrahedra are joined. When enough tetrahedra are joined in this way, the corners seems to build a helix (hence the name....). I hoped that five tetras would produce some interesting effect, yet joining pre-build tetras turned out more of a challenge than I hoped for. There's more than one way to connect modules to each other, and each yields different results.
Then I decided a new approach. Instead of perceiving the tetrahelix as compound of tetra modules, I tried to understand it as singular structure enclosing connected tetrahedral spaces. So if I take a tetrahedron and extend it to enclose to tetrahedral spaces, I get a tensegrity tetrahelix without a single tetrahedron remaining visible.
The first 5-stage tetrahelix required 30 struts, yet when my idea works out with only 21 struts I have six stages (6 struts for the 'seed' tetra, 3 for each additional stage). Instead of connecting an incomplete three-strut tensul 'somehow', I connected the extension into the corner loops. Although that meant some more steps per connection than usual, I found an easy way to do so. The first extension, however, brought the biggest surprise. I just build myself another trigonal dipyramid.
I used a magnetic model to figure out the corner configuration for the tetrahelix. Each additional tetrahedron adds only a single vertice to the helix. So the dipyramid has two tetras and five corners, lets make this a bit more systematic.
struts # tetra spaces # corners # type
6 # 1 # 4 # tetrahedron
9 # 2 # 5 # trigonal dipyramid
12 # 3 # 6 # 3-stage tetrahelix
15 # 4 # 7 # 4-stage tetrahelix
n*3+3 # n # n+3 # n-stage tetrahelix
I'm not sure whether an object encompassing 3 tetrahedral spaces already deserves the name tetrahelix, it takes at least 5 tetrahedra to have all possible vertice configurations (3, 4, 5 and 6 struts converging in one corner). However, knowing that the maximum number of struts meeting in a single is six made extending the helix a breeze. The corner at the end has 3 struts, on the second level there are either four, five or six struts meeting. Thus there are three different constellation for the faces connecting to the 'top' (or bottom) corner: 3-4-5, 3-4-6 and 3-5-6.
So all I needed to do when extending the helix was finding the 3-4-5 face, and adding an incomplete tensul to the vertices of this face. I think I went up to 69 struts, creating something quite floppy which might connect into a torus.
Unlike most other models, I didn't manage to create something self-balancing. Up to about seven or eight stages, the model kept straight while I held one end, before the elasticity of the string made it bend a lot. While it's not too suitable for mere display, it's a lot of fun to play with. At the moment I have it hanging around, with an EL-wire threaded through the corners. In this constellation, the helical structure becomes visible, which otherwise remains quite hidden in the chaos of sticks and strings.