Talk:Planar vs Convergent Assembly
From Wise Nano
Some possible elaborations by 143.182.124.3 14:42, 3 Dec 2004 (CST)
Chris:
You might get a the benefits of corner-wise assembly of cubes, by instead using long hexagonal-cross-section blocks (hexblocks). These could be brought in at a slight angle, the end pushed into the corner of previous "wall" of blocks, and tilted up to settle into position. The wider angle would be less likely to cause binding at incorrect angles. Alternately, the hexblocks could be laid in diagonally (analogous to corner-wise cubes), with the benefit of perhaps locking layers of hexbocks together more securely. (I don't think this should be necessary, and it complicates design and assembly.)
Hexagons allow producing strong hexagonal cross-section cells, as well as larger diameter cells with smoother walls than rectangular blocks of the same diameter - useful when you want to reduce mass, or create enclosures for other types of parts. And of course, hexagonal is a fairly natural shape for diamond, though I don't know as that matters much at the nanoblock scale.
Longer, somewhat skinny blocks - say 9x longer than thick - should allow packing more assembly points in a given area for a given volume of block - speeding assembly ~4x. For most structural components this would not be a major limitation. The smaller diameter (plus the hexagonal shape) would also make smoother surfaces when building rotating parts. Long blocks may also give the assembler more of a "handle" to grip while moving the hexblock into position from an angle, than a cube would being inserted corner-wise.
It should be reasonable to allow varying the length of hexblocks, keeping the thickness the same, without making the assembler arm or insertion process significantly more complex. This would be useful when making precise 3D curved surfaces. In fact, I suspect that with a bit of cleverness, one would need no more than two standard lengths to create most 3D smooth curved objects, precise to about 1/3 of a hexblock diameter. If the hexblocks are half the diameter of a cubical hexblock, this reduces the worst-case roughness of surfaces by 1/6th to 1/3rd, though at the cost of no longer having truly flat surfaces. If an atomically precise smooth flat surface is needed, special half-hex blocks might be attached to the surface - but generally if two surfaces must slide against each other, it should be sufficient to simply rotate them 90 degrees with respect to each other, so that the hexagonal "notches" do not engage like gears.
Most rigid construction might use a hexblock with a pre-inserted pin, that the assembler need only push further in to engage locking mechanisms to connect adjacent hexblocks. The beveled end of the pin could be left protruding a bit in most cases, so that a hole in the end of the next layer of hexblocks can be centered on it to further help in aligning the end of the hexblock as it is inserted, insuring that the block is positioned tightly against adjacent blocks. The protruding pin would also help hold blocks in position until locked against adjacent blocks. The pin itself is probably designed to be locked in position, preventing slippage except when unlocked by the assembler (or potentially by a recycling disassembler). The pin would be short enough to be completely pushed into the hexblock (a scond locking position) if the protruding pin is not desired - e.g. if two layers of hexblocks need to slide horizontaly with respect to each other for some reason.
Locking pins might be designed to push at angles out of two sides separated by one side without locking pins, with any side able to accept locking pins sliding into them. This allows surfaces with no protruding locking pins, and generally will mean four of six sides have pins locking the block to adjacent blocks. With proper planning, no block - even on an outer surface with three exposed sides - should have less than 3 sides locked. (I haven't proven this, but it seems right...) The same holes on the sides and top/bottom of the hexblocks would allow attachment of other functional nanoblocks.
The above comment was mine... by Tom Craver 14:47, 3 Dec 2004 (CST)
I guess I got logged out before I hit Save...
