• Google
    This Blog Web

October 2011

Sun Mon Tue Wed Thu Fri Sat
2 3 4 5 6 7 8
9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29
30 31          

RSS Feed

Bookmark and Share

Email Feed

  • Powered by FeedBlitz

« A Tidal Wave of Change | Main | Fear of Nanotech »

March 25, 2004


Feed You can follow this conversation by subscribing to the comment feed for this post.

Brett Bellmore

I highly doubt that "dangerous products" is going to be so narrowly defined as products capable of mechanochemistry. As I pointed out, I've got a machine shop in the garage, and I'm perfectly capable of turning out products there, with no self-replicating or chemical synthisis abilities at all, that would have a SWAT team kicking down my door at 3:00AM.

What I doubt is the ability of programs, short of artifical intelligence, to enforce the design rules we're likely to be presented with.

Brett Bellmore

I've had a chance now to review your paper on nanofactory design, and a number of aspects of the proposed design strike me as curious.

First, construction of products is dominated by the time to construct the nanoblocks making them up. Considering that the proposed design relies on a finite number of pre-designed blocks, wouldn't it make more sense to have a number of first in, last out buffers for each nanoblock design, with the nanoblock fabrication stages continually topping them off? In that case, product delivery could be dominated by assembly time, assuming that nanofactory isn't being continually used at it's maximum capacity. Even for designs containing unique blocks, the entire block fabrication capacity could be devoted to manufacturing just those blocks, which would presumably comprise only a tiny fraction of the product's mass.

Second, this "unfolding" business strikes me as a real pain. I picture the nanofactory as an array of prismatic rods, containing nanoblock fabrication systems, the bulk of the rod comprised of nanoblock buffers, and with convergent assembly stages at one end, ending in a system capable of placing the largest scale blocks anywhere on the end surface of the rod.

Such functional units could be capable by themselves of replication, while nanofactories capable of manufacturing larger products could be assembled out of arrays of such rods.

As I envision the operation of this, the surface of the array would be covered by a plate consisting of an array of units capable of powered sliding motion, while maintaining a good seal. The nanofactory would start manufacturing (Or just delivering; They could be recycled.) more of these around the perimeter of the area the product was destined to emerge from, causing a vacum chamber to rise out of it's surface. This would proceed at the same speed the product was assembled, and when the product was completed, the entire area would be covered with more of these sealing units, allowing the vacum chamber to be opened without compromising the nanofactory's vacum.

Chris Phoenix, CRN

Brett, my nanofactory design was the simplest one I thought could work. The cube assembly was extremely simple, because anything more intricate would take more complex robotics and planning. My version takes only a 3 DOF gantry crane (modified for VdW-dominated regime in the smallest branches). I have no doubt that better designs can be found.

Since each fabricator's nanoblock winds up in a predetermined place in the product, it's very simple to route the assembly instructions. But that means you can't start making the nanoblock until the product design arrives. (Somewhere in the depths of my paper, I think I suggest caching commonly used blocks. But maybe I took that out before the final version.)

If you want rapid assembly of products, start with the insight that most of the mass of a product will be structural/inert, simply because active mass is extremely power-dense (though also efficient). So start with a prefabricated aggregate block of 90% solid blocks, 2% computers, 2% motors, 2% signal-routing blocks, 4% miscellaneous, all evenly distributed through the block. Load it into the output port of a convergent-assembly system, and run it backward to pull it apart into individual nanoblocks, with each sub-branch getting a mix of parts. Then run it forward again while reshuffling the blocks. You could make a product in seconds.

Post-assembly unfolding makes several things easier: product delivery, handling during convergent assembly, size of internal nanofactory tunnels, and a couple more I can't remember right now. So we may be stuck with unfolding. Remember that it may be more like inflating than unfolding--not every motion has to be kinematically planned and constrained/powered.

Good questions! Thanks for asking them. Please start working on nanofactory version 2. It may turn out that I've underestimated the required flexibility of a "primitive" nanofactory by an order of magnitude or two.


Brett Bellmore

As I see it, the chief advantage of my concept is that it allows the nanofactory to manufacture another nanofactory (And any other product of it's own size.) without the design being required to incorporate the complexity of self-unpacking from a compact configuration. A further advantages is that the design is capable of scaling itself up to handle larger products, simply by manufacturing more copies of the subunit to add to the edge. A process which can actually proceed automatically.

At some point in the convergent assembly process, you reach a block size sufficiently large that the utility of continuing convergent assembly is overcome by the utility of not having to be able to fit the product within a fraction of the nanofactory's interior. This happens at a point where the assembled blocks are large enough that the added computation to manage non-convergent assembly is relatively minor, and is probably determined by the block placement frequency and size of the blocks. If you're placing blocks at a rate of, (hypothetically) 100 Hz, and they're a millimeter in size, you could complete a meter of assembly in ten seconds. Not much point in doing it faster at that point...

The buffers really just serve to smooth out the factory's power consumption and heat generation; Instead of drawing massive amounts of power for short intervals, block fabrication would proceed continuously at some low level.

In any event, I expect that the real chanllenge is to be found at the mechanochemistry fabricator stage, not in specifying the overall system arthetecture.


I am a product designer that is very interesting in the new nanotechnology and its possible appplications. I would like to have some more information about the issue.

Chris Phoenix, CRN

Daniel, the technology doesn't exist yet. If you're interested in a technology that may be ten years away, email me for more info.



i am also looking to do something in a nanotecnology like i am looking for someproduct that i can enter in the commercial market.so if u can suggests any nanoproduct it will be a great help.

The comments to this entry are closed.