Nano Product Design

Responding to the previous post about the "tidal wave of change" that a sea of new nanotech designers might bring, blog reader Brett Bellmore wrote:

Would your proposed regulatory regime REALLY let everyone be a designer? ... I can't imagine the uber-regulators letting people just design and make things, and then trade the designs. They'd want to at least review the plans, and unless they had AIs available by the millions, that would be quite a bottleneck.

Here is how CRN addressed that issue in our paper on "Safe Utilization of Advanced Nanotechnology":

Rapid innovation is a key benefit of nanotechnology. The rapid and flexible manufacturing process allows a design to be built and tested almost immediately. Because designers of nano-built products do not have to do any actual nanotechnology research, a high level of innovation can be accommodated without giving designers any access to dangerous kinds of products.

[A] design with billion-atom, sub-micron blocks—permitting specification of near-biological levels of complexity—would still pose no risk of illicit self-replication. The minimum building block size in a design could be restricted by the design system. A fully automated evaluation and approval process could also consider the energy and power contained in the design, its mechanical integrity, and the amount of computer power built in. The block-based design system provides a simple interface to the block-based convergent assembly system. A variety of design systems could be implemented using the same nanofactory hardware, and the designer would not have to become an expert on the process of construction to create buildable designs.

With a safe-design nanofactory, adults—and even children—could safely play with advanced robotics, inventing and constructing almost anything they could imagine. (Today, adults as well as children find it worthwhile to play with the Lego MindStorms™ system.) More powerful products would require an engineering certification. This could be given to any responsible adult, since even a malicious product engineer would be unable to bypass the factory's programming and cause it to make illicit fabricators. A product that included chemical or nanomechanical manipulation ability would have to be carefully controlled, even during the design phase, to prevent the designer from building something that could be used for illicit nanomanufacturing.

Risks and dangers associated with products could be assessed on a per-product basis. Many products, produced with simplified design kits, could be approved with only automated analysis of their design. Most others could be approved after a safety and efficacy assessment similar to today's approval processes. Only rarely would a new degree of nanotechnological functionality be required, so each case could be carefully assessed before the functionality was added to appropriately restricted design programs.

Product approval for worldwide availability could depend on any of several factors. First, unless designed with a child-safe design program, it could be evaluated for engineering safety. Second, if the design incorporated intellectual property, the owner of the property could specify licensing terms. Third, local jurisdictional restrictions could be imposed, tagging the file according to where it could and could not be manufactured. Finally, the design would be placed in the global catalog, available for anyone to use.

We'd be very interested in serious critiques and/or alternative proposals. Thanks!

Comments

[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.

Chris

[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.

[daniel]

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.

Chris

[avinash]

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.