Requirements for an assembler?
James Logajan
I'm having a small problem with the very first step designing a "Self
Replicating Nano Assembler Version 0" and need help with it:
What problem should it solve? That is, the first step in any engineering
project is to establish requirements, and to do that the objective needs to
be clear and useful. No point setting out on a long engineering journey
without a goal.
But it isn't as easy to answer as it seems (well, not for me anyway). The
problem should be such that you would need "billions and billions" of
assemblers (so only self replication is a viable solution). And it should
be such that bulk chemical processes could not accomplish the task. And a
solution to the problem or goal should be of sufficient value that the
investment in time and money is worthwhile. I realize that the previous
"requirements" are ass-backward, but this is an attempt to find a
legitimate rationale for development of a special-purpose assembler. Not a
Swiss-army-knife-like general assembler that can be programmed to perform
all sorts of tasks, but an assembler that self-replicates and has
capabilities that address _one_ macro-level problem of interest to humans.
It doesn't _have_ to address a problem a lot of people have - it can
address the problem of even a small group of people, provided that small
group has the resources and motivation to support such an endeavor.
One near-miss example is an assembler that mass produces high efficiency
solar arrays at extremely low cost. I _think_ this is a near-miss because
bulk processes may eventually be able do the job. The other important
missing piece of widespread solar power is a low cost, high efficiency,
high energy density electrical power storage. But again, there may be bulk
processes that will lower the cost of, say, modern ultracapacitors.
Other areas:
Anything in the biological realm that needs doing that probably only a
self-replicating assembler could handle? I'm not too knowledgeable about
current and future capabilities in that area.
Transportation? Maybe super-duper self-repairing roads? Pot-holes cause a
lot of damage, and road repair is a neverending job - would auto insurance
companies and government transportation departments have the long term
dedication needed to finance research along those lines?
Production of materials - bulk processes seem to excel in this area, so I
just don't see improved materials by way of replicating assemblers as being
a viable first goal.
Production of food - food is inexpensive, so far as I understand it.
Problems in this realm appear to be due to severe economic hardship whose
solution may not be found by technical means. But even if it could, the
people most in need of the solution do not have the resources to solve it,
and those with plenty of resources are not motivated to direct those
resources to solving a problem they do not suffer.
Elimination of poverty - same as the food situation.
Energy - other than what I mentioned above on solar, I think energy is
inexpensive enough that there is no motivation in making it even cheaper.
erincss
>James Logajan
>, but an assembler that self-replicates and has
>capabilities that address _one_ macro-level problem of interest to humans.
>Production of materials - bulk processes seem to excel in this area
Self Repairing Armor Materials for the military. We want a structure that is
light in weight, strong, hard, and reasonably flexible, which can straighten
itself out when folded or dented or deformed in any way. This would require the
material structure to be laced with assemblers. If the damage to the material
is very extensive, then new material (replacement atoms) will be required to
fill in what was lost. Material idea? Carbyne-Fullerene Composite, laced with
sapphire, too. Machine Phase Material.
Buy a product made of this once, and it can repair itself over a whole
lifetime.
Obiwan
How about artificial muscle?
Proper artificial muscle, that is, not stuff like shape memory alloy or
pneumatics, but molecular levers and ratchets, like biological muscle, but
powered by electricity. This might be a tall order for one design of
assembler, and may require several different ones working together.
The principle is simple, but it's realisation on a molecular scale is
difficult, and will probably need assemblers.
Solving this problem would open up lots of areas. Robotics, prosthetics,
transport, for starters would all benefit from replacing bulky, heavy,
noisy, inefficient electric motors with artificial muscles.
BEN
"James Logajan" <Jam...@lugoj.com> wrote in message
news:ar7ta...@enews4.newsguy.com...
>
> I'm having a small problem with the very first step designing a "Self
> Replicating Nano Assembler Version 0" and need help with it:
>
> What problem should it solve? That is, the first step in any engineering
> project is to establish requirements, and to do that the objective needs
to
> be clear and useful. No point setting out on a long engineering journey
> without a goal.
[ Remainder of quoted material elided by moderator. -JimL ]
John Michelsen
>Replicating Nano Assembler Version 0" and need help with it:
>
>What problem should it solve? That is, the first step in any engineering
>project is to establish requirements, and to do that the objective needs to
>be clear and useful. No point setting out on a long engineering journey
>without a goal.
>
>But it isn't as easy to answer as it seems (well, not for me anyway). The
>problem should be such that you would need "billions and billions" of
>assemblers (so only self replication is a viable solution). And it should
>be such that bulk chemical processes could not accomplish the task. And a
>solution to the problem or goal should be of sufficient value that the
>investment in time and money is worthwhile. I realize that the previous
>"requirements" are ass-backward, but this is an attempt to find a
>legitimate rationale for development of a special-purpose assembler.
In fact there are an astronomical number of products meeting these
requirements. Take a look at chemical synthesis - for each type
of molecule we want to make there are a series of bulk reactions that can
build it up. The problem is that each reaction has a fairly high defect
rate, 80-90% is considered a good percentage yield. Take your calculator
out and consider how much material you can waste in making a compound
if each step only has a 90% yield. A typical drug might take 20
reactions to make - thats 0.9^20 or a 12% yield.
Now consider the number of molecules that might have value (therapeutic,
industrial, etc). Most of them will be big (and so take a lot of reactions
to make) because the more atoms you include, the more possible variations in
structure are possible. But because of yield problems the vast majority
will be impossible to make with any efficiency. How many penicillins, etc.
are out there waiting to be discovered, but can't be because we can't make
them?
Now add in positional control at the atomic level and its potential to
reduce these type of yield problems.
>Anything in the biological realm that needs doing that probably only a
>self-replicating assembler could handle? I'm not too knowledgeable about
>current and future capabilities in that area.
Think of all the different types of molecules and structures in the
biological world, and all the possible interactions they have (for example)
in the human body. Bulk processes can't produce or control but a tiny
fraction of these currently. To come up with ways to heal people we mostly
we just throw in random molecules and hope to find one that does
something useful. It's not exactly a delicate or precise approach.
John
10of100
> Replicating Nano Assembler Version 0" and need help with it:
>
> What problem should it solve? That is, the first step in any engineering
> project is to establish requirements, and to do that the objective needs to
> be clear and useful. No point setting out on a long engineering journey
> without a goal.
Considering we are talking about first generation Assembler, one that
most likely will be primitive, how about a more specific purpose like
making a material to be later assembled into what ever we need. This
can range from metal alloy sheets with specific properties, such as
light yet strong, maybe tune in specific properties. Such as shining
ultraviolet light on it for an hour changes it from a malleable form
to a super hard form. Allowing inexpensive deployment of Space
habitats. Just an idea.
James Logajan
eri...@aol.com (erincss) wrote:
> Self Repairing Armor Materials for the military. We want a structure
> that is light in weight, strong, hard, and reasonably flexible, which
> can straighten itself out when folded or dented or deformed in any way.
> This would require the material structure to be laced with assemblers.
> If the damage to the material is very extensive, then new material
> (replacement atoms) will be required to fill in what was lost. Material
> idea? Carbyne-Fullerene Composite, laced with sapphire, too. Machine
> Phase Material.
Perhaps. Since it is a military application I'd consider it only if nothing
else more constructive can be thought up. It at least has the nice
qualities that the requirements are palpable enough that they can be
fleshed out in more detail with little effort, and a well-heeled
constituency for such a product exists.
In general I'm leery of the market for "smart" materials, though they have
a couple of properties that ease the engineering task. For example, the
assemblers may always be considered in physical contact with each other, so
design of navigation, communication, and transport systems are greatly
simplified over applications that require free-roaming assemblers.
Aaron F Stanton
Having recently worked in the pharmaceutical research world, I can tell you
that it's much less sane than merely throwing random chemicals at the body
and seeing what happens. The only reason that we don't all die horribly
from some odd "medicine" are the equally insane number of hoops you have to
jump through to get a drug approved. There are so many possible points of
failure in the process that it boggles the mind. Additionally, to get a
statistically useful sample size is very difficult even for pretty simple
sounding stuff. "Sure, it doesn't kill rats, and it seems to help
them...so, would you like to try it?"
Sorry to rant.
Anyway, the primary requirement for any assembler is that it can be built in
bulk quickly. Maybe it doesn't even have to be able to build itself if you
design a factory that can only build assemblers, but does so very quickly.
Maybe your assembler can be made via spontaneous self-assembly in a shake
and bake lab. It really doesn't matter, as long as your assembler can build
whatever it is you are loking for and you have a lot of them.
Heck, you might need a multi-step process...probably will. To build the
10^23 machines needed to make the sapphire rocket engine, you need to build
10^12 factories, which means you need 10^6 factory assemblers, which means
you need 10^3 factories that can only build those...which then means you
need only one special purpose seed assembler that can only build factories.
I'm sure it's at least that much of a pain.
Aaron
Incidentally, there are different regulations regarding medicinal devices
and drugs. Nano will make that get really interesting.
"John Michelsen" <jmic...@mail.com> wrote in message
news:arh5e...@enews2.newsguy.com...
Michael Richards
> Replicating Nano Assembler Version 0" and need help with it:
>
> What problem should it solve? That is, the first step in any engineering
> project is to establish requirements, and to do that the objective needs to
> be clear and useful. No point setting out on a long engineering journey
> without a goal.
"Clear and useful."
And it would probably be a good idea to start with somthing simple -
the complex things can come later.
As some people have observed, you have a large set of possible
choices. But you have to start somewhere.
How about making a material that could have the highest possible
strength to weight ratio? Something much stronger than steel and
lighter than balsa wood? Something made entirely from one element in a
straight-forward fashion, but something that is (probably) impossible
to build without an assembler?
It is called nanoDiamond, and the exact description of how to build it
is at http://nanoDiamond.info
This solves a useful problem, and is a clear description, so matches
your criteria.
And yes, this is a shameless plug - I designed it.
Michael
James Logajan
"Obiwan" <obi...@foosh.com> wrote:
>
> How about artificial muscle?
[ ... ]
> Solving this problem would open up lots of areas. Robotics,
> prosthetics, transport, for starters would all benefit from replacing
> bulky, heavy, noisy, inefficient electric motors with artificial
> muscles.
Unless I'm mistaken, macro-sized electric motors and actuators seem to be
adequate for industrial and transport applications. I'm not so sure the
niche applications in robotics and prosthetics is a compelling enough
market opportunity, but I'll add it to my list.
William DiMenna
An assembler will not be "nano" any time soon. All microtechnology has taken
something in existance on a larger scale and shrunk it down. (Remember
vacuum tubes?) Do we currently make complex assemblers on a macroscale? How?
Out of what?
I believe there may be a way to make an assemler that is very material and
shape specific in about one square inch of size. The amount of work would be
limited by its ability to carry a power supply and material to repair the
substrate.
MEMS engineers are having trouble coming up with anything useful besides
simple sensors. I hope we can focus this discussion on a more practical real
world problem. Any suggestions?
"James Logajan" <Jam...@lugoj.com> wrote in message
news:ar7ta...@enews4.newsguy.com...
>
> I'm having a small problem with the very first step designing a "Self
> Replicating Nano Assembler Version 0" and need help with it:
>
> What problem should it solve?
[ Excess quoted material elided by moderator. -JimL ]
James Logajan
jmic...@mail.com (John Michelsen) wrote:
> In fact there are an astronomical number of products meeting these
> requirements.
I agree the number is certainly large! But there is an even larger
astronomical number of dead ends!
> Think of all the different types of molecules and structures in the
> biological world, and all the possible interactions they have (for
> example) in the human body. Bulk processes can't produce or control
> but a tiny fraction of these currently. To come up with ways to heal
> people we mostly we just throw in random molecules and hope to find one
> that does something useful. It's not exactly a delicate or precise
> approach.
Any preference on what disease(s) or affliction(s) should be targeted
first? Ideally something that many people want fixed and whose cause is
known (or almost known) and for which there are no good treatments expected
any time soon. I'll say right off that I suspect that biological goals are
not a good first goal for the first assemblers or non-replicating
nanorobots. The requirements needed for free-floating nanorobots operating
in the soup that is a biological system just seem rather daunting to me.
James Logajan
Greg_...@hotmail.com (10of100) wrote:
> Considering we are talking about first generation Assembler, one that
> most likely will be primitive, how about a more specific purpose like
> making a material to be later assembled into what ever we need. This
> can range from metal alloy sheets with specific properties, such as
> light yet strong, maybe tune in specific properties. Such as shining
> ultraviolet light on it for an hour changes it from a malleable form
> to a super hard form. Allowing inexpensive deployment of Space
> habitats. Just an idea.
Space habitats aside, what earth-bound markets are in need of such material
and for which there are no good non-nanotech substitutes? Super-strong is
nice, but not compelling. Super-cheap is better, but aren't most materials
cheap enough that in most products it is something else that accounts for
the bulk of the cost?
For example, if I build an auto-frame and body out of practically free,
super strong, lightweight material, the price of such a car and its running
costs may be reduced, but wouldn't such an advance only cut the cost of the
car in half, at most?
James Logajan
"Aaron F Stanton" <ari...@in-motion.net> wrote:
> Anyway, the primary requirement for any assembler is that it can be
> built in bulk quickly. Maybe it doesn't even have to be able to build
> itself if you design a factory that can only build assemblers, but does
> so very quickly. Maybe your assembler can be made via spontaneous
> self-assembly in a shake and bake lab. It really doesn't matter, as
> long as your assembler can build whatever it is you are loking for and
> you have a lot of them.
It seems unlikely that one can design a complex system like a molecular
robot that can then be manufactured via shake-and-bake.
> Heck, you might need a multi-step process...probably will. To build
> the 10^23 machines needed to make the sapphire rocket engine, you need
> to build 10^12 factories,
So each factory needs to build 10^11 machines. If a machine has only 10,000
parts, then the factory must perform 10^15 operations. At 10^6
operations/second, that means it will take 10^9 seconds, or over 30 years!
> which means you need 10^6 factory assemblers,
> which means you need 10^3 factories that can only build those...which
> then means you need only one special purpose seed assembler that can
> only build factories.
Your first steps aren't as bad as the "last". But the nature of the beast
that is molecular manufacturing of macro sized objects is that the
traditional linear growth curve of factory systems is grossly inadequate.
Hence the need for exponential or higher rate production systems; e.g. self
replication.
> Incidentally, there are different regulations regarding medicinal
> devices and drugs. Nano will make that get really interesting.
Yup!
John Michelsen
Michael....@Gartner.com (Michael Richards) wrote in message:
>
> How about making a material that could have the highest possible
> strength to weight ratio? Something much stronger than steel and
> lighter than balsa wood? Something made entirely from one element in a
> straight-forward fashion, but something that is (probably) impossible
> to build without an assembler?
>
> It is called nanoDiamond, and the exact description of how to build it
> is at http://nanoDiamond.info
Hmm, without calculations you cant really make any claims about the
strength of these things. Also, I would imagine that you would get a
lot of junk stuck in all the holes. Water vapor might get trapped
inside and make it as heavy as a water filled sponge if it didnt have
an air-tight surface. If filled with air, it might make a good
explosive. Otherwise, its an interesting type of material - it might
be very crack resistent for example. I was thinking of similiar
things a few years ago:
http://web.archive.org/web/19981206025534/sugar.ps.uci.edu/~jmichels/implem/analog.html
John
Aaron F Stanton
"Michael Richards" <Michael....@Gartner.com> wrote in message
news:arkk...@enews2.newsguy.com...
Pretty interesting. It reminds me of zeolites made from pure carbon.
Yes, it would be a good explosive - insanely flammable.
I strongly believe that the strain on the double walled system would be too
great.
I noticed the minimization method you used, but what was the potential you
picked? Also, what program was used to draw the pictures?
Thanks,
Aaron
P.S. It's not so much a description of "how to build it" as "what it's
structure would be" - two very different things.
erincss
The structures at http://nanoDiamond.info remind me of this idea: A cubical
array of stacked short carbon nanotubes, built via assembler method (about the
only way you can do this) so that there are no holes/empty spaces, in a regular
cube formation. Perhaps the nanotubes are filled with some material or bonded
with something like sapphire.
John Michelsen
> first? Ideally something that many people want fixed and whose cause is
> known (or almost known) and for which there are no good treatments expected
> any time soon. I'll say right off that I suspect that biological goals are
> not a good first goal for the first assemblers or non-replicating
> nanorobots. The requirements needed for free-floating nanorobots operating
> in the soup that is a biological system just seem rather daunting to me.
Have you looked closely at Robert Freitas's work yet? He's desiged some fairly
simple nanobots that might make a big difference. Respirocytes and scavengers
of various types would be very nice to have.
Worldwide Causes of Mortality
Percentage
http://www.infoplease.com/ipa/A0779147.html
ischemic heart disease (inability to provide adequate oxygen to the heart)
13.7
cerebrovascular disease (inability to provide adequate oxygen to the brain)
9.5
=================================================================================
vascular disease
23.2
repirocytes, fat, cholesterol, hdl, blood clot scavengers
acute lower respiratory infections (pneumonia)
6.4
hiv/aids
4.2
diarrheal diseases
4.1
tuberculosis
2.8
cancer of trachea/bronchus/lung
2.3
=================================================================================
infectious disease
19.8
microbivores
chronic obstructive pulmonary disease (smoking)
4.2
repirocytes, soot particle scavengers
road traffic accidents (physical and bleeding damage)
2.2
clottocytes, repirocytes
perinatal conditions (various complications of pregnancy and childbirth)
4.0
Aaron F Stanton
"James Logajan" <Jam...@lugoj.com> wrote in message
news:arm9h...@enews2.newsguy.com...
>
> "Aaron F Stanton" <ari...@in-motion.net> wrote:
> > Anyway, the primary requirement for any assembler is that it can be
> > built in bulk quickly. Maybe it doesn't even have to be able to build
> > itself if you design a factory that can only build assemblers, but does
> > so very quickly. Maybe your assembler can be made via spontaneous
> > self-assembly in a shake and bake lab. It really doesn't matter, as
> > long as your assembler can build whatever it is you are loking for and
> > you have a lot of them.
>
> It seems unlikely that one can design a complex system like a molecular
> robot that can then be manufactured via shake-and-bake.
I agree completely. Shake and bake is not the way to go unless you want to
wait an insanely long time for random chance to do what design can do
faster. (I'm reminded of the "infinite improbability device" from the
Hitchhiker series.)
>
> > Heck, you might need a multi-step process...probably will. To build
> > the 10^23 machines needed to make the sapphire rocket engine, you need
> > to build 10^12 factories,
>
> So each factory needs to build 10^11 machines. If a machine has only
10,000
> parts, then the factory must perform 10^15 operations. At 10^6
> operations/second, that means it will take 10^9 seconds, or over 30 years!
Many thanks for doing the math there. I was pulling numbers out of nowhere,
not thinking clearly. Laziness was pretty much in charge at this point.
>
> > which means you need 10^6 factory assemblers,
> > which means you need 10^3 factories that can only build those...which
> > then means you need only one special purpose seed assembler that can
> > only build factories.
>
> Your first steps aren't as bad as the "last". But the nature of the beast
> that is molecular manufacturing of macro sized objects is that the
> traditional linear growth curve of factory systems is grossly inadequate.
> Hence the need for exponential or higher rate production systems; e.g.
self
> replication.
Self replication is the ideal method, most certainly. If the design for a
self replicating assembler is too difficult for an initial
design/manufacture, though, it will likely be necessary to go through a
multistep process before a system that can bootstrap itself is acheived.
(Compilers for programming languages sometimes go through iterations like
this, where the first stage is written in an entirely different language,
and all it can do is compile a subset of the language in question - a subset
that is just barely sufficient to write a compiler. Then a full
implementation of the compiler is written and then it bootstraps itself.
Admittedly, most compilers aren't done that way anymore, but that's how many
have been created in the past.)
If a fifteen step process (or more) is needed before even one self
replicating assembler arises, it's worth it. At that point, you win the
game.
Michael Richards
jmic...@mail.com (John Michelsen) wrote in message
news:<armim...@enews2.newsguy.com>...
> Michael....@Gartner.com (Michael Richards) wrote in message:
> > It is called nanoDiamond, and the exact description of how to build it
> > is at http://nanoDiamond.info
>
> Hmm, without calculations you cant really make any claims about the
> strength of these things. Also, I would imagine that you would get a
> lot of junk stuck in all the holes. Water vapor might get trapped
> inside and make it as heavy as a water filled sponge if it didnt have
> an air-tight surface. If filled with air, it might make a good
> explosive. Otherwise, its an interesting type of material - it might
> be very crack resistent for example.
> John
Correct, it needs calculations to determine the exact strength to
weight ratio (or more exactly, strength to mass ratio).
Water vapor could be a problem, I hadn't thought of that - I will have
to design an air-tight surface.
Why would it be a good explosive if filled with air?
Yes, hopefully extremely crack-resistant.
Glad you found it interesting.
Michael
erincss
>John Michelsen
>Have you looked closely at Robert Freitas's work yet? He's desiged some
>fairly
>simple nanobots that might make a big difference.
Regarding nano medical machines, I was pondering something. Would it be
feasible, before "free range" nano medical bots were developed, to have a type
of nano medical bot made of say fullerene spheres and rods, which would be
attached to a "feeder line" made of nanotubes that extend outside the body,
into a macroscale machine, allowing easier control of their motions? Imagine
the picture of what I'm describing. You'd have these nano or micro machines
made from fullerene and other material, connected to an exterior machine (maybe
a medical computer/bed) rather than microscopic nano medical bots that swim
through the bloodstream on their own.
Michael Richards
"Aaron F Stanton" <ari...@in-motion.net> wrote in message
news:<arn8q...@enews2.newsguy.com>...
> "Michael Richards" <Michael....@Gartner.com> wrote in message
> > It is called nanoDiamond, and the exact description of how to build it
> > is at http://nanoDiamond.info
>
> Pretty interesting. It reminds me of zeolites made from pure carbon.
> Yes, it would be a good explosive - insanely flammable.
> I strongly believe that the strain on the double walled system would be too
great.
> I noticed the minimization method you used, but what was the potential you
picked? Also, what program was used to draw the pictures?
> Thanks,
> Aaron
>
> P.S. It's not so much a description of "how to build it" as "what it's
> structure would be" - two very different things.
Why would it be insanely flammable? (John may already have answered
this.)
The strain on the double-walled system may be too great - I have yet
to compare it with other variations to see which has the minimal
energy (for example, two nanojacks, one inside the other, but not
tightly bonded).
By "potential", do you mean something other than the "RMS gradient of
0.01" I mention at End:Technical Bits? I pretty much optimized them
until they stopped moving, and then some more in case they were just
moving through a local minimal.
I used POV-Ray to draw the pictures, with similar spot-lights and area
lights on all models to get the highlighting just right.
And you are right, this description gives you no idea how to actually
build it, but I have a few ideas on that as well that need a bit more
work.
Michael
Aaron F Stanton
"Michael Richards" <Michael....@Gartner.com> wrote in message
news:arum7...@enews2.newsguy.com...
>
> "Aaron F Stanton" <ari...@in-motion.net> wrote in message
> news:<arn8q...@enews2.newsguy.com>...
> > "Michael Richards" <Michael....@Gartner.com> wrote in message
> > > ...
> > > It is called nanoDiamond, and the exact description of how to build it
> > > is at http://nanoDiamond.info
> > > ...
> >
> > Pretty interesting. It reminds me of zeolites made from pure carbon.
> > Yes, it would be a good explosive - insanely flammable.
> > I strongly believe that the strain on the double walled system would be
too
> great.
> > I noticed the minimization method you used, but what was the potential
you
> picked? Also, what program was used to draw the pictures?
> > Thanks,
> > Aaron
> >
> > P.S. It's not so much a description of "how to build it" as "what it's
> > structure would be" - two very different things.
>
> Why would it be insanely flammable? (John may already have answered
> this.)
Flammable for a few reasons that work together. First, it's nearly pure
carbon. Second, it's extremely porous - mostly hollow cavities. I
understand that's a design feature meant to make the structure lightweight,
but the effect is to make nearly every atom a surface atom, exposed to
anything that was floating through. If this was not sealed from the
environment, it would be highly exposed to oxygen. Part of why paper is so
much easier to light than a nice heavy wood log is how much surface area is
exposed to the air - sure, you can light the outside of the log, but it
takes a while for air to get to the interior. A structure like this could
have nearly every carbon atom exposed.
Terminating the structure with hydrogens doesn't help make it any less
flammable. You would then have a hydrocarbon that was exposed. It does get
rid of the double bonds, but those are not necessarily a bad thing - if you
have aromatic ring structures in places you want to be flat, they could be h
andy in maintaining stability.
Sealing the entire structure would help solve this, but then you have a
problem with a vacuum inside a sealed system. I'd fill it with nitrogen
gas, which is pretty non-reactive, and then seal it. Maintaining the seal
might be tough in the long run.
> The strain on the double-walled system may be too great - I have yet
> to compare it with other variations to see which has the minimal
> energy (for example, two nanojacks, one inside the other, but not
> tightly bonded).
I'm not sure what you mean by tightly bonded here.
> By "potential", do you mean something other than the "RMS gradient of
> 0.01" I mention at End:Technical Bits? I pretty much optimized them
> until they stopped moving, and then some more in case they were just
> moving through a local minimal.
No, you specified a method of optimization, but what was the equation you
used to determine the forces between atoms? There are many of them, some
better than others, and some are nearly useless. (A simple harmonic
potential between only two atoms, for example, is really only useful for
guessing frequencies of spectra associated with a structure that's already
minimized. A simple harmonic has bonds that cannot be broken, and thus you
can't ever get chemical reactions with it.)
Some potentials include factors for angles between bonds (3-body terms) and
torsion (4-body terms), along with another 4-body term that I can't recall
the exact name of but involves 3 atoms connected to a central atom.
Potentials like that works fairly well, but it can be a pain to program them
unless you already have a library written.
Then you get to ab initio (quantum mechanics) calculations, and I know you
didn't use those. The systems are way too large for that to be used in a
reasonable length of time on most PC's.
> I used POV-Ray to draw the pictures, with similar spot-lights and area
> lights on all models to get the highlighting just right.
> And you are right, this description gives you no idea how to actually
> build it, but I have a few ideas on that as well that need a bit more
> work.
>
> Michael
>
Overal this is interesting. Please don't take my comments as destructive
criticism.
Aaron
William DiMenna
Micron is manufacturing a device that is a camera with a self contained LED
light source that is swallowable. It is the size of a large pill. The device
is used to take pictures of the small intestine. All it does is take
pictures. There are no moving parts it. It is completely passive. We are a
long way from a nanosized robot that can function actively in the human
body. There are HUGE compatibility issues to be worked out. Not to mention
feasibility, toxicology, assembly, clinical trials, ect. In some cases it
takes up to 10 years for a known chemical compound to pass clinical trials.
Without an extremely major technical advance we are a very very long way
away from anyhting that has been discussed in this posting.
"erincss" <eri...@aol.com> wrote in message
news:arum7...@enews2.newsguy.com...
>
10of100
>
> Space habitats aside, what earth-bound markets are in need of such material
> and for which there are no good non-nanotech substitutes? Super-strong is
> nice, but not compelling. Super-cheap is better, but aren't most materials
> cheap enough that in most products it is something else that accounts for
> the bulk of the cost?
Your looking for established markets to improve?! Well let's see,
there's the bicycle sector with possibility of folding bicycles, and I
don't mean hinged. Auto sector with self repairing dents (you can
include almost any other device that can get banged up with this).
If we can somehow make the metal alloy reversible in properties,
such as a material that can go from the flexibility of mylar to super
hard back to mylar, you could have flexible fixed wing crafts that
don't need rudders or control flaps.
Then there's body armor that harden's when under fire but flexible
when not.
How about emergency support struts, a building that is about to
collapse can be held up for minutes longer to evacuated people.
How about emergency shelters could be put up in no time, and can
weather out any disaster that it is in.
True most of the things I mentioned might be done without nanotech but
it would be at a greater cost, not just with the production of the
material, but with the cost of the control mechanisms themselves.
An example, shape changing wings can be done with the technology we
have now but the control mechanisms that have to be used add weight
and complexity, thus increasing the cost of the device.
> For example, if I build an auto-frame and body out of practically free,
> super strong, lightweight material, the price of such a car and its running
> costs may be reduced, but wouldn't such an advance only cut the cost of the
> car in half, at most?
What your looking at, I think is, the cost of the labor. But wouldn't
an assembler reduce the cost of labor by making products easier to
maintain and assemble? Think of the cost of a house if you could put
up the frame of the house within a few hours, by yourself or assembly
of vehicles, where factories have computers directing the mylar type
metal alloy that I mentioned, into certain shapes and then hardens it,
reducing the time and effort to make a vehicle.
Michael Richards
> > The strain on the double-walled system may be too great - I have yet
> > to compare it with other variations to see which has the minimal
> > energy (for example, two nanojacks, one inside the other, but not
> > tightly bonded).
>
> I'm not sure what you mean by tightly bonded here.
By tightly bonded, I mean that no double-bonds exist on the
double-walled system - rather than one nanojack inside the other,
without covalent bonds between them, there would be a covalent bond
between each carbon atom on the inside nanojack and the corresponding
atom on the outside nanojack. (As well as the three covalent bonds to
neighboring atoms on the same nanojack.)
The reasoning for doing this was to reduce unwanted reactions and to
increase strength.
On another point, do you think it would be worthwhile scrapping the
design of http://nanodiamond.info and creating a similar one with
sapphire (aluminum oxide) which apparently is not insanely flammable,
as this design is?
(PS I had originally mentioned flammablility as a potential problem
with the design in the FAQ, but did not realize how much of a problem
it would be.)
> > By "potential", do you mean something other than the "RMS gradient of
> > 0.01" I mention at End:Technical Bits? ...
>
> No, you specified a method of optimization, but what was the equation you
> used to determine the forces between atoms? ...
If I understand you correctly, the answer is: I used Molecular
Mechanics + as defined in the Help of HyperChem thus:
"MM+
The most general method for molecular mechanics calculations,
developed principally for organic molecules as an extension of
MM2(tm). This is an all atom force field. HyperChem assigns atom types
and parameters not normally available to MM2 users, extending the
range of chemical compounds that this force field can accommodate. MM+
also provides cutoffs for nonbonded interactions, solvation,
constraints, and molecular dynamics not normally associated with MM2
calculations."
> Overal this is interesting. Please don't take my comments as destructive
> criticism.
Not at all, your suggestions are very welcome.
Michael
Aaron F Stanton
"Michael Richards" <Michael....@Gartner.com> wrote in message
news:asgta...@enews1.newsguy.com...
>
> "Aaron F Stanton" <ari...@in-motion.net> wrote in message
> news:<as073...@enews2.newsguy.com>...
>
> > > The strain on the double-walled system may be too great - I have yet
> > > to compare it with other variations to see which has the minimal
> > > energy (for example, two nanojacks, one inside the other, but not
> > > tightly bonded).
> >
> > I'm not sure what you mean by tightly bonded here.
>
> By tightly bonded, I mean that no double-bonds exist on the
> double-walled system - rather than one nanojack inside the other,
> without covalent bonds between them, there would be a covalent bond
> between each carbon atom on the inside nanojack and the corresponding
> atom on the outside nanojack. (As well as the three covalent bonds to
> neighboring atoms on the same nanojack.)
> The reasoning for doing this was to reduce unwanted reactions and to
> increase strength.
>
I think I understand - you mean that the two layers would be bonded to each
other. I think that it's still a strained situation. You might try
introducing some "spacer" atoms into the outer layer to reduce the strain,
although I can't immediately visualize where they would go.
> On another point, do you think it would be worthwhile scrapping the
> design of http://nanodiamond.info and creating a similar one with
> sapphire (aluminum oxide) which apparently is not insanely flammable,
> as this design is?
> (PS I had originally mentioned flammablility as a potential problem
> with the design in the FAQ, but did not realize how much of a problem
> it would be.)
>
I would do both if I were you. Properly sealed, this design isn't bad, I
think. (For certain applications, the flammability could be a plus...if you
don't mind military applications. Actually, if the burning can be
controlled, it might make for an interesting scaffolding. Fireworks, too,
though you would need to dope it with various other elements for pretty
colors.)
> > > By "potential", do you mean something other than the "RMS gradient of
> > > 0.01" I mention at End:Technical Bits? ...
> >
> > No, you specified a method of optimization, but what was the equation
you
> > used to determine the forces between atoms? ...
>
> If I understand you correctly, the answer is: I used Molecular
> Mechanics + as defined in the Help of HyperChem thus:
>
> "MM+
>
> The most general method for molecular mechanics calculations,
> developed principally for organic molecules as an extension of
> MM2(tm). This is an all atom force field. HyperChem assigns atom types
> and parameters not normally available to MM2 users, extending the
> range of chemical compounds that this force field can accommodate. MM+
> also provides cutoffs for nonbonded interactions, solvation,
> constraints, and molecular dynamics not normally associated with MM2
> calculations."
>
Got it. I'm familiar with MM2 (though I don't recall the exact form of it),
and HyperChem has a good reputation for solid power and good accuracy.
> > Overal this is interesting. Please don't take my comments as
destructive
> > criticism.
>
> Not at all, your suggestions are very welcome.
>
> Michael
>
There are a couple things you should think about - while potentials like MM+
are very handy for an initial first stab at things, and can be reasonably
accurate, ab initio methods are more precise. They are also unfortunately
quite computationally expensive, relative to other methods. Even the
concept of a "bond" is somewhat malleable in a quantum mechanics
perspective, especially the distinction between covalent and ionic bonds. A
bond can be thought of as a region of space with a high probability of
finding one or more electrons that are being strongly interacted with by two
or more nuclei.
That's rather an aside...for structures like these that are reasonably stiff
and strongly connected, molecular modeling should suffice. I anticipate
that quantum chemistry will only really come into play when designing the
portions of a system that are supposed to react in highly specific ways,
such as the tip of an assembling tool.
Aaron