Re: [EE]: Reprap circuit printing
Bryan Bishop
> The second generation of reprap is supposed to support printing
> circuit. The method proposed is printing low-melt metal such as field's
> metal or woods metal. This has a number of drawbacks, the low
> melt-point is generally incompatible with any form of power electronics
> since high temperatures could cause melting, woods metal is toxic and
> the alternative, field's metal, melts at an even lower temperature.
> Because the molten metal has high surface tension, it must be poured
> into channels rather than being extruded onto a smooth surface.
One of the big deals about RepRap is that it's designed ad-hoc,
meaning this stuff wasn't thought about in terms of "this component,
which would make it self-replicating, can't be made because we're
starting with _blah_ material, which can't be made by that other
component". Let me briefly explain-- Freitas previously published in
"Advanced Automation for Space Missions" and "Kinematic
Self-Replicating Machines" (KSRM) about this problem of
closure-engineering, which requires a little foresight on behalf of
the designers (to make a long story short, there's an 'open
manufacturing' mailing list out there working on a system to evaluate
a design for 'closure'):
Excerpts from KSRM on "closure engineering" --
http://groups.google.com/group/openmanufacturing/msg/4ff7a92e2425dde2
This doesn't immediately answer your question, of course, because it's
more long term rather than the "pick a material, any material and run"
method that you're asking help with.
In the beginning of integrated chips and the semiconductor industry,
guys were using huge sheets of wax and lenses from cameras bought from
down the street, and the circuits were millimeter scale, not
micrometer scale. While we're clearly not going to write out a digital
circuit at this large of a scale, it might still allow for analog
circuits. Ben Lipkowitz and I have been interested in clay pit
clanking replicators because of this, the aluminum and silicon being
useful for the glass or possibly other ways at getting very simple
circuits, maybe.
Another idea floating around is microfluidic logic, either through the
shrinky-dink methods where you throw it into an oven and it shrinks
down to size, or through the more conventional fluidics like you see
inside of pond assemblies when you break those down.
And then there's graphene and nanocrystals. The nanocrystal synthesis
methods are getting pretty interesting. To my knowledge there's no
simple semiconductor nanocrystal synthesis method, but I do know of a
few --
http://heybryan.org/~bbishop/docs/nano/eznanoparticles/
.. requiring low temperature plasmas, though. :-/ CNTs require CVDs,
which is a big turn off for a clanking replicator IMHO.
Graphene might be interesting however. You can use scannng tunneling
microscopes or atomic force microscopy to very delicately position
graphene into the circuits that everybody has been raving about:
http://heybryan.org/graphene.html
There are amateurs (or professionals?) who have worked on building
ridiculously simple piezo-actuated STM setups for ~$100 with things
like broken speakers and some screws.
http://heybryan.org/instrumentation/instru.html
The resolution isn't quite down to what would be needed for graphene
transistors (10 nm). That's a fairly large expectation for amateur STM
equipment; presumably the overall system needs to be more robust than
relying on 10 nm accuracy of tacky/hacky equipment. ;-)
Uh, there's probably a number of other materials I'm not remembering.
http://heybryan.org/semiconductor.html
> Based on these issues, I've been looking for another solution for
> printing conductive material.
> The broad requirements for the material are these:
> 1. Must be printable, that is it must be possible to load a syringe up
> with the material, squeeze it out onto the substrate, and let it cure.
> Two-part compounds would be alright provided that a mechanism for
> clearing the print nozzle is practical.
> 2. Must be conductive, at least as conductive as gold (Ag>Cu>Al>Au)
> 3. Must be readily available, this material should be something that can
> be made easily or purchased easily. Ideally, I'd like it to be less
> expensive than conductive epoxy. ($26 for 6mL seems pricey to me)
> 4. Must not require high temperature baking (>150C)
> 5. Should not require lengthy cure time, this will slow fabing
> significantly on complex parts
> 6. Ideally, should be solderable
Good luck. One area that you might want to look into is proteins and
biological material. These could be grown and harvested, but the trick
is finding something sufficiently conductive and not requiring a lot
of specialty chemicals for feedstock, harvesting, that sort of thing.
(remnants of the "self-replicating bioreactor" concept - even the
plastic tanks would be made from the cell membranes, heh).
> There are four broad possibilities I've found:
> 1. Conductive particle suspensions, such as conductive epoxies.
> 2. Conductive thick films
> 3. Conductive ceramics
> 4. Conductive polymers such as polyacetylene
>
> I've not been able to find much information on thick films, conductive
> ceramics, or conductive polymers. I have found information relating to
> some of the properties of polyacetylene, but no melting points, etc.
> which are necessary to consider integration into a reprap.
>
> I'd like to know if I've missed something obvious or if there's another
> material out there which fits this bill. Reprap could have a great
> future before it if we can get past these initial growing pains.
Yeah, so to help solve the growing pains, just to throw in an ad here
I guess :-), what the OSCOMAK/SKDB and other projects (Freitas?) are
doing is this simple 'manufacturing web' where processes and materials
are cross-referenced to detail which process requires what. This then
allows a search algorithm to go through and do 'dependency checking'
much like in the debian APT system. Imagine going through and just
"fishing out" a replicating mechanism from the possibility space. At
the moment some aspects of the project have NSF funding for cyber
repositories for open hardware designs, so things are moving well.
Best of luck,
Smári McCarthy
Hash: SHA1
Rapid Prototyped Electronic Circuits,
http://staff.bath.ac.uk/ensab/replicator/Downloads/report-01-04.doc
- Smári
Smári McCarthy
sm...@yaxic.org http://smari.yaxic.org
(+354) 662 2701 - "Technology is about people"
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iEYEARECAAYFAkktsvAACgkQ9cJSn8kDvvH63ACeMfnPX35WHh2Yc5DjzdR6IOhb
PqcAn3OjEXc3prKfov4PGAbQx2PAiNy5
=A0dO
-----END PGP SIGNATURE-----
ben lipkowitz
> Rapid Prototyped Electronic Circuits,
> http://staff.bath.ac.uk/ensab/replicator/Downloads/report-01-04.doc
I converted the doc to pdf (3MB on ~70kB/s up DSL):
http://fennetic.net/pub/irc/reprap_circuits.pdf
Bryan Bishop
Heh, and I converted it to HTML the other day ..
http://heybryan.org/books/Manufacturing/rapid_prototyped_electronic_circuits/report.html
Who the heck writes in DOC anyway?
- Bryan :-)