Rapid manufacturing

Rapid manufacturing (also known as direct manufacturing)
is an additive fabrication technology that makes complex
parts automatically from a CAD STL file. There are several
terms similar to rapid manufacturing: 3-Dimensional Printing,
additive fabrication, freeform fabrication, solid freeform
fabrication, and stereolithography. Most rapid manufacturing
systems make objects from flimsy materials (plastic, paper,
starch...) but a few can make strong metal parts. Some of
them can make inexpensive rocket engines.

Here is a brief review of rapid manufacturing:
http://www.triz-journal.com/archives/2003/10/l/12.pdf

The best source of information about rapid manufacturing
is the annual Wohlers Report ($390,
http://www.wohlersassociates.com)

The cheapest rapid manufacturing method is a process
similar to inkjet printing. The fastest printer of this
kind is ProMetal R2. It makes aluminum 6061-T6
(45 KSI = 310 MPa) parts having the maximum size of
8"x8"x6". It can make one such part in 24 hours.
ProMetal web site: http://www.prometal.com/equipment.html

Rapid manufacturing is limited to small parts because
the technology is still expensive. The 8"x8"x6" size
is rather small for a rocket engine. The engine would
probably have four small exhaust nozzles. The metal
parts made by the ProMetal R2 printer are dirty --
they are covered with unbound aluminum powder which
must be removed manually. The clean part is coated
with bronze to make it impervious and then sintered
in a small oven. I imagine that a large oven could
fuse the small parts into one large engine.

Metal parts produced by any sintering process have
poor corrosion resistance, so they should be coated
with a refractory metal if they are going to be in
contact with the hot exhaust gas. A laser-melting-
-powder method used by Optomec LENS-850 machine makes
corrosion-resistant parts, but the process is slow
(0.5 cubic inches per hour) and the machine is
expensive. The maximum part size is 18"x18"x42".
Arcam EBM S12 electron-beam-melting-powder machine
(http://www.arcam.com) also makes corrosion-resistant
parts having the maximum size of 8"x8"x7". Arcam is
three orders of magnitude faster than Optomec, but
it makes only steel and titanium parts. (Its web
site claims that aluminum alloy powder will available
soon.) 3D Systems Sinterstation selective laser
sintering machine is as slow as the Optomec LENS-850.
Parts made by the Sinterstation are of the same quality
(poor corrosion resistance) as parts made by the
ProMetal R2 printer. EOS machines are slower than
Sinterstation.

The best materials for regenerative rocket engines
(aluminum and copper) reflect the laser beam (albedo
up to 98%) rather than absorb it. The high albedo and
poor energy efficiency of lasers (typically less than
10% of electric energy is converted to laser beam
energy) strongly favor the Arcam electron beam system.
A powerful electron beam is easier to generate and
deflect than a powerful laser beam. Laser beams are
deflected by moving parts which cannot match the
scanning speed of the electron beam and require too
much maintenance.

It is theoretically possible to make an oversize
Arcam machine that can fabricate one big (3 ft. dia.)
aluminum, pressure-fed rocket engine in a week. The
speed of Arcam fabrication is now limited by the
speed of delivering thin layer of metal powder to
the spot where the electron beam melts it. If this
bottleneck is eliminated, the Arcam-like machine can
fabricate several big rocket engines in one day...

Which mean nothing in your case as you are too lazy to even work out the
dimensions of your designs.

Earl Colby Pottinger


--
I make public email sent to me! Hydrogen Peroxide Rockets, OpenBeos,
SerialTransfer 3.0, RAMDISK, BoatBuilding, DIY TabletPC. What happened to
the time? http://webhome.idirect.com/~earlcp

Andrew, I found your posting really interesting. Thanks for including the
URLs. Arcam's process (http://www.arcam.com) sounded really interesting.

The ProMetal (http://www.prometal.com/equipment.html) was interesting
also. It can make sponge metal, I wonder if it would be useful for
transpiration cooling of the engine. Or, maybe with the arcam process
capillary tube (transpiration) cooling.

http://www.afrlhorizons.com/Briefs/Oct04/ML0312.html


Craig Fink

I thought this was a moderated group? Nothing but a personal slur is
something I would have expected to see in sci.space.shuttle or .history,
not sci.space.tech.

Craig Fink

On Sun, 13 Feb 2005 21:35:17 -0600, Earl Colby Pottinger wrote:

Which mean nothing in your case as you are too lazy to even work out the
dimensions of your designs.

Earl Colby Pottinger

Craig Fink wrote:
Earl Colby Pottinger wrote:
Which mean nothing in your case as you are too lazy to even work out the
dimensions of your designs.

I thought this was a moderated group? Nothing but a personal slur is
something I would have expected to see in sci.space.shuttle or .history,
not sci.space.tech.

I had originally held Earl's posting for further review, but in
the end decided it was marginally inside the acceptable range.

Part of that decision was that there is a factual component to
Earl's post, that Andrew has not in fact posted actual dimentions
to the engines he is proposing and has stated that he hasn't
even completely worked them out. That is not directly relevant
to this particular topic, and Earl is not being entirely polite
about pointing it out, but it's a valid technical commentary.

In the future I will probably ask people to rephrase posts
that are that far into the grey area even if they do include a
valid technical comment. The "lazy" comment was not necessary
and did not contribute to a positive discussion here.


-george william herbert

(George William Herbert) :

Craig Fink wrote:
Earl Colby Pottinger wrote:
Which mean nothing in your case as you are too lazy to even work out the
dimensions of your designs.

I thought this was a moderated group? Nothing but a personal slur is
something I would have expected to see in sci.space.shuttle or .history,
not sci.space.tech.

I had originally held Earl's posting for further review, but in
the end decided it was marginally inside the acceptable range.

Part of that decision was that there is a factual component to
Earl's post, that Andrew has not in fact posted actual dimentions
to the engines he is proposing and has stated that he hasn't
even completely worked them out. That is not directly relevant
to this particular topic, and Earl is not being entirely polite
about pointing it out, but it's a valid technical commentary.

In the future I will probably ask people to rephrase posts
that are that far into the grey area even if they do include a
valid technical comment. The "lazy" comment was not necessary
and did not contribute to a positive discussion here.

To George William Herbert, understood. And if you had said the message
should not go thru I would understand your choice in that direction.

To C.F. and A.N. You are right. The words 'too lazy' are very personal and
should not have been in a public message of this nature. I let my emotions
get carried away again. Sorry. I may say the same elsewhere but this is a
TECH usenet group, and I should not have posted such here.

However, I think there is a very important point why I get carried away.
There is no magic tech for building any type of spacecraft that does not
involve crunching the numbers, or atleast blindly building a prototype and
testing it (HARD!). Promoting a hardware design that has no design specs,
then promoting a tech to build that hardware with the specs missing still
means not reaching the goal of a working motor. I have been hearing about
this cluster design for what seems years now and have seen no work done in
refining that design in all that time.

This bugs me personally as so many of my designs that look good on paper have
turned in real messes when tried as working hardware - details matter. One
of my designs failed because I forgot to add an O ring to the design.
Another, the screws were the wrong lenght. My latest designs have reached
the point of being very dangerous if I have a design fault. I spend weeks
tweaking diffirent ideas to see what can be done to meet my goals and still
be safe and it is not easy. And I know there is still lots more work needed
after reaching the first set of goals.

A 3D printer that can work in metal or other high temp materials will do
nothing to speed up development if care is not done in the design stages
first - just like a computer GIGO.

Earl Colby Pottinger

PS. If I have a 3D printer that worked with metal I would be more interested
in building a water cooled TPS. It would be ideal for the large number of
branching feed channels needed.

--
I make public email sent to me! Hydrogen Peroxide Rockets, OpenBeos,
SerialTransfer 3.0, RAMDISK, BoatBuilding, DIY TabletPC. What happened to
the time? http://webhome.idirect.com/~earlcp

On Thu, 17 Feb 2005 23:13:23 -0600, Earl Colby Pottinger wrote:


This bugs me personally as so many of my designs that look good on paper
have turned in real messes when tried as working hardware - details
matter. One of my designs failed because I forgot to add an O ring to
the design. Another, the screws were the wrong length. My latest
designs have reached the point of being very dangerous if I have a
design fault. I spend weeks tweaking different ideas to see what can be
done to meet my goals and still be safe and it is not easy. And I know
there is still lots more work needed after reaching the first set of
goals.




Happens to all of us, our own frustration with something overflows, or may
be projected on someone else.

I've never built any hardware, so from my point of view I'm somewhat
ignorant of the real world problems. That said, I would think that a 3-D
metal printer would really free up or eliminate some of the design
considerations wrt to just being able to manufacture something. Multi-part
objects can become one because there is no longer the requirement to fit
the manufacturing tool in to produce it. Kind of like making hardware the
way we make silicon chips. I would think that the process could even be
stopped in the middle of producing a part. That way, internal areas of the
part could be machined if necessary to meet tolerance requirements. Then
restart the 3-D building process to finish the part. Or print head
changes, like changing colors, print different metals or transition slowly
from one metal to another. A transitional alloy. Or, build internal stress
relief areas to isolate thermal stress to one or two dimensions.

When space transportation cost become reasonable, I would think there is
going to be another industrial revolution when many things will be
manufactured in space and then brought back to earth for sale. A vacuum
chamber is used in one of the machines so the electrons flow freely and
impurities in the metal are kept to a minimum. Space has plenty of vacuum.
Power shouldn't be a problem, why try to figure out how to bring it down
to earth, just use it up there. Zero gravity might be a big plus in
producing light weight objects. Huge parts might be possible, like unibody
airframes for aircraft.

Only time will tell.

PS. If I have a 3D printer that worked with metal I would be more
interested in building a water cooled TPS. It would be ideal for the
large number of branching feed channels needed.

Yeah, I agree, complex cooling channels within metal TPS would be much
easier. You might not even have to dump the coolant, just pump it to the
hot areas, where it vaporises. Then, the gas flows to cooler areas of the
vehicle to recondensed to be pumped back to the hot areas. Vastly
increasing the area available for radiant cooling, and using more vehicle
mass as a heat sink.

Craig Fink

Earl Colby Pottinger wrote:

PS. If I have a 3D printer that worked with metal
I would be more interested in building a water
cooled TPS.

It would be ideal for the large number of
branching feed channels needed.

I'm going to answer these in the other order.

1) You'd probably want the channels to be a lattice
rather than a tree, to provide redundant paths, and
just let the openings to the nozzle be much less
capacious than the feed channels, to meter the flow.

2) [I am not a rocket scientist, so this next may
make no sense whatever, but then again...]

Why use water? With a complex enough channeling
available, in my ignorance I can envision a
throatless nozzle, where the fuel itself is all used
as coolant before it is used as fuel, and emerges
even perhaps from the entire interior surface of the
nozzle, or else flows in (essentially) non-cooling
channels down to the lip and then does a backflow
from lip to base before emerging, like a tuna's
retia mirabilia counter-current heat exchanger(*),
moving the heat cooled from the nozzle to where it
will most help ignition.

This might, however, have the coolant enter the
nozzle at really immense speed, eroding the
pores through which it emerges, so the nozzle
had better be cheap indeed to build, it might
not be reusable.

Moreover, it wouldn't be any extra manufacturing
complexity once you have 3D "print to build"
technology as your base, to have the channels be two
separate but interwoven sets, unconnected
internally, and to have the oxidizer and fuel emerge
and mix at the nozzle interior surface.

I suppose one would need something the moral
equivalent of a glow plug at the top to set things
going and keep them going, which would make for an
attractive restartable engine, too.

FWIW

xanthian.

(*) http://www.the-aps.org/press/conference/tuna.htm

Craig Fink wrote:


I've never built any hardware, so from my point of view I'm somewhat
ignorant of the real world problems. That said, I would think that a 3-D
metal printer would really free up or eliminate some of the design
considerations wrt to just being able to manufacture something.
Multi-part objects can become one because there is no longer the
requirement to fit
the manufacturing tool in to produce it.

A quick scan of the Pro-Metal pdf indicates that it may be early to make
this claim.

For instance, it appears that voids in the part either require good old
drilling out, or making the part in 2 pieces, each with half the void.
When a variation on "lost wax" casting gets included with this, that issue
may disappear.

I would think the sintering process would also have some negatives for
critical-dimension parts -- how much shrinkage would the part undergo, and
can it be predicted accurately enough to maintain tolerances? The obvious
answer is to finish up with machining, but that may negate the "no need to
fit the manufacturing tool in" advantage.

Also, can grain production be controlled appropriately with this process?
Many parts with severe service requirements require precise control of the
grains ("tempering").

I think the print-a-part scheme currently wins only for low volume or
prototype production where making jigs or molds would be prohibitive.



[...] I would think that the process could even be
stopped in the middle of producing a part. That way, internal areas of
the part could be machined if necessary to meet tolerance requirements.
Then
restart the 3-D building process to finish the part. Or print head
changes, like changing colors, print different metals or transition
slowly
from one metal to another. A transitional alloy. Or, build internal
stress relief areas to isolate thermal stress to one or two dimensions.

This may be possible, but it would require being able to accurately
recalibrate the positioning mechansism to the part after each change.
Probably within fidiciual art these days, but do the current print-a-part
machines have that capability?

/dps


--
Using Opera's revolutionary e-mail client: http://www.opera.com/m2/

D Schneider wrote:

A quick scan of the Pro-Metal pdf indicates that it may be early to make
this claim.

For instance, it appears that voids in the part either require good old
drilling out, or making the part in 2 pieces, each with half the void.
When a variation on "lost wax" casting gets included with this, that issue
may disappear.

Not true. Very long and convoluted holes
are troublesome because special tools may
be needed to remove the loose powder from
the holes.

I would think the sintering process would also have some negatives for
critical-dimension parts -- how much shrinkage would the part undergo, and
can it be predicted accurately enough to maintain tolerances? The obvious
answer is to finish up with machining, but that may negate the "no need to
fit the manufacturing tool in" advantage.

True

Also, can grain production be controlled appropriately with this process?
Many parts with severe service requirements require precise control of the
grains ("tempering").

Sintered objects are not as strong as the
objects made by conventional methods, or the
laser melting (Optomec), or electron beam
melting (Arcam).

As of now the only the Optomec machine can make
strong metal parts having gradually changing
composition.

PS. Behrokh Khoshnevis is trying to make houses
using the rapid manufacturing technology. He calls
it "contour crafting." (www-rcf.usc.edu/~khoshnev)