Analog SFF, November 2008

by Dell Magazine Authors

  But early machine tools were human-controlled. It took a great deal of expertise to use them, and a fair amount of time to make each duplicate, even when using mechanical templates. In the late 1940s, John Parsons (head of a company that produced helicopter rotors) devised a way to make punch-card-operated electromechanical calculators generate templates for human-operated machine tools to follow. He then envisioned an extension of the system that would have automated machine tools follow the templates on their own. He became known as the father of numerical control technology and was awarded the National Medal of Technology in 1985.

  As computers developed, numerical control became quite sophisticated. One major step came when machine tools were designed to record the motions made by a human operator while that operator made an object. Then the machine could play back the recorded motions to make a duplicate of the object, and to do so as many times as desired. Another step was machine tools that could trace the contours of an object to be copied and carve material away from a chunk of metal or other material until the contours—and the object—were matched. Later it became possible to design an object on a computer and feed the design to automated equipment to make the object; this is CAD/CAM (computer-aided design/ computer-aided manufacturing). But the process remained expensive—"machine tools” meant drills and lathes capable of working wood or metal—and it remained impossible to make hollow objects in one piece.

  A common household decoration is artificial fruit and vegetables. How could a bell pepper be duplicated as an artificial? You could carve it out of wood. You could also make a mold and pour in plaster or melted wax or plastic. To save on raw material, you could pour in just enough wax or plastic to coat the inside of the mold. Then the result would be a hollow pepper, much like most of the imitations you are likely to run into. But it would not have inside it the whitish partitions, the seeds, or the green lumps that you can find in real peppers. To make an imitation with those internal features would require molding two half peppers and gluing the halves together. And of course that is done, perhaps not with imitation fruits and veggies, but with things like—for instance—rubber duckies with internal squeakers. The evidence is easy to find, for the lines where parts are glued together remain visible.

  There is an alternative way to make hollow objects with internal features. It is based on the recognition that a hollow object can be sliced, and then the slices can be made separately and fastened together, either on the fly or later on. Each slice, of course, will show a portion of the internal structure. Cast the slices in wax or plastic, stack them up, and glue them together, and you would have an imitation pepper with all its internal bits (as well as seams).

  "Rapid-prototyping” (also known as “instant manufacturing") and “3D-printing” tools both begin by representing an object as a series of slices in the form of computer images. The first step is to generate a three-dimensional image, either by scanning the object or by drawing the object with CAD software. Software, instead of a knife, then generates the slice images. One method of turning the slice images into physical slices—known as “selective laser sintering"—is based on the fact that a thin layer of powder (plastic, metal, or ceramic) can be heated with a laser so the powder particles are softened or partially melted and stick to each other. “Stereolithography” is the term for a process using a thin layer of liquid plastic hardened by exposure to ultraviolet light from a laser.4 “Two-photon 3D lithography” starts with a tank of liquid plastic and uses two laser beams to deliver just enough energy to harden a tiny spot of plastic where the two beams meet. Focal Point Microsystems (fpmicro.com) in Atlanta, Georgia, is developing this approach for making tiny objects, including electronic chips. At present stereolithography and selective laser sintering are much more suited to making large objects. There are other methods as well, such as fused deposition modeling or FDM (see below).

  One type of rapid prototyping machine or 3D printer consists of a platform across which a powder or liquid plastic can be spread, as in Figure 1A. The laser then draws the first slice, fusing the powder or solidifying the liquid (Figure 1B). Once that is done, the platform is lowered the thickness of one slice and another layer of powder (or liquid) is spread across it, covering the first slice with the material for another (Figure 1C). Then the laser draws the second slice. (Figure 1D). The process is repeated until all the slices have been printed and the complete object is finished. The level of detail, as well as the smoothness of the surface, depends on how thin the slices are.

  * * * *

  Figure 1A. The platform with a layer of powder on it.

  * * * *

  Figure 1B. Drawing one image slice with a laser.

  * * * *

  Figure 1C. Adding a second layer of powder.

  * * * *

  Figure 1D. Printing a second image slice.

  * * * *

  Do you see a problem? Making an object with holes in its surface allows the unsolidified powder or liquid to be shaken out when the process is finished. But a pepper has no holes in its surface. A pepper full of powder or liquid is not really what we're after. It's also a waste of raw material. Another approach is to use, instead of powder or liquid, a pasty material that will harden rapidly, such as plaster, beeswax, or a goopier liquid plastic (among other materials)5. Instead of a laser, mount above the platform a syringe full of this material, and squirt it out to draw the slices, as in Figure 2. The accuracy of detail in the slice depends on how fine a line of material the syringe can squirt.

  * * * *

  Rapid-Prototyping in Industry

  Many companies make rapid-prototyping or 3D-printing devices for industry. One of these companies is the Z Corporation, which uses a method originally developed at MIT in 1993. Like the machines described above, it has a platform that can be lowered the thickness of one layer at a time.6 It spreads a layer of powder across the platform. But instead of using a laser to melt the powder grains enough to stick together, it uses a form of ink-jet printer with four print heads that deposit a thin layer of binder material (in three colors and clear). The machines are the size of office copiers. The printed items must be treated with a liquid “infiltrant” that hardens the material enough to support handling. The items are not, however, sturdy enough to install in products intended for sale and use. They are prototypes, demonstration models, and the originals around which molds can be formed for making plastic, metal, and ceramic items for actual use. These uses are valuable enough to make the company a growing concern.

  * * * *

  Figure 2. Writing an image slice by extruding material.

  * * * *

  The Stratasys company uses a different method—Fused Deposition Modeling—akin to the syringe method described above. Inside machines such as the FDM200mc (Figure 3) is a heated chamber in which two moving heads extrude melted plastic to form each layer of an object being printed. These objects can be quite complex, as seen in Figure 4. They can also be formed of several types of plastic in several colors. What is more, the use of melted plastic gives the items much greater solidity. According to the brochure, “Unlike parts created by competitive processes, the dimensions won't change and parts won't distort, which means you can use them in demanding applications.” They are thus well suited to making products with small production runs, as well as prototypes, demonstration items, and molds. More advanced Stratasys machines, such as the FDM Titan, can use more than one material at a time.

  What do these machines cost? The Zprinter 450 was touted as a price breakthrough when it came out in 2007—and it sells for a bit less than $40,000. The smaller Zprinter 310, which uses only one color of binder, costs about $25,000. Stratasys's FDM200mc sells in the same range. Stratasys's Dimension line starts at under $19,000 and goes up to $33,000. Such prices don't sound like much of a breakthrough until we note that high-end stereolithography rapid-prototyping machines can run $180-500,000, and selective laser sintering machines can run $270-325,000. High-end versions of fused deposition mo
deling, the Stratasys approach, can cost up to $300,000.

  Even the low end of this price range is a bit much for most home and small business budgets, and that's a shame. 3D printing has an enormous amount of appeal to anyone who likes to make things. Medical labs have used versions of the technology to print precisely shaped bone implants from bonelike material and hearing aids that precisely fit a patient's ear canal. Under development are rapid prototyping machines for dental labs and even dentists’ offices, where they could make crowns and false teeth. Dr. Stephen Schmitt, a dentist in San Antonio, Texas, is already using such a machine to make custom dental implants. A machine based on a modified inkjet printer has produced sheets of biodegradable gel with embedded cells; the eventual aim is to make on demand custom-designed tissues and organs for use in transplants. Researchers have even begun to develop techniques for “printing” skin and blood vessels. If everything pans out, the future of health care should be very interesting.

  * * * *

  Figure 3. The Stratasys FDM200mc. Photo courtesy of Stratasys.

  * * * *

  Considering the high prices of much medical equipment, the high prices of industrial 3D printers are no obstacle for health care. For the rest of us, these machines are out of reach. However, smaller machines with much smaller price tags are now available, and they will change the nature of the game.

  * * * *

  —

  Figure 4. Products of the FDM200mc.

  * * * *

  Taking 3D-Printing Home

  Rapid prototyping or 3D printing is a fascinating technology for two chief reasons. One, it gives people the power to make a great many things they now have to pay for (if they can find them). Two, because it prints things, not just pictures, it can print its own parts. The first 3D printer can then become two, which quickly become four, and so on. Costs become extraordinarily low, and the means of production spread quickly throughout society. Since some users will tinker with their machines to improve them, the best improvements will spread fastest, in a process akin to Darwinian natural selection.

  * * * *

  Figure 5. A RepRap 3D printer. Photo courtesy of Adrian Bowyer.

  * * * *

  This is the basic idea behind the RepRap project (reprap.org/ bin/view/ Main/ShowCase), whose founder, Adrian Bowyer of the University of Bath, says “the replicating rapid prototyping machine will allow the revolutionary ownership, by the proletariat, of the means of production. But it will do so without all that messy and dangerous revolution stuff, and even without all that messy and dangerous industrial stuff. Therefore I have decided to call this process Darwinian Marxism."

  The ultimate goal of the project is a von Neumann machine, a machine that can reproduce itself. The prospect of such machines makes some people nervous, for the one thing that seems likely to keep robots from ever taking over the world is their dependence on humans to make them. At the moment, however, the von Neumann machine goal is a long way off. RepRap machines such as the one in Figure 5 are made of many plastic parts the printer can produce, using the FDM technique, but many other parts must be supplied by a human being. These include metal rods, screws, motors, power supplies, and computer chips.

  The RepRap project makes available to all who would like to build their own 3D printer parts lists (the parts should cost less than $600) and instructions for building and programming the machine. And there's no charge. The project's motto is “wealth without money,” and it begins with the project itself. For future users, the motto means they will be able to satisfy a great many needs and desires without worrying about whether they can afford to pay for them. Eventually, it may mean that human civilization is built upon an industrial infrastructure that continually builds, rebuilds, and improves itself, without the need for investors.

  * * * *

  Figure 6. The Fab@Home fabber. Note the printed objects on the platform in the middle of the machine. Photo courtesy of Floris van Breugel.

  * * * *

  Unfortunately, the RepRap machines are not yet ready to take home. For that we must look to two other 3D printer projects, Fab@Home and Desktop Factory.

  * * * *

  Fab@Home

  In 2006, Hod Lipson of Cornell University launched the Fab@Home project with Ph.D. student Evan Malone. By January 2007, they were able to announce that their “Freeform"7 fabber—about the size of a microwave oven—could be assembled for about $2,400. The parts list, with a list of suppliers, instructions on how to build and operate it, and all the necessary software are available for free—from their web site (fabathome.org). As with the RepRap machine, the idea is that people should feel free to modify and—hopefully—improve the machine. Also like the RepRap machine, it uses a version of Fused Deposition Modeling, but in addition to plastic, it can print using PlayDoh, cheese, silicone caulk, plaster, chocolate, cake frosting, metal-impregnated plastics (for printing wires), and other soft substances that will harden quickly. I have talked to artists who would love to have one for use in making the wax cores for use in lost-wax casting. The one in Figure 6 has two syringes, which can use two separate substances or the same substance in two different colors.

  By the summer of 2007, it was already possible to buy fully assembled versions of the Fab@Home fabber from Koba Industries and Automated Creation Technologies for about $3,600. This is a great deal cheaper than any of the industrial machines, and though performance is not as good as with the industrial machines (the Fab@Home fabber is slower, and the surfaces of the objects it makes are not as smooth), with its ability to use many different materials, it has an astonishing versatility. It will only improve over time. Eventually ... Well, Evan Malone has already built a version of the machine that uses a rack of syringes and can make things out of several materials at the same time. He has used it to make a working battery, and his ultimate goal is to use his fabber to make a complete, working robot. If it can do that, of course, it can probably make another fabber, just as the RepRap folks intend. At that point, the fabber becomes the robot's womb and—just maybe—we will need to find some other way to keep the robots from taking over.

  In November 2007, Popular Mechanics gave the Fab@Home fabber a “Breakthrough of the Year” award.8

  * * * *

  Desktop Factory

  Idealab was started in 1996 “to create and operate pioneering technology companies.” In 2004, it gave birth to Desktop Factory (www.desktopfactory.com), whose goal “is to one day make 3D printing as common in offices, factories, schools, and homes as laser printers are today. Just as desktop publishing exploded as prices dropped and laser printers became ubiquitous, so too will new uses for 3D printing emerge as devices become inexpensive and widely available.” Their first product was in beta testing by the summer of 2007, and soon thereafter they began to take orders for delivery in 2008.

  The Desktop Factory 125ci 3D printer easily fits on a desktop and weighs less than 90 pounds. It can make things up to five inches on a side with layers a hundredth of an inch thick, slightly thicker than those laid down by the industrial machines. Speed of printing is comparable to that of the industrial machines. For raw material, it uses a proprietary plastic powder that can be fused by light from a relatively inexpensive halogen bulb to make things sturdy “enough to throw across a conference table."

  They claim to be offering the cheapest 3D printer on the market, but that is true only if they ignore Fab@Home, which is not entirely fair. The Fab@Home fabber may seem more like a hobbyist's toy, but it is on the market. It also has a very significant feature that the Desktop Factory does not: it can use multiple raw materials, none of which are proprietary. According to Desktop Factory CEO Cathy Lewis, they are working on a number of future improvements, of which this may be one. As it is now, Popular Science listed it among the best new technologies for 2007.9

  * * * *

  Limitations So Far

  3D printing technology is still a long way from being a science-fictional replicator that can make anyt
hing the human heart desires. For one thing, it can't make anything very big, except by making multiple small pieces that can be glued, screwed, or bolted together. For another, even a faster machine such as the Desktop Factory will take hours to make something. For a third thing, the surfaces are not as smooth as we are accustomed to from factory-made items. And for a fourth, they can print using only one or two materials.

  These shortcomings identify four obvious directions for future improvements. Changes in surface finish and speed seem likely to work against each other, for the finish improves as a printer renders an object as a series of more, thinner layers, and every additional layer takes more time to print. Improving the number of materials a printer can use is now under way, for a multi-syringe Fab@Home fabber has already been tested. Size will be difficult to improve without making the machines much more expensive, but keep an eye on Khoshnevis's house printer!

  * * * *

  3D Scanners

  An ordinary computer printer prints documents, meaning reports, e-mails, and pictures, every one of which is sent to the printer by the computer as a computer or digital file containing the information to be printed. 3D printers have to be sent similar files, though the information in them is a little different. The file has to describe a three-dimensional object in enough detail to permit the object to be recreated. It has to specify the position in three dimensions of every particle in the object's surface. It has to specify color, shine, and curvature. If there are internal details, it has to specify those. In the future, it may even have to specify what every particle is made of. The version of the file stored in the computer may not break the object into layers because one may wish to vary the thickness of the layers. A 3D printer can produce an object with thick layers faster, and that may be good enough for some purposes. Other purposes may demand thin layers and a better surface finish.