Since the introduction of practical rapid prototyping (RP) about 20 years ago, this technology has become an important part of the design development process in many industries. RP benefits are well established. Prototype parts can be quickly produced at low cost for validation, measurement, and in some cases actual trial. RP can be a major cost-saver because it allows the designer and manufacturer to see what a part will look like in early development stages. It also allows design changes or product cancellations when such moves are least expensive, particularly with highly complex designs.

More recently, the basic prototype-making technology has spawned two other manufacturing aids, rapid tooling (RT) and rapid manufacturing (RM). RT provides methods for rapidly manufacturing molds while RM involves techniques for low-volume part production.

Some industry observers see these three concepts (RP, RM, and RT) as ultimately taking two paths. First is the communications aspect. It involves those RP techniques that show what parts will look like, and stress the faster, lower cost technologies. The second, more related to RM and RT, involves the manufacturing side with parts that are stronger and more precise.

Designer creates program that the 3-D system can use to create a prototype.

Toolmaking. RT produces molds in two ways, either by forming a mold over an existing part (indirect) or creating a mold from a CAD file (direct). For example, the 3D Keltool rapid moldmaking process from 3D Systems (Valencia, CA) is used chiefly for die casting and injection-molding applications. The manufacturing process begins with a CAD file that describes the products. Using stereolithography, a pattern is made. Next a transfer mold is made using silicon rubber. The mold is filled with a slurry of tool steel and tungsten-carbide powder along with a epoxy binder. After curing, then firing in a hydrogen-reduction furnace, the metal is sintered and binder burned off. In a second firing, powdered copper infiltrates the part and gives it full density. The final part is 70% steel and tungsten and 30% copper. Resulting tools can last one million shots or more. The entire process takes eight days.

Big benefits are short time-to-market and the ability to quickly change product design. Currently size is limited to 5.9 X 8.5 X 4" (150 X 216 X 102 mm) or a total of 144 in.³ (2360 cm³⁾. Shrinkage is 0.6%, because it is linear and isotropic. This can be compensated for with an oversize master pattern. The process can reproduce details down to 0.04 mm, and has a flatness of ±0.025 mm per 25 mm. Maximum operating temperature is 650° C.

Model is "printed" in a matter of hours for evaluation.

Another popular toolmaking method is Rapidtool from DTM (Austin, TX), which is used to make complex molds. The process begins with a CAD model of the cavity. This model is converted to an STL file and sent to a Sinterstation system to make the mold insert.

In the machine, a laser fuses powered metal particles by melting a polymer coating on the particles. This binds the stainless steel particle into a porous metal skeleton.

Once design is approved, a more durable model is made using a 3-D stereolithography process.

The parts are transferred to a reducing atmosphere furnace where the binder decomposes and the steel powder sinters to form bridges between particles. At this stage the part is 60% dense. In a second furnace cycle, bronze infiltrates the mold producing a fully dense part that may be capable of producing 100,000 molded parts.

Currently the company is investigating a new metal material for plastic injection molding.

Part making. RM is still in its infancy and its successes have been in low-volume niche markets, such as medical parts or space-related items. Two factors driving this industry are the growing demand for low-volume production and product customization.

Final prorotype

Some predict RM will ultimately be as simple as running a fax or Xerox machine. A user will just enter a CAD file and a part would appear. The US Navy is particularly interested in this idea because much of the space aboard a ship is taken up by part storage.

This project is said to be already underway. Some ships have machines from 3D Systems that generate part copies which can be used to make investment casting molds.

The space station program is another case where a production run of one will be essential. Being able to create parts from a single RM machine would save a lot of space and weight.

One apparently successful RM company, Fusion Models (Fresno, CA), makes molds for low-volume customers. In one case they were able to produce four complex molds needed for a medical appliance in four weeks.

Boeing is another example of RM success. They use DTM model 2000 and 2500 Sinterstations. Most of the work is in low-volume part manufacture. "People don't realize how many opportunities are out there for low-volume production," says Roger Spielman, Rapid Prototying Lead Engineer, Rockdyne Power & Propulsion, Boeing Co. (Canoga Park, CA). "It's a big advantage to be able to cut time and expense from the typical test, fail, re-design, re-test, and optimize loop."

Boeing engineers are currently making small parts (Sinterstation size) for the International Space Station, the Space Shuttle Main Engine Program, and several aircraft programs. All of these parts are end-use, qualified, and man-rated components.

"Our Selective Laser Sintering (SLS) metal program is in the final development phase for several new rocket engine systems, and is projected to replace castings for some of the parts, including rotating hardware [fuel and oxidizer impellers]," explains Spielman.

For the parts made of polymers, the company stays with material supplied by DTM. Their original work for the space station and shuttle was based on a nylon composite (LNC-7000) material. Some of the newer work will be based on DTM's Duraform products. "We are currently performing test and evaluation on these materials which should lead to flight certification status. Polymers are similar to ABS with strengths in the 7 ksi [48 MPa] tensile range," says Spielman.

On the metal side, Boeing has successfully sintered many different materials, but are focusing on high-temperature, high-strength superalloys (nickel-based generally). Many configurations, from injector bodies weighing up to 18 lb (8.2 kg), injection molding dies, forming tools, pump components, to impellers, are being tested and evaluated. Superalloys are approaching the properties of wrought steel with a tensile strength of 7 ksi tensile at 25% elongation.

"All parts meet or exceed casting tolerances. I generally look for an easy target, usually ±0.005" (±0.13 mm)," concludes Spielman.

Material progress. All RP techniques, particularly those that are well established, are placing a new emphasis on materials. According to a 3D Systems spokesperson, "Initially we stressed how well our machines worked, now the emphasis is on materials. In fact, material properties are influencing how we design our platforms."

Eventually, the major thermoplastics such as ABS, polycarbonate, and nylon, will offer mechanical and thermal properties that can be used for even more demanding applications.

Software simplification. One of the initial problems with RP was converting CAD files to the form that most RP machines needed to convert a design to a prototype. This was often the job of a system specialist and a production engineer. Programs have now greatly improved and the conversion is much easier. The developers are still striving for "pushbutton" simplicity.

Product improvement. Many of the established RP companies are now stressing incremental improvements. For example, Stratasys Corp. (Eden Prairie, MN) uses the Fused Deposition Modeling (FDM) process that creates 3-D parts of ABS, wax, elastomer, and polyester compound directly from CAD files.

The company recently offered their model FDM3000 with a build envelope of 10 X 10 X 16" (254 X 254 X 406 mm). But the major innovation is a new support removal system for use with an ABS modeling material called WaterWorks. This system allows virtually "hands free" production of concept models, precision prototypes, and tooling patterns and masters.

With WaterWorks, a completed model with supports is immersed in a water-based solution and, after a brief period of time, the supports simply wash away. The result is a clean model with smooth surfaces.

Low-end success. "I believe there is a strong divergence in the market. At the high end are those working on RM and RT," says Marina Hatsopoulos, CEO of Z Corp. (Somerville, MA). "We, on the other hand, are working in a much larger computer peripheral market. Our units are inexpensive, easy to use, and designed strictly for part presentation. Accuracy and strength is not an issue because it's an appearance model."

There is no need for high-cost equipment for this work, and speed is one of the system's biggest advantages. It generates parts in a matter of hours, and is based on well-established ink-jet printer technology. "Basically, we lay down a layer of powder, then use the ink-jet to deposit a fluid that links the powder together," explains Hatsopoulos. "The part materials couldn't be simpler: plaster or corn starch and sugar. However, we are working with our suppliers to generate more diverse materials."

Two new products. As an example from the user's side, Bastech Inc. (Dayton) has successfully completed several rapid tooling projects. One large manufacturer enlisted the company's aid to produce a set of small, complex inserts for injection molding. Using DTM's RapidSteel 2.0 they completed a 3 X 3 X 2" (76 X 76 X 51 mm) set of inserts in 16 days. This timeline included engineering of the cavity and core geometry from a 3D model, building the inserts on a DTM SLS 2500 Sinterstation, (15 hr), running through two 22-hr furnace cycles, polishing, mounting in a base, adding core and ejector pin systems, and injection molding. (The customer had to wait over four additional weeks to benchmark the mating parts produced with traditional prototype tooling methods.)

In a second project, a recreational equipment manufacturing company needed tools to bring a new product to market in a short timeframe. The market segment was a new one for the company; and in addition to producing parts, the tooling would be used to justify further investment in the marketplace. Bastech created a two-cavity, fully automated tool, and a seven-cavity family tool in about seven weeks. First-article parts were produced at this time, then an additional four weeks was taken for customer engineering modifications, mold texturing, and plating operations.

The customer is running these tools today and they have produced over 40,000 parts. The thermoplastic material being molded is glass-filled nylon. With this material, the plan is to run these rapid tools up to 100,000 parts. At that point the revenue generated by the rapid tools will fund multicavity production tools. By using Bastech's rapid tooling solution, this customer was able to enter and establish itself in a new marketplace (one season early), make real-time tooling modifications (based on customer input), and eventually fund traditional production tooling.

Composite part casting. Basic RP technology may expand well beyond its current limits, and combine with other advanced technologies. For example, a manufacturing technology effort supported by the Air Force Research Laboratory's Materials and Manufacturing Directorate at Wright-Patterson AFB (Dayton, OH) may improve the fabrication of metal-composite parts.

A metal-matrix-composite casting process being developed by Metal Matrix Cast Composites (MMCC) Inc. (Waltham, MA), converts CAD programs into high-quality, net-shaped finished products in a matter of days.

The new process expands on conventional RP, allowing the manufacture of durable parts for engines and brakes at nearly half the weight, and at a competitive cost. This task is especially difficult in the manufacturing of composites, where non-recurring expenses such as engineering and tooling can be very high. The new process achieves uniform dispersal of ceramic particles, and can selectively tailor sections of the part being cast to specific customer specifications. Unlike conventional casting processes, it can utilize a multitude of material reinforcements and alloys with many "architectures," or shapes, which produce a broad range of choices for designers and engineers.

The first step is machining carbon molds to hold the porous ceramic preforms to be cast, then the molds are loaded with an alumina, boron carbide or silicon-carbide preform, which is placed into a sealed container. Next, the container holding the casting is evacuated and heated, and molten alloy is transferred to the casting vessel. The can is pressurized in an autoclave to drive the molten alloy into the porous preform region, and the casting is then cooled and the part extracted and finished.

Engine block from Z Corp. uses ink-jet technology to build prototypes.

This process can take from several days to a few weeks, depending on complexity and scheduling considerations. Parts produced by this process have an excellent surface and, since parts can be cast to near-absolute net shape, they require little or no secondary machining. In some cases, the mold is constructed around a preform, not unlike conventional sand casting.

Parts currently made by this process include connecting rods for two-stroke outboard marine engines, brake calipers, water-cooled brake disks for heavy trucks and aircraft tow vehicles, brake caliper pistons, brake rotors, push rods, racing-bicycle pedal cranks and circuit board heat sinks.

Possibly this work will lead to a means for developing cost-competitive, metal-matrix-composite products that can replace steel and other high-density materials. Compared to parts manufactured by traditional processes, it yields products that weigh only about half as much, are quick to manufacture, and are cost competitive due to accelerated design and engineering time.

For the RP industry, about the best thing that can be said about the 1990s is that they're over. The first decade of commercial availability of additive fabricators has seen tremendous technical advancements, global proliferation of the technology, and widespread initiation of academic research programs. Yet these solid accomplishments have been accompanied by lackluster sales for the machine vendors, painful price wars for the service providers, and dismal financial performance by the handful of companies that have put themselves up on the stock market.

For manufacturing engineers that use fabricators to design and develop new products, the problems in the industry have been a mixed blessing. Vendors are desperate for sales and willing to cut deals at every turn. Prices for outsourced fab services are very affordable. Yet the doldrums of the industry do not play well with the senior management that approves the purchase of new capital equipment. And depressed unit sales make it difficult for vendors to push machine prices down into the neighborhood that would allow middle managers to purchase the equipment on their own. It's a vicious cycle.

Against the background of a financial market generating huge fortunes with unprecedented speed for companies conducting any manner of business on the Internet, the fabricator industry has been struggling to stay afloat. Despite tremendous gains in price/performance, ease of use, and quality of materials, sales have been lackluster and market appreciation of the technology has been disappointing.

The cycle is similar to what happened with personal computers in the early 1980s, when some doubted the future of that industry. But like computers, fabricators are too valuable a technology to remain down for long. It is difficult to know whether the new up-tick will come from new technology, a new personality, or a new surge in demand. Most likely, it will be a combination of all three. In the meantime, there is much technology that manufacturing engineers can take advantage of today.

The most exciting new applications are showing that RP is definitely not just about prototyping anymore. Among the new applications are medical modeling, aerospace parts, and building construction.

Dr. Marshall Burns
Ennex Corporation
Los Angeles, CA

Stark contrasts and conflicting trends have characterized the RP industry during its brief history. Dominating its early boom were innovation, numerous successes, strong investor interest, a large pool of early adopters, and a mindset that "anything is possible."

An environment marked by many opposing trends gradually replaced this dynamic one. The industry posted a lower-than-projected growth, investor interest waned, and innovations declined rapidly. Today, the technologies driving the RP industry show an overwhelming resemblance to the base technologies that launched the industry. Through my role at the National Center for Manufacturing Sciences (Ann Arbor, MI), I interact repeatedly with hopeful start-up RP companies that come to me at the end of their rope, having been unable to attract the interest of venture capitalists whose attention is dominated by e-commerce opportunities.

While this disturbing trend needs to be addressed, the existing ranks of RP companies do show some improvement. These more stable OEMs are refocusing their attentions on new inventions and are well equipped to exploit them. The road to recapturing the imagination of a marketplace that has turned its attention elsewhere is lined with the type of head-jerking innovations that originally launched the industry.

Ken Johnson
Program Manager
National Center for Manufacturing Sciences
Ann Arbor, MI

Bastech Inc DTM Corp Fusion Models Metal Matrix Composites Co 3D Systems Thyssen Production Systems Inc

Z Corporation