Rapid prototyping for pattern and foundry tooling.

Changing the landscape of all of manufacturing, this technology has vastly streamlined conventional foundry patternmaking.

Rapid prototyping, the process of producing three-dimensional (3-D) models directly from computer-aided design (CAD) images, has been in practice for the last decade. As more developments have come to the forefront, it appears that this process will have a great impact on the metal-casting industry. It is now possible to make prototype castings with a reduced lead time and at less cost as compared with previous methods.

In the metalcasting industry, rapid prototype parts are now used to create patterns. A rubber mold can be produced from the rapid prototyped part. A duplicate polymer model can then be produced. In some cases, the rapid prototype part itself is used as a pattern to produce metal parts. For complicated parts, the rapid prototyped part can be used like a pattern in the investment casting process.

Diminishing skills in the pattern-making and modeling trades are making rapid prototyping more and more attractive. Some feel that this process will make inroads in the patternmaking trade in the future.

This article discusses how rapid prototyping can be used to aid in the parts development process by streamlining patternmaking. It describes how three of the most popular rapid prototyping methods - Stereolithography (3D Systems Corp.), Solider (Cubital, Inc.) and Laminated Object Manufacturing, or LOM, (Helisys, Inc.) - can be used to obtain a pattern. In addition, examples of how patterns have been produced by these various processes are included.

Applications

There are many reasons for using rapid prototyping. An automotive company executive recently noted that two of his firm's priorities were to cut development costs and shorten development time. Both of these objectives are achieved through rapid prototyping.

In comparing rapid prototyping to the conventional approach to making parts, three development areas in which rapid prototyping can be put to use will be addressed. They are design verification, preprototype tooling and proto-type-production.

Design Verification

Design verification was the market that early rapid prototyping pioneers targeted to complement developing CAD systems. When comparing costs that are the direct result of design verification as well as optimization, potential savings are best calculated on a case-by-case basis.

Design verification savings using rapid prototyping vary by firm and application. For certain "high risk" products, design verification has been practiced for many years.

In this area, after a 2-D CAD drawing is designed, a 3-D model is produced to verify if the design is correct. In the near-term, the opportunity exists for a low-cost, design verification system. Conventional methods for design verification have been manual and numerical control (NC) machining of aluminum and machineable plastics.

Preprototype Tooling

Preprototype tooling can be described as silicone robber or epoxy tooling for cast polyurethane parts. Preprototype tooling generally can be viewed as low-cost materials for a limited number of patterns that would otherwise never stand up to a foundry environment.

As companies learned and began to experiment with rapid prototyping, many needed more than one or two models from the rapid prototyping systems. This created a natural opportunity to build silicone rubber tooling for additional parts, as well as parts from this soft tooling that would have a wider range of mechanical properties for fit and function and some physical testing.

The cost comparisons now become even more significant as many firms begin to slowly replace traditional prototyping procedures with soft tooling. This requires material experience and assumes the product validation is low risk based on historical material performance.

Prototype-Production

A recent development with the use of rapid prototyping patterns for tooling is the direct result of a breakthrough with epoxy resins. In the past, the acrylics had a high shrink rate and coefficient of expansion. Therefore, early work with rapid prototyping outputs for investment castings provided marginal results. Often, the shell would crack due to the pattern expansion and leave a high ash content within the shell.

Epoxy resin can be used in this application to build accurate patterns for investment castings. This tooling process utilized a rapid prototyping pattern as a tooling insert for A-2 or H-13 tool steel castings.

Stereolithography

Stereolithography is the most widely used rapid prototyping method. It takes a solid CAD model and forms the model by solidifying the liquid polymer from a vat with a laser.

The laser beam cures the polymer by exposure to the laser ultraviolet light. The beam, which is directed only to areas that are solid on the CAD model, can solidify a layer of 0.002-0.010 in. As the layers are being made, the object is lowered in the bath of liquid polymer. Thus, it is easy to make complicated objects without coring.

Figure 1 illustrates how QuickCast patterns (stereolithography models produced for direct use as patterns) were used to create investment casting tooling for a water shield housing that was to be produced by injection molding. Conventional kirksite tooling from a wood pattern costs about $33,000 and took 17 weeks. QuickCast tooling, created from an A-2 cast tool, was produced in just six weeks at a cost of $18,000.

The benefits from this example are production materials, further development of investment cast tooling and high potential for prototype to production requirements. More development must be done in this area, however, to prove a repeatable process. Further projects are now in place for using QuickCast patterns for aluminum prototype tooling.

Solider

This method produces prototypes from what is called a Solider machine. Across-section image of the CAD model is sent to the mask generator. A mask plate is charged through an "image-wise" ionographic process. The mask plate develops the CAD cross-section image by charging the plate with electrostatic toner.

The base material is prepared in tandem while the mask is being made. The material is made by spreading a thin layer of photopolymer of any viscosity. It is then placed under the prepared mask. A strong flash of ultraviolet light passes through the mask. All the exposed areas in the layer (0.004-0.006 in. thick) are cured.

After the exposure, the part is wiped and any residual, unsolidified material is vacuumed off. At the next station, wax is filled in all the cavities to replace the removed photopolymer. The material is then cooled to solidify the wax.

Next, the layer is milled to an exact thickness to produce a flat solid surface so that it can receive the next layer. Finally, a thin layer of photopolymer is spread on the material. It is then ready for exposure. When the part is built, the wax surrounding the part is melted or rinsed away. The wax can be retrieved and used again.

This process can make many parts at the same time. Parts can be nested together in the machine's envelope. Additionally, this process needs no final curing. Shrinkage and distortion are minimized. Finally, by building the part in a solid environment, no artificial support structures are needed. The current envelope for this is 20 x 14 x 20 in. Dimensional accuracy is 0.1% up to 0.02 in.

LOM

LOM performs like a 3-D laser plotter. To make parts, the machine sequentially bonds layers of sheet material such as plastic, paper or composites that are coated with polyethylene on one side. These materials range in thickness from 0.002-0.020 in.

The layer of material is placed on top of the previous layer. A focused laser beam cuts the material to the CAD generated outline. The laser cuts around the periphery of the computer-generated cross section of the part. Hot rollers laser melt the polyethylene and bond the layers together. This method of making parts is fast and produces no internal stresses.

One Midwestern pattern shop specializes in making patterns that produce castings for the aircraft and hydraulic pans industries. These castings can be complicated - with some having as many as 25 cores.

Making patterns from 2-D drawings can be laborious and shapes can be misinterpreted. Considerable time and expense can be saved by seeing the part in 3-D before further processing is done.

Because the pattern shop produced tooling for complex castings and creating the masters was very time-consuming, it began exploring rapid prototyping three years ago. Its customer base was also designing in 3-D more and more, which made rapid prototyping more feasible.

In comparing rapid prototyping to traditional pattern-making and CNC machining, the shop decided rapid prototyping was the fastest, most economical way to produce a part.

When the shop started experimenting with rapid prototyping, it used service bureaus to create parts. During this process, the shop experienced a couple of major failures along with a few successes.

Integrating rapid prototyping into the business was a fairly smooth evolution. At first it was apprehensive about bringing in a new and competitive method of producing patterns. After working with this new technology for awhile, the apprehension disappeared - much like it did when CNC machining was introduced.

Eventually, the pattern shop purchased its own LOM system so it could guarantee control of deliveries and quality.

An example of the pattern shop's experience with LOM is illustrated in Fig. 2. Used as part of an aircraft fuel pump, this casting requires 25 cores to produce.

Using the computer, the customer made a stereolithography (STL) file for the exterior and an initial graphics exchange standard (IGES) file for the interior. To create the interior, the IGES information was downloaded to a CAD system. Toolpaths were created and the pipe cores were machined on CNC machines.

The exterior was created using LOM directly from the customer's STL file. This part is a good example of cutting lead times. This exterior would have taken 3-4 weeks using conventional methods. With LOM, it was produced in about 2 weeks.

The final product was aluminum and epoxy tooling which was used to produce the sand molds and cores.

In this example, LOM models were created and sent to an investment foundry to produce six stainless steel castings. This enabled the customer to purchase the castings without having to go through the expense of building tooling. The models were burned out of the molds just like wax is normally done.

Another example shows the great advantage of having an actual part to hold and examine instead of only seeing it on the screen. Shown in Fig. 3 is a small impeller that is used to pump oil, and its LOM model. Normally, the pattern shop would have machined this pattern on a five-axis milling machine. Due to scheduling conflicts, however, the pattern shop was forced to find another method to meet the delivery, and decided on LOM.

Earlier, after producing an LOM model of the impeller, the engineer examined the part. Luckily, he noticed that the blades were going in the wrong direction. Since no drawings were used, nobody would have caught the error. The problem occurred in the design stage when the IGES file was created.

The engineer was then able to take this part back to engineering so that others could review the final product. Once an acceptable model was produced, a robber mold for reversing was made, A hard epoxy part was created out of iron-filled epoxy. In production, this part is now a steel investment casting.

Quicker, Cheaper

In comparing rapid prototyping to conventional prototype methods, significant savings can be recognized both in cost and timing. Ongoing innovation with technical processing now permits companies to integrate product development through to real-time production. What began in 1988 as visual design verification is now a tooling aid.

Further, it can be seen that a continual direction toward lower-cost design verification tools and higher-cost, highly accurate tooling helps the entire metalcasting process. The cost savings are significant.

Foundrymen and patternmakers are clearly involved in technology that has the near-term potential to change the manufacturing landscape. As these tools become more readily available and processes refined, there will be a natural transition into production validated materials.