Rapid Prototyping (MCAD Modeling Column)

When you absolutely gotta have it fast!

Time to market is critical. Today's modern firms move heaven and earth to be the first, and it's not easy. The pressure is enormous. Design takes time. Tooling is a challenge. Marketing wants everything now. And the CEO is watching you.

Other than buying stock in an antacid company, how do you handle it all? Some available tools can make the process a whole lot easier for you. So, sit back, put up your feet, pop the top on a cold Mylanta and read on.

Designing for the Masses

You need 3D physical models for many reasons. Perhaps you recall the story I like to tell about showing 3D shaded images to a customer and so impressing him that he wanted parts shipped overnight. After explaining that the product existed only digitally, he was very disappointed. We kept the job, and it was a successful product, but it taught me a lesson.

I learned this same lesson other times. My colleagues and I designed a handheld product, and the image on the screen looked cooler than we'd imagined. It was going to sell big-time, we told ourselves. We sent the data to our model shop and within a week those master wizards turned out the full-scale model. When we got it in our hot little hands, we discovered that the grip was all wrong. It didn't feel right at all: for one thing, it was too heavy. We had to redesign the whole thing, which didn't make us happy.

Another time we designed a part and sent it out to the shop. When it came back it was almost microscopic. The machinists hadn't made an error; 3D models just look huge on screen. We had no idea the model was going to turn out so tiny. Suffice it to say, we could've saved ourselves a lot of time if we'd have invested in some RP (rapid prototyping) muscle. I have many examples of how RP could've saved my bacon, but they're too embarrassing to mention.

RP models are a great return on investment pretty much any way you look at them. You can sell an idea better if you show someone a model of it. It's really hard for someone to say, "That won't work" if a physical version is sitting in front of him or her. RP models are fast to produce, fairly cheap and usually fairly accurate. They can tell you when there's a problem you haven't caught, and that ability saves all sorts of trouble. But there is a bewildering number of competing RP technologies available. How do you choose which one is right for you? You have to get familiar with the technology and what it can do for you. Let's take a look at a few of the most common examples.

Stereolithography

STL (stereolithography) was the first true RP technology. It works by curing a light-sensitive resin in layers. Your computer model is split into multiple layers and each is traced on the surface of the resin. An elevator in the vat of resin lowers and more resin flows over the partially hardened layer. When the next layer is traced the elevator lowers again. Over time, your part is built up to its finished dimensions. This process allows you to build parts that are impossible to create using traditional machining methods, such as a completely enclosed but hollow sphere (although you need to allow whatever support material to drain). You can make snap features with huge undercuts very easily. But you're going to have to be very careful what you do. Unless you are going into production with the STL part directly, the production part will have to rely on the traditional limitations of the toolmaker. Just because you can model it and make an STL part doesn't mean it can be mass-produced or produced at all.

STL is relatively fast compared with traditional machining. A part that might be machined by a skilled machinist in a day might only take a few hours to build on an STL machine. Of course, complexity counts against the machinist. But accuracy and resolution go against the STL machine. Because the process is layer-based, a natural stairstepping effect occurs on the finished part. It can actually look pretty cool (figure 1), but most often it's unacceptable. You will need to swallow some of your time savings by going back and manually sanding and finishing your part. Also, resolution is an issue for low-cost machines. The going standard for STL is approximately ± .003"–.005". Some machines on the market now have resolutions as small as 16 micrometers. They're very impressive and fast, but they're not cheap.

Figure 1. Look closely and you can see the stairstepping between layers of even the most accurate STL part. The small support features are .020" diameter.

Another thing to consider when looking at STL technologies is the material you can use for your models. Remember, you are dealing with UV curing resins. They tend to be fragile, though they're better than those in days past. You can perform some functional testing on the STL parts, but they aren't going to be as strong as an injection-molded part. Another factor is the environment. STL machines typically need a controlled environment because the resins are hydrophilic. They suck water right out of the air, which means that the parts will distort on a really muggy day. They won't tie themselves into knots or anything, but the few percent length or width change may be a problem in some designs. And finally, you will need to knock off the support structures that carry the part's weight while it's building. This step will mean some extra finishing work and a bit of risk of damage to your part. You also usually need to bake the STL parts for awhile to finish the curing process.

Some flexible materials available for STL machines sweetly simulate rubber. But even the best of these will degrade and eventually break over time and use. I have found STL to be just the ticket for small, highly detailed parts.

Selective Laser Sintering

Close on the heels of STL is SLS (selective laser sintering). This technology works almost identically to STL but uses powdered materials in place of liquids. Your model is laser drawn atop a thin layer of plastic powder, which fuses, or sinters, the particles to one another. The elevator moves down, and the first layer is buried under another layer of powder, which is in turn fused and so on and so forth, until your part is finished. The surface finish on an SLS part is somewhat grainy because there really isn't a good way to control the flow front edges on the melting plastic. An advantage is the ability to create multicolored models, which is a major selling point (figure 2).

Figure 2. The ability to mold in different colors in the same model enables a lot of design capability. If you can model it or wrap an image around it, it's doable. The surface finish can be an issue, though.

The parts aren't super-rugged, so they are mostly for touchy-feely applications. If they sell your designs, however, these machines are worth their weight in gold! SLS machines aren't what you'd want for creating extremely tiny parts.

Fused Deposition Modeling

FDM (fused deposition modeling) is another additive process that competes with STL. It works by extruding a tiny molten bead of plastic and drawing the layers of your model. The process is not unlike caulking your bathtub. FDM machines start with the STL file format to input the 3D model. Cleanup is easier, and the machine's environment isn't as big an issue. The materials are actual engineering materials. The downside is that the finished part is very porous. It won't have the same density as an injection molded part. You can perform some functional testing with FDM parts, but they won't be as hearty as production parts.

Generally, FDM machines are very fast. Because FDM parts are made of actual engineering materials, they don't need postcuring. When the machine switches off, the parts are ready to be handled. FDM machines use two different materials when building parts: the modeling material and the support material. You can get water-soluble support materials that come off very easily in a heated water bath. Some FDM machines even produce metal parts (figure 3). Tiny parts can be an issue for FDM, but as long as you keep your expectations reasonable, FDM units are great RP machines!

Figure 3. FDM allows you to use more actual engineering plastics than STL. Some FDM machines even let you build in metal.

Laminated Object Modeling

LOM (laminated object modeling) is an older technology. I always mention it in the hope that this technology will see a revival. I just like it—even with its inherent problems. LOM works by laminating large sheets of paper and cutting the model's layers out of each layer. No supports are necessary, but the cut-out material must be removed before the next layer is adhered, which means that LOM isn't the fastest RP process. Why do I like it so much? It's well suited for larger parts such as the housing of a piece of exercise equipment. The finished part has the characteristics of wood. It is surprisingly strong but obviously vulnerable to moisture. You can paint it for protection or finish.

Computer Numerical Control

What can you say about CNC (computer numerical control)? Unlike all of the preceding technologies, it's a subtractive modeling method. You start with a solid block of whatever material you want and mill away what you don't. It's been around forever. It can produce and reproduce identical parts all day long with a surface finish that's hard to beat (figure 4).

Figure 4. You just can't seem to beat a machined part for surface finish. Accuracy also is a CNC strong point.

Tool changers and other options on some of the newer machines let you "set it and forget it." Obviously, you can create production tooling or even parts with a CNC machine. It excels at fine detail, but you're limited by what cutters you have and what the machine can see of your part. If you try to make undercuts, you may have to get creative.

Conclusion

RP is a great technology to have at your disposal. Before you buy such capability, though, you need to avail yourself of service bureaus to learn what to expect. Look into RP. It's worth your time. You'll find it one of the best modern design tools you'll ever see.

Mike Hudspeth, IDSA, is an industrial designer, artist and author based in St. Louis, Missouri.