A mere 17 years ago this month, stereolithography debuted at the Autofact show in Detroit. This ushered in rapid prototyping (RP), the process of creating physical models out of plastic directly from computer-aided design (CAD) geometry files. Gone would be the clay, wooden, and metal models that took lots of time and more money to produce. Back then, trade show attendees all stood in awe as cute, little gears and propellers and chess pieces and dinosaurs and so many more things essentially grew in the RP machines—right before our eyes. Better than pens, rulers, and other trinkets handed out at trade shows, attendees coveted these semitransparent or opaque prototypes out of plastic resin.
A lot has changed since then. The number of RP vendors has expanded and then consolidated. The number of RP materials has grown to satisfy all sorts of applications. More impressive, RP is no longer just for small parts, and RP parts are even being used in final assemblies. (Although those clay, wood and metal models have yet to disappear.)
Fit to be tried
Hyundai Mobis (Seoul, Korea) got around the limits of RP size when it needed to check for mechanical interferences and evaluate the airflow in a dashboard it was designing for a Kia being introduced in North America at the end of this year. Needless to say, car dashboards are larger than your average chess piece. Much larger. The new Kia dashboard measures 20 in. x 18 in. x 54 in.
Mobis has been using RP for about three years now on such projects as modeling ABS and airbag cases. These prototypes help Mobis verify part designs, mold patterns, and the master model of a mold, according to Tae Sun Byun, principal research engineer at the Autotech Div. of Hyundai Mobis. In the past, Mobis would have milled a prototype of the dashboard in plastic. While effective, producing an accurate prototype for testing would have taken more than 20 days. Much of that time would have been spent cutting out the complex backside surface of the dashboard. Explains Byun, some of the walls in the dashboard measure only 1-mm or 2-mm thick, and some of the cavities in the dashboard are deep and difficult for a cutting tool to reach.
So, Mobis deployed an FDM Maxum, an RP machine from Stratasys Inc. (Minneapolis, MN), to make the dashboard out of ABS plastic. This material, says Byun, "gives us the durable parts we need for assembly and functional testing." The designers had a prototype dashboard ready for testing in less than a week. It was an assembly of four pieces glued together. The dashboard model was mounted on a fixture and inspected with a coordinate measuring machine. The greatest deviation in the RP model was just 0.030 in. over its entire length of 54 in. In the next phase of part verification, the dashboard was mounted in a Kia cockpit mock-up. This revealed a few mating problems between the initial dashboard design and the ventilation ducting and related subassemblies. Sensors in front of the exhaust ports in the dashboard confirmed the effectiveness of the ventilation system.
All this checking identified 80 design problems—before the dashboard's release to tooling production. The designers went back to the CAD system to make the appropriate changes. Byun estimates the company saved more than $70,000 through RP. "If we did not make the RP dashboard," he says, "we would have had to pay that money to fix the finished injection molding tool used for the dashboard."
Parts on demand
S&S Cycle, Inc. (Viola, WI) is an OEM and aftermarket supplier to motorcycle enthusiasts. About five years ago, it decided to take its RP work in-house by buying a Vanguard Selective Laser Sintering system from 3D Systems (Valencia, CA). (3D Systems was the company that debuted RP at Autofact in 1987.) Since then, S&S has been using its RP machine to make prototypes, to help in making molds, and to make one-offs for actual races.
For example, using OneSpace Designer Modeling, a 3D solids modeling system from CoCreate Software, Inc. (Fort Collins, CO), S&S engineers and designers applied a shrink factor to a 3D model of a new crankcase. This model was then used to produce a RP master that perfectly replicated the aluminum engine part as it would have come out of the mold. S&S's casting vendor then rammed loose sand around the RP plastic model, removed the model, poured in aluminum, and made engine parts they can test before actually committing to tooling. These are mostly destructive tests. S&S puts the parts in motorcycles and engine dynamometers and runs "the absolute hell out of them," says Eric Wangen, Concept Designer for S&S. But in the process, he continues, "we have simultaneous engineering and testing going on, rather than test at the end design cycle and hold up the sale of our product."
This is how S&S learns about its newly designed parts—before tooling is ever made. Wangen admits that without RP, he doesn't know how S&S would shorten product development from concept to production. He doesn't relish going back to the "old way": six months design, six months testing, and another six months or so to make tooling. And when all is done, S&S may still need to make modifications to that tooling.
Sometimes RP is used directly to make tooling. For a new cylinder head, S&S used RP to make a head for a sand cast mold that was used in production until a low-pressure permanent mold in steel was ready 26 weeks later. This provided two broad benefits. First, S&S was able to work out any glitches in the mold design before committing half a million dollars to the final mold for high-volume production. RP is "smart money," says Wangen; "all the changes to the mold design were done before we started making tooling." Second, S&S could produce 1,000 or so aluminum engine parts from the sand-cast mold, using it as a bridge between low-volume and high-volume manufacturing. This "bridge tooling" let S&S ship product six months earlier than expected.
But here's a big win for RP: Producing finished parts. The gearheads that race bikes are constantly changing something, tweaking a little here or there, looking to get more power out of their bikes. Last year, at the Mopar Parts Mile-High Nationals in Denver, the team behind rider Dave Feazell changed the location of the port where it comes out of the cylinder head. They sent the head to S&S, which designed a new manifold in 48 hours. Another day or so was spent using RP to make the new port out of 3D Systems' DuraForm glass-filled material. This material is resilient enough to withstand the violent vibrations that are the norm in bike racing. Because the intake system of the manifold was running cool air, all Feazell's team had to worry about was the resident heat after the engine was shut off. For that, S&S applied silicone around the manifold ends to insulate the part and keep it from melting.
Feazell's team got the part on his bike in time for the qualifying rounds. The RP part held its own; Feazell shaved over 0.003 seconds off his race time.