Some users are discovering that combining direct digital manufacturing with existing manufacturing techniques is better than replacing them
By Bruce Morey
Wouldn’t it be great to get an order and ship out a completed part with little pre-planning or manufacturing engineering required beyond plugging a CAD file into a machine that spits out the finished part? Especially if it’s a single machine that produces an unlimited variety of parts. That is the promise of direct digital manufacturing (DDM) or “rapid manufacturing.” Size, surface finish, and material limitations remain, though system developers are continuing to improve this technology. Manufacturing engineers are just beginning to understand how to use it. The fact that it’s so radically different from current, established manufacturing techniques also means that engineering knowledge needs to evolve before DDM can achieve its full potential.
Guided by data from a CAD (or related STL) file, DDM methods range from ‘printing’ plastic parts to more recent metal-based processes such as laser sintering, laser melting, or electron beam melting. Development is now going into technologies that produce parts intended for service, according to Terry Wohlers, president of Wohlers Associates Inc. (Fort Collins, CO). Metals (especially titanium) in particular are under development, he says. They are intended for functional prototypes and finished products. As reported in Wohlers Report 2007, an industry study on the subject, worldwide revenues from DDM (as contrasted to machines sold for prototyping) were 3.9% of the total in 2003, a figure that grew nearly three fold to 11.7% in 2007.
“Direct digital manufacturing is best used when production volumes are relatively low and part complexity is high,” says Wohlers. “The four areas where DDM is satisfying current market needs are aerospace, motorsports, medicine, and dentistry.” He believes that more industries, such as automotive and consumer electronics, will embrace DDM in the future as the materials and machines that it requires improve.
One class of such machines, called Powder Bed machines, builds up a part by spreading a thin layer of powder metal in a bed, and then sintering (melting) select sections before depositing the next layer on top of it. The finished part emerges from a loose volume of powder that is recycled. Part sizes for powder bed machines tend to be limited to one foot in any direction, comments Wohlers. The size of the powder particles determines resolution—think surface finish.
EOS Inc (Novi, MI), supplies the EOSINT M 270 for metals, which uses their Direct Metal Laser-Sintering (DMLS) process. According to Jim Fendrick, a director with EOS, DMLS currently can produce parts from bronze-based alloys, tool steel, stainless, cobalt-chromium, and titanium. “These materials are especially useful in the medical, dental, and aerospace industries,” says Fendrick. “Both parts and tooling can be made with the M 270.”
Using a 250-W fiber-laser fixed in the roof of the build area, the M 270 provides a build volume of 250 x 250 x 215 mm, and a build speed, depending on materials, of 2–20 mm3/sec. Powder thickness ranges from 20 to 40 µm. The M 270 can complete a part in as little as a few hours, though large or complicated parts may take a day or more. Material choices also affect speed.
Fendrick agrees that there is increased interest in advancing the technology from prototypes to in-service parts. “Industry is seriously looking at rapid manufacturing right now,” says Fendrick. “The customers are pulling us into it, even while we are pushing the technology forward.
“Direct and rapid manufacturing is a disruptive technology,” Fendrick observes. “It inspires people to think differently about how to design and build parts. We reduced the part count for a blood centrifuge from 32 parts to just two parts. Even if the part costs more, the time savings may be worth it.”
Still, Fendrick says that there are routine market challenges for early DMLS adopters, particularly in aerospace, where the industry has lengthy qualification processes for part certification.
Another powder bed technology, Electron Beam Melting (EBM), forms the core technology for machines made by Arcam AB (Mölndal, Sweden), which are marketed in the US by Stratasys (Eden Prairie, MN). Electron beams fuse the metal powder in a vacuum chamber. Arcam currently advertises three metals: titanium Ti6Al4V, titanium Ti6Al4V ELI, and ASTM F75 cobalt chrome. Thicknesses of layers laid down by the Arcam EBM S400 can vary from 0.0019 to 0.07″ (0.048–1.78 mm), according to Fred Fischer of Stratasys. “Where accuracy is critical, postprocess machining is generally required.” The build volume is 200 x 200 x 180 mm. The A2, also offered by Arcam, varies from the S400 primarily in its build volume. Two different build tanks are offered: one is rectangular and has a build envelope of 200 x 200 x 350 mm; the cylindrical tank has dimensions of 300 mm in diam by 200 mm.
Another approach to building up a part from layers of material in a powder bed is to deliver powder by nozzles directly to the point where a focused laser melts the powder, fusing it into a part line-by-line, layer-by-layer. Called Powder Deposition, the technique typically offers larger working envelopes and the ability to either make parts or repair existing parts.
“I describe it as free-form fabrication,” says Richard Grylls, product manager of the Laser Engineered Net Shaping (LENS) system provided by Optomec Inc. (Albuquerque, NM). Offered commercially since 1998, LENS was originally developed at Sandia National Laboratories. The LENS 850-R system offers a working envelope of 900 x 1500 x 900 mm, positional accuracy of ±0.25 mm across the working envelope, and linear resolution of ±0.025 mm, according to Optomec. The included tilt-rotate table offers angular resolution of 0.01°. Spatial resolution of the features deposited using LENS can vary between 300 µm and 1 cm, depending on the laser power used. Using a 1 or 2-kW fiber-laser provided by IPG Photonics (Oxford, MA) as standard equipment, it can deposit material at up to 200 cm3/hr.
“The LENS process is housed in a chamber purged with argon such that the oxygen level stays below 10 parts per million, to ensure there is no impurity pickup during deposition,” says Grylls. The control head for the laser and powder metal deposition has three-axis linear X, Y, Zmotion control, while a tilt-rotate table gives the system five-axis control. Two additional axes are available through an optional pitch/yaw wrist control. Optomec also offers the smaller LENS 750 with a 300 x 300 x 300-mm working envelope, with similar accuracy and deposition rates.
“Not only can we build a part from scratch, we can repair existing parts by adding thin layers where parts are worn or damaged,” says Grylls, noting that, at present, repair applications are more popular than building original parts. This capability is especially useful in military and aerospace applications, where parts may have been in service for decades, and the original design data and tooling may be unavailable. LENS also provides value repairing manufacturing defects in high value parts, thereby improving yield and reducing cost. Enabled by the fiber laser, Optomec has built a special DeepRepair head specifically to reach into tight spaces and repair non-line-of-sight areas.
Another company that uses powder metal deposition is the POM Group (Auburn Hills, MI). They enhanced existing laser cladding technology by adding precise motion-control. The company’s systems use a sophisticated combined laser and powder metal head. Cameras mounted on the head closely monitor melt-pool height. A patented feedback control system, using information from the cameras, modulates laser energy to maintain the melt-pool height at predetermined levels. This keeps deposition to the desired dimension, and limits the amount of distortion from layer to layer, according to J. Mazumder, CEO of the POM Group.
“We designed our systems for building and repairing heavy molds and dies,” says Mazumder, “which is why we developed the fully five-axis head. It was not practical to move these heavy molds on a two-axis bed system.” Feeding up to four separate powders into the melt-pool, end-users can experiment with metal alloys.
POM Group offers two standard machine configurations: the larger is called the DMD 505, which offers a 48 x 23.6 x 24″ (1219 x 600 x 610-mm) working envelope. Equipped with a 5-kW CO2 laser from Trumpf Inc. (Farmington, CT), it boasts a deposition rate of 0.6–9 in.3/hr (2.73–40.9 mm3/sec.) The smaller DMD 105 offers a 32 x 20 x 15″ (813 x 508 x 381 mm) working envelope with a 1 or 2-kW fiber coupled-diode laser with positioning accuracies of ±0.001 in/ft (0.098 mm/m) in X, Y, and Z and 0.015° in B and C. Powder thicknesses, which control the ultimate resolution and finish of the part, range from 40 to 100 µm.
The POM Group also reports that repair of existing parts, molds, and dies comprises the largest part of their business, along with precise coating of conventionally made parts enabled by the precision of their machines.
The POM Group is building a third machine based on an articulated-arm ABB IRB 4400-60 robot (from ABB Robotics, Auburn Hills, MI) rather than a Cartesian CNC machine. A diode laser from Laserline (Los Gatos, CA) provides the power. “The advantage of the six-axis robot is its much larger working envelope. While it sacrifices some of the accuracy we have on the CNC-based machines, it has a 2-m cubic working envelope, and will be truck-mounted,” says POM Group Vice President Bill Swenson. “We are building this for one of the automotive OEMs to do die-repair on-site.” With dies weighing up to 50 tons (45.4 t), according to Swenson, the truck-mounted robot repair machine offers the maximum mobility option for die repair. Swenson reports that this process provides one-tenth the heat-affected-zone of traditional welding techniques used to repair such dies, with material properties very close to or even better than wrought properties.
The hybrid nature of the repair process has led Optomec’s Grylls to speculate that hybrid manufacturing may be the most effective use of systems that employ “free-form fabrication.” He envisions a future machine shop that uses CNC machinery to cut shapes of parts that then have features, like bosses or flanges, added. “With the rising costs of certain metals, like titanium, and the availability of these metals continuing to be tight, a hybrid approach is going to be a real benefit to these additive types of manufacturing processes,” says Grylls.
Hybrid manufacturing seems to be precisely how a small, innovative company, Linear Mold and Engineering (Livonia, MI), is building their business. Founded four years ago to provide prototype parts to the automotive industry, it now provides metal prototype parts to a variety of industries, as well as tools and dies used both for prototype and production parts.
To build the capacity for expansion, they knew they needed to provide parts that were more complex. They needed to go beyond the capabilities of their existing three-axis CNC machines. “While we were looking for either a small five-axis CNC or EDM machine to help build our tools and dies, we discovered the DMLS process and purchased an EOS M-270,” explains Geoff Watz, general manager for Linear Mold and Engineering. They now have two EOS M-270s. Besides delivering metal prototype parts, they also use their M-270 to ‘grow’ complicated inserts for use inside tools that have been milled using their stable of three-axis CNC machines.
“Our three-axis machines provide a cost-effective way to build most of the tool. We cut pockets inside the tool in complex areas. While it’s being machined, we grow tool inserts with the M-270 for those places where the tool is complex. This used to be done primarily by EDM, but our technique of using tool inserts grown by DMLS is more cost-efficient,” explains John Tenbusch, the company’s president and CEO. It saves time, because the inserts can be grown in parallel while the rest of the tool is on the CNC machine, whereas with EDM the process was sequential. They note that some EDM work is still required, but not enough to justify the expense of owning their own equipment. They outsource what EDM work they need.
Experienced users of DMLS, they have not found any significant material or size limitations in the DMLS process. They report that they use three metals provided by EOS extensively: Direct Metal 20 (a bronze alloy), 17-4 stainless, and a cobaltchrome. Surface finish has not been an issue, with most parts meeting surface finish requirements after benching. “As for size,” says Tenbusch, “the current build envelopes have not been a problem for us using our machines efficiently. It’s not practical to build larger parts anyway, given the current growth speeds.”
Linear Mold is keeping their DMLS machines busy. Although the DMLS investment was originally made for their die-mold business, a lucrative benefit of their machine is making direct-metal prototype parts, which has become a growth business for them. They have produced everything from prototype aerospace turbine engine parts to prototypes of bone and joint implants, as well as new surgical-tool concepts. “When we’re not using the machines to make metal test parts for outside customers, we’re using them to make mold inserts,” explains Tenbusch.
“We are just scratching the surface on using DMLS,” says Tenbusch. “This is a part of our business that is going to keep growing.”
This article was first published in the February 2008 edition of Manufacturing Engineering magazine.
Published Date : 2/1/2008