In today’s competitive marketplace, time to commercialization is key. Thermoplastic rapid prototyping machines like those produced by Stratasys build up detailed models by extruding a continuous bead of liquefied plastic in layers as fine as 0.005 in. An x-y gantry propels the extrusion head in a series of raster moves, depositing a layer on the build platform (z stage), letting it cool, then applying the next layer to form the finished prototype. To build machines robust enough to stand up to the kind of continuous operation required by rapid prototyping departments and services bureaus, Stratasys turned to compact, long lifetime, high-torque motors from MICROMO.
As the first step in thermoplastic prototyping, Stratasys converts plastics ranging from acrylonitrile-butadiene-styrene (ABS) to polycarbonate into filament that is wound on spools and sealed in protective canisters. The actual deployment of the filament is a two-part process. First, a motor in the material bay pulls the filament from the canister and feeds it up a channel to the extrusion head. At the extrusion head, another motor pushes the filament through a liquefier that heats the plastic as high as 450° C to soften it enough to be extruded in a fine bead.
It’s an exacting process, says Paul Leavitt, systems engineer at Stratasys. “It has to be precision controlled because we need to extrude the exact amount of material in exact timing with the gantry. The pump control has to be synchronized with the gantry motion so that the bead stays a constant width.”
The x-y gantry is positioned by NEMA 34 brushless servo motors. For the filament feeding and extrusion, Stratasys uses brushed servo motors from MICROMO. It’s a demanding application. The high-speed moves made by the gantry dictate that the extrusion motor be compact and lightweight. In addition, the motors need to supply high acceleration, along with the flexibility to handle up to 25 different types of plastic, ranging from stiff and brittle to highly pliable. The motors also need to be robust. “It isn’t just a motor’s ability to translate electrical energy into torque,” says Leavitt, “it’s got to last. The extrusion device is on a gantry that gets pushed around at one or two G’s, so it’s got to be able to survive all those shock loads, and it's got to do that for years and years.”
The machines are subjected to punishing duty cycles: typically 50 to 60% utilization, with 4000 to 5000 hours of operation per year. An average job runs eight to 12 hours, but with the advent of machines with build envelopes as large as 3 ft ´ 2 ft ´ 3 ft, job windows are stretching. “In the past, a build of 24 hours was a pretty long part,” Leavitt says. “Now, with these big machines that we’re putting out, a seven day part is not out of the question. We've got customers doing 90% utilization, so they’re using it 7000 hours per year.” That’s an average of nearly 20 hours a day, 365 days per year, so the motors have to be tough.
Each machine features an extrusion head with two liquefiers, one to supply material for the model and one to supply material for any support structures the model requires during the build. Each liquefier is delivered filament by a pair of canister motors synchronized to allow automatic changeover from one canister to the other, allowing the machine to run without interruption. The filament passes between an idler wheel and the servo-motor-powered serrated feed wheel that pushes the filament up an 8 ft channel to the extrusion head.
The canister motors are slaved to the extruder head motors, which hand down current commands. The constant current places the filament in compression so the head motor does not have to do any work to pull the filament. Initially, the design used a plastic spur gear and a MICROMO coreless DC 2230 motor but the assembly could not stand up to the extreme demands of the job. The engineering team turned to the more powerful model 2224 DC motor, fitted with a steel planetary gear. “We went to a much higher factor of safety,” says Leavitt. “Typically they size [the motors] for 2000 hours of continuous use but we want five to 10 times that. How do you do that? You’ve got to de-rate the motor, basically.”
With the new gearhead in place, the filament feeding process became robust. The extrusion head presented problems of its own, however, with motors failing after only four to six months of operation. At first, the engineering team thought the issue was an insufficiently powerful motor but analysis by the MICROMO’S Test Group showed that the problem was actually rooted in premature failure of the gearhead, which put excessive stress on the motor and caused it to fail in turn. Further investigation, though, revealed the real culprit: early breakdown of the lubricant. “It was a really lightweight oil that would break down and turn into varnish in two or three months,” says Leavitt. “Pretty soon it was just gone.”
MICROMO installed a high-pressure, high viscosity synthetic grease in the gearbox that stood up to the wear. Perplexingly, tests showed no change to the failure rate. Part of the issue was the new grease -- it protected the gearhead but it added significantly more resistance. “The Klubersynth made our gearboxes last about five times longer,” says Leavitt. “The downside was that it took three times the current to drive that grease. The earlier motor just couldn’t do the job any more.”
The real problem went beyond the grease, though. The real problem was the application itself. “We've learned that we can't just design to meet motor requirements because we’re not working with a continuously operating device,” says Leavitt. To create a part, the gantry must make raster moves, going forward then reverse many times a second. When the gantry reverses, the extrusion motor must at the same time accelerate the filament in the reverse direction to keep the plastic bead from coming out, then rapidly accelerate it forward when the gantry advances again. “When that happens many times a second, it's a very difficult thing for a motor to do,” says Leavitt. “It's not just sitting there spinning at a constant velocity under a load, it's full-current acceleration forward and then full-current acceleration in reverse. Those extrusion motors really get beat up.”
The solution was switching to a bigger motor. The model 2224 motor used initially features precious metal brushes that are designed to minimize electrical noise. For some applications, that is essential, but the Stratasys team didn’t care about electrical noise -- they needed lifetime, and the strong copper-graphite brushes of the coreless 2342 DC motor gave it to them. “We went to the coreless DC 2342 motor and wow, that was the solution,” says Leavitt. “Those precious metal brushes are like fine wispy hair and the copper graphite brushes are probably a hundred times the mass. You can put a lot more current through the motor without it breaking down.”
Once they’d chosen the new motor, they were able to change to a gearbox with a 66:1 reduction ratio. This improved responsiveness by decreasing the amount of force applied to the plastic for a given acceleration.
Testing it out
The next step was accelerated life testing in a 60° C oven, subjecting the assembly to one cycle of 10 raster segments (forward and reverse) every 1.3 s. Leavitt ran tests of 500, 1000, 2000, and 4000 hours, the last of which is equivalent to five years (11.2 million cycles). “We got wear analysis for every one,” says Leavitt. “It was pretty impressive. At 4000 hours, [the gearmotor showed] only 10% wear. The gear teeth still had sharp edges on them. The commutator was hardly worn at all and the brushes looked brand-new.”
In the first nine months after deployment, they only had three failures but that wasn’t good enough for Leavitt. “We went from 15 to 30 failures a month to three in nine months, so it was much better, but we’d done all of our life testing and said, ‘This thing can't fail, it’s just impossible. What's going on?’” What was going on was a size mismatch. The MICROMO coreless DC 2342 motor is 42 mm long compared to the model 2224’s 24 mm length. As a result, the encoder on the back end of the motor was projecting out enough to be hit in normal shop-floor operations.
MICROMO’s dedicated machine shop went to work, shortening the flange, the motor shaft, the pinion, and the screws to reduce the length of the overall assembly. They also added a protective cap over the encoder so that any impact got transferred to the motor housing. In addition, they tapped holes into the motor shaft and added custom cabling harnesses for the motor.
The result was a success. Overall, the motor changes have resulted in a robust, responsive solution that allows prototyping houses to run their machines at astoundingly high duty cycles without fear of downtime. “To date, the coreless 2342 DC motors have performed very well with few field issues,” says Leavitt. “I don't know how long they're going to last, maybe 10 or 15 years.”
According to Leavitt, any additional expense for the larger motor is more than offset by the new opportunities it has opened up. “It's got more capability from a velocity point of view,” he says. “We’re replacing [the old motors] in our old products but also in our new products we use that capability to go faster, so that's a benefit. Meeting the performance requirements has enabled us to develop our FDM 400mc and FDM 900mc, products geared toward direct digital manufacturing (DDM). DDM has higher reliability, throughput, and accuracy requirements than rapid prototyping.”
“It takes a true partnership between customer and motor supplier to optimize standard products into customized solutions,” says MICROMO applications engineer Michael LeBlanc. With the help of MICROMO, Stratasys has developed thermoplastic prototyping and DDM machines that can stand up to the toughest workloads day after day, year after year.