Thermally conductive plastics beat the heat

With electronic devices gaining power and losing bulk, design engineers have really started to feel the heat. Thermal management once meant a few well-placed fans, vents, and aluminum heat sinks. But today's small, hot electronics can benefit from new ways to keep cool-like using thermally conductive plastics.

"Why bother?" you might ask. After all, thermally conductive plastics don't come cheap. Per-pound material prices run anywhere from 5 to 50 times more than aluminum. And these highly filled thermoplastics, the best of which offer a thermal conductivity that tops out at about 100 W/mK, still don't beat the 80-160 W/mK offered by diecast aluminum. Yet in the right applications, conductive thermoplastics can still compete against metals with much higher thermal conductivity.

Just ask David Coutu, president of Intelligent Motion Systems (IMS) and the company's chief engineer. For a new motor with an integrated microstepper drive, Coutu first designed the combination enclosure-and-heat sink in diecast aluminum. Later on, he switched to a thermally conductive PPS from Cool Polymers Inc. (Warwick, RI). "Aluminum may have a higher thermal conductivity, but our application really lent itself to the benefits of plastics," Coutu says.

A fitting application. Even though the PPS compound IMS used has a thermal conductivity of just 20 W/mK, the new enclosure nearly matched the thermal performance of its aluminum predecessor. According to Coutu, both the plastic and the aluminum managed to meet IMS' most important thermal target by keeping the housing temperature under 60C with the motor running at its top speed and torque. In terms of the absolute temperature difference, the plastic housing runs only two to five degrees hotter than the aluminum. "The difference between the two materials was negligible," Coutu says.

The reason why goes right to the heart of what makes the IMS application a good fit for thermally conductive polymers. As Cool Polymers Engineering Manager Mikhail Sagal points out, total heat transfer (AT) encompasses not just conductive, but also convective and radiant heat transfer. "Many electronics applications aren't conduction limited," he says. "But they are convection limited." And IMS' stepper motor, the MDrive 17, fell squarely into that category. "Conduction wasn't the bottleneck in getting the heat out," Coutu says, explaining that the new polymer housing matched the convective cooling capability of its aluminum predecessor because both had about the same surface area.

Sometimes, plastics can even outperform aluminum because their inherent design freedom can help engineers increase the surface area of cooling components-by adding more fins or other thin-walled features. "In these cases, a plastic part can easily surpass aluminum," Sagal says.

Aside from thermal performance, plastics can help meet other design goals too. The housing's PPS compound is also electronically conductive to contribute to shielding requirements. And while the MDrive doesn't take advantage of plastics' strong suit in parts consolidation, Coutu notes that this capability will be important in other applications. On an upcoming controller, for example, he plans to eliminate an internal heat sink and vents by employing a thermally conductive polymer in the housing. "Not only do we eliminate a component, but we improve shielding by getting rid of the vent," he says.

Analyze it. To see if thermally conductive plastics would work in the first place, IMS and Cool Polymers started out with an extensive modeling and prototyping program. First, engineers at Cool Polymers used CFD software to model the enclosure in both diecast aluminum and a thermally conductive polymer. "Our main concern up front was that we not lose any thermal performance," says Coutu. The results showed that plastics, in theory, could transfer enough heat. Just to be sure, IMS and Cool Polymers engineers went on to machine enclosure prototypes out of stainless steel, which has roughly the same thermal conductivity as the polymer enclosures. Physical testing on this model confirmed the analysis results.

All of the CFD modeling had an added benefit of helping Coutu take a systems approach to designing the MDrive-by allowing him to optimize the efficiency of device's electronics in light of the enclosure's heat-dissipation capabilities. "Thermal management for us was a two-pronged strategy," he says. It involved not just the housing design but also an effort to minimize the heat from the device's integrated drive and power supply. At the end of the day, the MDrive motor had to be about 40% more efficient than a comparable stand-alone motor. But without the thermal analysis work to tie the electronics and housing design together, that efficiency target could have been even greater, adding cost.

Keeping a lid on cost. Once Coutu knew that a plastic enclosure would work, he focused on the economic justification for using a material that, at first glance, costs more than aluminum. IMS, however, actually lowered its total costs by about half when it went to plastics, according to Coutu. "Even the best stepper motors are a basic product, so it was very important that we keep costs low," he says.

To offset the price disadvantage of thermally conductive polymers, Coutu picked a material thermally conductive enough for his application-but not more. "We picked the most thermally conductive polymer we could without crossing the line to a much more expensive material," he says. That strategy meant choosing a PPS that offered a thermal conductivity far lower than the 100 W/mK that Cool Polymers can supply.

The bulk of the cost reduction can be traced to manufacturing advantages. According to Coutu's estimates, the MDrive's relatively low production volumes would have made the initial diecast housings twice as expensive as the plastic ones, mostly due to tooling costs. "Ramping up to diecasting is much more expensive than injection molding," he says. Investment casting, another option, operations would have required expensive secondary machining operations in order to meet Coutu's aesthetic goals or the housing's plus or minus 0.002-inch dimensional tolerances. "Plastics gave us a smoother ride to tight tolerances and good aesthetics than metal could," he says.