"The Start of a New Movement," Feature Article, Aug. 2002

the start of a new movement

A NASA invention is one of several poised on the brink of commercialization. But it needs outside help to get there.

By Paul Sharke, Associate Editor

It isn't often that engineers who design machine components get to speak about new architectures. That's something left to the designers of sexy computer chips, not for practical nuts-and-bolts folks. But "architecture" is exactly the word John Vranish sprinkles into the conversation as he's describing his latest invention. The NASA engineer has found a novel way of combining two fundamental machine elements—gears and bearings. With it, he may indeed be looking at the construction of a new way of moving things.

Four years ago, Vranish's bosses at the Goddard Space Flight Center in Greenbelt, Md., asked him to come up with an "over-achieving" planetary speed reducer, something that could step down a motor output dramatically in a simple and lightweight form. It seemed they had many small mirrors they wished to move on the next space telescope, but they weren't necessarily willing to pay the weight penalty that using many conventional reducers would invoke.

What Vranish came up with doesn't look much different than any ordinary planetary transmission at first inspection. It has planets and rings and a sun gear. But there is a difference. It has no bearings—not in the conventional sense of pairs of rolling-element components supporting every gear.

Instead, Vranish has combined gear and bearing into a single machine part he calls a "gear bearing." In its simplest form, a gear bearing sandwiches a crowned spur or helical gear between two cylindrical rollers. The gear and the rollers share a centerline. The radii of the rollers equal the pitch radius of the gear.

By giving up its need for rolling-element bearings, a gear-bearing planetary transmission achieves a light, compact speed reducer. Rollers handle radial loads while tooth ends take thrust.

When the teeth of one gear bearing roller are meshed with the teeth of another, their respective bearing surfaces touch also. The gear teeth transmit mechanical force while countering thrust. The roller takes the radial loads.

Vranish displayed two of his planetary reducer prototypes at the recent National Design Engineering Show in Chicago. With the first, he achieved a 70:1 reduction in a single stage. With the second prototype, Vranish employed "phase tuning" to build a 325:1 reduction. More about phase tuning in a moment.

If only spur gear teeth make up the gears in mesh, the thrust-carrying capacity is limited, Vranish explained. In that scenario, the top surfaces of the teeth carry any thrust loads as they roll along the bottom of the roller. Where thrust loads are large, gear bearing architecture lends itself to the use of herringbone, or double helical, gears to take advantage of their superior thrust capacity.

Exploded view of the planetary transmission depicts two ring rollers. The colors here correspond to those in the front view, shown on previous page.

Although NASA's telescope application needed a lightweight, compact, super-reducing planetary gear train, the basic elements that make up the transmission are incorporated readily into a number of other motion devices, including slides, reversers, linear-to-rotary converters, and even speed increasers. The same approach can handle a doubling of speed or a 2,000:1 reduction in a single stage, along with everything in between.

Vranish said that the gear bearing assemblies eliminate sliding friction even under load. Devices constructed from gear bearings encounter rolling friction only. Another advantage gear bearings have over conventional roller bearings is their elimination of carriers normally employed to keep friction-reducing elements apart.

Obvious differences crop up immediately upon placing a gear bearing differential transmission alongside a conventional planetary transmission. First, of course, the new reducer eliminates eight ball bearings needed to support the planets and suns in a three-planet reducer.

Two different load paths distinguish the reducers as well. Loads travel through the conventional reducer by way of the relatively weak bearings. In the gear bearing reducer version, the load goes from output ring to upper planet to lower planet to ground ring, a much stronger path, according to Vranish.

Nuance and Performance

In addition to producing huge speed reductions in small spaces, gear bearings list smoothness and precision motion control among their attributes. Smoothness, because they reduce micro-chatter of ordinary gear and bearing combinations. Precision, because they eliminate a rotational wobble built into many transmissions.

Elements in the mechanism underlying micro-chatter are the small size and loading differences that distinguish one ball or roller from another. Another contributor to the phenomenon is the carrier that spaces the rolling bearings evenly around the race. Together, the inconsistencies between individual bearing-element speeds and their fixed spacing can make the larger balls or rollers overdrive the smaller ones, causing skidding, or micro-chatter, along the raceway. If no carrier is used, faster bearings collide with slower ones, making matters even worse. By eliminating balls, rollers, and carriers, gear bearings suppress micro-chatter.

According to Vranish, gear bearings self-adjust to eliminate the slight assembly misalignments that come with installing bearings between shafts and housings. Bearings are installed within fairly close tolerances: too tight, and they'll squeeze the balls into premature wear; too loose, and the component will perform poorly at reducing friction. Sometimes, the bearing cocks slightly during installation, leading to shaft wobble. Gear bearings, by avoiding any secondary assembly, reduce chances for misalignment.

But these are traits that Vranish called "nuances" of the new design. His enthusiasm built as he spoke about several high-performance techniques, such as phase tuning and rifle true anti-backlash, which work with gear bearings.

Turning a prototype by hand produced a smooth output and a great decrease in speed.

In explaining phase tuning, Vranish said that a typical planetary reducer with three planets would require at least a difference of three teeth between the ground and output rings. That's because the planets would be spaced equally around the ring in 120-degree increments. Each tooth on the planet ordinarily has to engage a full tooth between input and output rings, leading to the minimum required three-tooth difference.

Phase tuning changes that. To enable their assembly, the input and output sides of the planet gears are cut as separate elements. Because of this, the gear sides can be rotated fractionally with respect to each other, then fastened together. The possibility of creating a difference of only one tooth between the ground and output rings thus opens. Vranish has proved the concept with a rapid prototype model.

Rifle true anti-backlash produces a planetary transmission with zero backlash. To visualize it, imagine the top and bottom halves of a planet gear bearing. The top one consists of a shallow helical gear and bearing machined from a single length of stock with an internal hex running through it. The bottom half also has a gear and bearing, as well as a mating hexagonal stem from which the points have been lopped off. Enough clearance exists for the upper gear bearing to slide over the stem.

First, though, drop a pair of back-to-back Belleville springs over the stem.

Now, nest this gear bearing pair, along with two other identical pairs, between the rings and sun of a planetary transmission. With the transmission unloaded, note that the spring pushes the upper gear and bearing into the helix of the ring until it can go no further. It's taking up the backlash. Under load, the hexagonal stem and its mating hexagonal hole generate ample frictional locking force to hold the gear teeth against counter torques. If the teeth wear, the Belleville springs take up any additional backlash as soon as the load is removed.

Technology Transfer

Of course, gear bearings don't interest everyone. They were merely one example of technologies unveiled at the Chicago show by various NASA agencies. One inventor from the Kennedy Space Center discussed new data acquisition devices and instruments; another, from Marshall Space Flight Center, discussed a quick-connect ball and swivel joint that had been used for satellite docking.

Many NASA inventions provide opportunities for commercial developers to work with the various agencies through licensing or cooperative agreements. According to Darryl Mitchell, commercial technology manager for Goddard, the idealized version of technology transfer puts a private sector firm to work furthering a NASA invention. Be it through licensing or collaborative agreement, the commercialization firm develops a product that not only benefits the private sector, but returns technology improvements to the agency. Such "dual use" technologies pay off in many ways: first, by solving a particular NASA problem; then by advancing a technology in the commercial domain; and, finally, by providing NASA access to innovations that stem from the original invention.

Through phase tuning, a one-tooth difference between ground and output rings makes extraordinary reduction ratios possible.

Although NASA has been in the technology transfer business for many years, the program, of late, has gained more emphasis in the light of shrinking budgets. Increasingly, NASA sees the value inherent in its technologies. At the same time, "The public expects more bang for its buck," Mitchell said. For instance, it's not enough anymore for NASA to merely acquire scientific data on the rings of Saturn, he explained. The public expects tangible benefits.

Perhaps the greatest success for the program was the commercialization of Nastran, the NASA structural analysis system whose development began in the 1960s. Today, MSC Software of Santa Ana, Calif., sells a proprietary version of NASA's finite element analysis software.

NASA technologies enter the commercial sector by one of two main routes, Mitchell said. The simplest, straight licensing, is more or less a matter of qualified companies filling out paperwork, he explained. More complex collaborative arrangements are hammered out with "Space Act" agreements, which define the product to be commercialized, divide responsibilities among parties, and decide on the handling of intellectual property.

(The National Aeronautics and Space Act of 1958, amended Oct. 30, 2000, formed NASA and encouraged commercial development of space, while devoting space activities to the peaceful benefit of mankind.)

As an example, Goddard is actively seeking partners to share in the commercialization of gear bearings, Mitchell said.

Meanwhile, Vranish has taken the invention about as far as he needs to for meeting the agency's requirements on a future space telescope. Now is the time to bring in outside experts who understand manufacturing, or marketing, or any of the many steps that must be negotiated on the way to commercial product success.

As for new architectures, you get the idea that a hands-on guy like Vranish might be a little less comfortable using words like "vision" than he is with terms such as "reduction ratio." But he admits that, after spending as much time as he has with the little gear bearings, he's begun to see a range of motion problems where they might work just fine—and perhaps work even better than what's being built today.