Small Wonder: The Magnificent Ball Bearing
On an August afternoon in 1943, a long column of American B-17s arrived over the Bavarian city of Schweinfurt. The bombardiers watched the city crawl by under the lenses of their top-secret Norden bombsights, made delicate adjustments, and, from four miles up, dropped eighty bombs into a mile-square target. The remarkable accuracy of their bombsights was due partly to the dozens of tiny and precise ball bearings in each one. Hundreds more ball bearings went into the bombers’ big radial engines. The importance of ball bearings to the war had not escaped the planners of this raid—their target was the Reich’s largest complex of ball bearing factories. The Army Air Force lost 36 airplanes out of 194 and more than 350 men in this attempt to cripple German war production at the heart of its machine-tool industry.
Ball bearings are still as important. Were someone to take away the ball bearing and its next of kin, the roller bearing, America would grind to a halt in a shower of sparks. You trust your life to them when you drive your car. Bearings for jet engines are so vital to safe flight that they rank as one of the most intensively quality-controlled products in our economy. Ball bearings keep computer disk drives turning smoothly. Bearings machined to tolerances equal to the thickness of a fingerprint spin in the hearts of the inertial guidance systems that steer our airliners and intercontinental ballistic missiles.
In the last hundred years, five products—the bicycle, the automobile, machine tools, the gyroscope, and the jet engine—have pushed ball bearings to new extremes of accuracy and efficiency. The balls rotate between two steel rings (the “races”), allowing the races to turn so smoothly that it seems the bearing has a waiver from the laws of friction. A “cage” of thin sheet metal keeps the balls neatly distanced from one another. The primitive predecessor to the ball bearing, the simple journal bearing, is exemplified by a greased wagon hub. It is sticky, wobbly, and a nightmare at high speeds.
Ball bearings were used in turntables under statues as early as about A.D. 40, and by the end of the eighteenth century, inventors were employing them in carriages, water turbines, and windmills. Still, the ball bearing lay in obscurity until after the Industrial Revolution, stymied in part by the lack of inexpensive steel. With steel costly and even rare, the few ball bearings in use before about 1860 were typically of cast iron, which was irregular in composition and brittle. Under stress the balls would split. After the invention of the Bessemer process and-the pen-hearth furnace, steel became widely available, but the ballbearing industry would still not emerge until a product appeared that could use the bearing to advantage. That product proved to be the bicycle, which when perfected, about 1885, would meet a tremendous demand for fast, individual, non-horse transportation.
The advantage of ball bearings over the simple greased hub for bicycles became obvious when a velocipede equipped with them won an international race in 1869. Within a few years, bicyclists were reaching speeds of nearly twenty miles per hour on ball-bearing hubs and, by the 189Os, bicycle ball bearings were all built on the same common plan. The balls ran around a track formed by a cup and a cone. The bearings were indifferently sealed, though, and water and dirt entered to abrade the parts. The grit “lapped” the balls smoother but also smaller, so most designs allowed the rider to adjust the cone inward as the balls wore down. Higher-priced models had cups and cones cut by lathes from tube or bar stock; cheaper models had cups stamped out of sheet steel. After heat treating to harden them, the races were brought to size against powered grinding wheels.
The grinder was vital for finishing the bearing parts, but the development of good abrasives and grinders was at first held back by the mistaken notion that grit from abrasion would lodge in the steel and later ruin the bearings. As the grinder came into wide acceptance in the 189Os, there remained uncertainty about the best way to form the initial rough, round shape of the balls. The Cleveland Machine Screw Company used two techniques. One was to crush steel bars between hemispherical dies with power hammers; balls would emerge in beadlike strings that workmen separated with chisels. The other was with a surprisingly automated and accurate lathelike device called a screw machine. It used cams to control different cutting edges, and some screw machines could control ball diameter to within a thousandth of an inch.
Either way, all the balls went through the same steps after rough-forming: a series of twin-disk abrasive grinding machines took them down to size. Operators stopped the machines now and then, pulled out a few balls, cleaned them, and measured for size. When the balls were sufficiently small and smooth, they went into furnaces for case hardening. Baths cooled them and relieved internal stresses, and revolving oak barrels tumbled them against one another for a final polishing. At a factory in Philadelphia, a final lapping, during which balls were rotated between disks in a compound of oil and fine emery powder, could bring tolerance to better than a ten-thousandth of an inch.
Between 1897 and 1914 the bicycle industry collapsed, defeated by the rise of the automobile. Autos brought their own tremendous demand for ball bearings, but until World War I most automobile bearings came from Europe. The likely reason is that tried-and-proven designs for bicycle bearings did not work very well in cars. Europeans moved ahead with new designs, using twin rings with curved inner surfaces to better support the balls. By wartime the use of bearings had spread far beyond cars and cycles to baby carriages, naval gun mounts, electric motors, sewing machines, fans, typewriters, gyroscopic compasses, and machine tools.
Bearings for machine tools represented the ultimate challenge to postwar ball-bearing manufacturers: the tools’ shafts and spindles had to spin with complete firmness under heavy loads and at high speeds. Many lathes still turned on simple oil-lubricated journal bearings; ball bearings were too loose at high speeds. This shortcoming was a major spur to better designs and manufacturing techniques, and by 1930 the problem had largely been overcome. Lathing of balls had nearly vanished in favor of forging—lathing was too slow, and forged balls were stronger. Steel was now hardened clear through, not just on the surface.
The technology was sophisticated— 1930-era techniques of forging, machining, hardening, and grinding are still followed today, and the steel used for most bearings is the same chrome alloy, SAE 52100. The alloy is prepared differently today though, with much more attention to purity—it is in such fine points that most further progress has been made, and seemingly minor changes have multiplied the durability and strength of the ball bearing many times while allowing it to be built ever smaller and to run ever faster.
As World War II approached, the inventor Winslow S. Pierce, Jr., son of a New York attorney, did the seemingly impossible. At age fourteen Pierce had built his first auto transmission, cutting the gears himself. He dropped out of Yale to pursue a life of invention and perfected an important new ball bearing in 1921 using split races. In 1936 his heirloom pocket watch, a Waltham, stopped running. Told by a repairman that a cracked jewel could not be replaced, he designed and built the world’s smallest ball bearing to replace it—just one-sixteenth of an inch in outside diameter, complete with ten balls.
Word of this achievement spread rapidly among the designers of precision machines. In 1940 Pierce founded Miniature Precision Bearings, Inc., of Keene, New Hampshire, and some of the company’s first bearings went into the Norden bombsight. The war increased business for all ball-bearing manufacturers— Strom Steel Ball Company was rolling out sixty million balls a month by V-J Day, and one four-engine bomber could require as many as thirty-four hundred balls in its bearings. The war also brought greater attention to quality control. The inspection of bearings dates back to the early manufacture of bicycles. In those days ranks of young women at the Cleveland Machine Screw Company rolled the gleaming balls back and forth across trays of shellacked pine looking for the occasional misshapen ball. One observer called the work “simply maddening to any but the strongest eyes.” The balls passed to a sizing station, where they then rolled down a track made of two diverging blades, each ball dropping into the appropriate tube when it reached the width corresponding to its diameter.
Bearing makers recognized the importance of sizing early on. If one ball in a bearing is much larger than the others, it will carry a disproportionate share of the weight and will therefore crack sooner—bearing life depends very strongly on weight carried, since the pressure between a ball and a race can reach hundreds of tons per square inch. The fragments of one cracked ball will quickly ruin the other balls and the races.
Early quality-control methods were poor at detecting flaws in roundness, and by 1913 a new but still crude technique had emerged: the balls were rolled down a trough, where they then fell onto an anvil and bounced toward a bin. Lopsided balls, it was reasoned, would miss the target. A later variation was to roll the balls down a sheet of glass.
As late as the 1920s the only way to measure a ball’s hardness and freedom from internal flaws was to crush it under a hydraulic press and watch the pressure gauge. Testers would lift a few balls out of each batch for smashing. Within ten years, though, machines were detecting some surface flaws electrically.
Because of their more complex shapes, races were even harder to scrutinize. In the early 1920s, inspectors measured races on micrometerlike gauges, determining both size and proper centering. They rubbed the inner grooves with silver-alloyed steel, hunting for soft spots, and dropped the races edgewise onto anvils to test for height of bounce and a proper bell-like tone. This was too labor-intensive a method to last long —one major manufacturer employed half its force in manual inspection—and the results were erratic.
In the past fifty years quality assurance has changed markedly. Measuring equipment today includes ultrasonic and magnetic testing for inner flaws, laser reflection to measure surface roughness, chemical etching to find grinding burns, and magnetic-particle detection of surface cracks. And there has been a change in the whole philosophy of quality control. The old method was to employ ranks of workers to inspect parts for flaws. This worked, but it didn’t get at the sources of those flaws, and the cost of discarding several percent of a production run could eat up all the profits. The current alternative is Statistical Quality Control. In a bearing factory it works this way: first, a record is made of the quality of work emerging from each machine, by time and batch. When enough information is available, charts are then drawn to show the average accuracy of each machine. As production continues, workers watch for times when quality significantly exceeds or falls short of the main trend. The goal is to get the machines “in control,” so that they produce consistent results—whether or not those are satisfactory— and then to improve their performance. Aberrations provide the information needed for improvement. The lure of SQC is that major savings can usually result with very little capital outlay.
When quality problems do appear, detective work the same day can often trace them to definite causes: perhaps the cooling oil is the wrong temperature, or minute floor vibrations have thrown off a machine’s gauging instruments. Such careful attention is the only way to produce bearing components meeting the requirements for accuracy. The most demanding tolerances are reserved for bearings that go into inertial guidance systems for airliners and missiles, and according to Jack Riley, a consultant to the bearing industry, it is now even possible to make bearings so precise that current instruments to measure performance cannot do them justice. Also, returns diminish quickly at accuracies tighter than one-millionth of an inch—the deformation caused by gravity itself enters the equation.
Statistical Quality Control emerged as a theory in the 1920s and saw some application in wartime industries but declined again with the postwar American business boom. “In the fifties and sixties, anybody here could make anything and make money,” says Riley. “They thought it was management skill, but it wasn’t. It was lack of competition.” As Americans drifted away from SQC, Japanese manufacturers increased its use. Ultimately, competition reawakened interest in SQC. American auto manufacturers “are jamming it down the throats of their suppliers,” Riley says.
Among the most intensive quality checks in the ball-bearing industry are those for jet-engine main-shaft bearings. These ball bearings, typically nearly a foot in outside diameter and costing as much as six thousand dollars each, have to tolerate six-hundred-degree temperatures and terrific combinations of thrust, centrifugal loads, and gyroscopic effects. As each new generation of jet engines demands higher rotation speeds and bigger main shafts, and as centrifugal forces rise with these, the ability of the bearing steel to hold together is sorely tried. If pushed too far, these bearings tend to crack catastrophically, shutting down the engine in seconds.
The steel in most main-shaft ball bearings is a blend called M-50, which is rich in chromium and molybdenum. It has been melted and cooled under a vacuum to remove impurities, and it is tested at the steel factory and again on receipt by bearing makers before forging begins. “For some jet-engine ball bearings, there may be fifty inspection operations at the tail end,” says Steve Martin, a senior vice-president at the Fafnir Bearing Division of Textron, Inc. Record-keeping is remarkably complete —an accident investigator can take a bearing serial number and trace each manufacturing step to a particular operator, day, and machine years before.
The consequences of aircraft bearing failure can be so serious that the manufacture and inspection of jet-engine bearings is comparatively labor-intensive and slow to change. Ordinary commoditytype bearings (for computers, appliances, vehicles, and the like) are now made in computerized factories that use robots and fully automatic cutting and assembling machines.
Most of the smaller commodity bearings consumed in this country now come from overseas, and price competition is fierce. “Take a bearing that sold for two dollars in 1982. Now there’s pressure to sell it for eighty or ninety cents,” says John Mussman, a vice-president of New Hampshire Ball Bearings, a maker of small bearings. Several American manufacturers have left the market.
Besides automation, another way to cut labor expense is to consolidate production runs, so that a given factory produces a minimum of different models and a maximum of each model. This reduces setup time, when highly trained workers must shut down the machines to readjust them for different models. Makers of commodity bearings are also seeking alternatives to the forging of race blanks, hoping to cut down on wasted steel. In the manufacture of some bearings, 50 percent of the raw stock leaves the factory as shavings.
One high-tech ball-forging experiment didn’t work out: in June 1973, Skylab astronauts tried cooling blobs of molten metal in free fall to see if perfect spheres would result. The balls were misshapen, the product of uneven cooling. Apparently nothing so far can beat the forge, which has been mashing out balls for over a hundred years now.
In fact, the ball-bearing industry is still using many principles developed fifty to a hundred years ago—considerably improved, of course, but still recognizable. Yet, over this period, ball bearings have become vastly advanced. Some can spin more than six thousand times a second; others can turn for years with absolutely no maintenance, sealed from weather and grit. If, as the poet William Blake suggested, one might “see a world in a grain of sand … and find eternity in an hour,” then to hold a ball bearing is to see the closest our machines have come to perfection.