What are rolling bearings, and what is new about them?

What a rolling bearing is

A bearing is the part that lets a shaft or wheel turn while holding it in place. A rolling bearing is one common type. It places small balls or rollers between two rings, so the moving surfaces roll over each other instead of sliding.

The basic idea is older than most people expect. A 1981 NASA technical memorandum by D. Dowson and B. J. Hamrock, "History of ball bearings", says in its abstract that by 1800 the essential features of ball and rolling-element bearings had emerged, and that it remained for precision manufacture and mass production to confirm their value (NASA). We have read the abstract only.

How they work

In a plain bearing, a moving part slides against a fixed surface. The Wikipedia article on ball bearings explains that because the balls roll, they have a much lower coefficient of friction than two flat surfaces sliding against each other (Ball bearing). The same article notes a trade-off. Because the contact area between balls and races is small, ball bearings tend to carry less load for their size than other rolling-element types, although they can tolerate some misalignment.

Tapered roller bearings use cone-shaped rollers between cone-shaped raceways, and can take thrust (axial) loads as well as radial loads. Wikipedia describes the geometry: the raceway cones and the roller axes, if extended, meet at a common point, so the rollers roll along their whole length without differential scrubbing (Tapered roller bearing). The National Inventors Hall of Fame says Henry Timken's aim was a bearing that could cope with heavy loads from the side, as when vehicles turn (Hall of Fame).

What the primary records show

US patent 606,635, dated 28 June 1898, was granted to Henry Timken and Reginald Heinzelman of St Louis, Missouri, for a "Roller-Bearing for Vehicles". It describes conical rollers running between a conical sleeve on the axle and a conical box fixed to the wheel hub. Ribs on the sleeve fit grooves in the rollers, and the patent says these help stop the rollers coming out endwise or twisting, and help resist side thrust on the axle (Google Patents).

An earlier design came from John L. Scott of North Webster, Indiana. His US patent 552,008, applied for on 23 March 1895 and dated 24 December 1895, covers a bearing for wagon and buggy wheels with two separate series of rollers on cylindrical seats of a tapering axle skein, plus a series of balls (Google Patents).

A patent shows what was described and claimed. It does not show how well or how long a bearing worked in service, and none of the sources we used gives measured life, noise or friction figures for these designs.

What is reported in histories

The points below come from a company history, encyclopaedia entries and the National Inventors Hall of Fame. We have not checked them against original records.

• Philip Vaughan, a British inventor, received a patent for a ball bearing in 1794, which Schaeffler calls the first modern ball bearing (Schaeffler; see also Ball bearing). Schaeffler says its hand-crafted cast-iron balls had to be de-flashed by hand and ground, and that the bearing tended to rumble rather than run smoothly.
• Schaeffler credits Friedrich Fischer of Schweinfurt with a ball grinding mill in 1883 that produced hardened cast-iron balls of the same size on an industrial scale. It says Fischer reached accuracies of down to 20 micrometres and received Patent No. 55783 for the mill in 1890. This is the company's own account of its history.
• Wikipedia says Jules Suriray, a Parisian bicycle mechanic, was awarded French patent No. 86,680 on 2 August 1869 for ball bearings in bicycles, and that bearings were fitted to the bicycle that won the Paris to Rouen road race in November 1869 (Jules Suriray). The same page notes, citing an 1891 history of cycling, that ball bearings had already been patented in France in 1857 and in the United States in 1861. We have not seen the patents themselves.
• The Hall of Fame says Timken began experiments in 1895, received two roller bearing patents in 1898, and set up the Timken Roller Bearing Axle Company in 1899 (Hall of Fame).

What is new

Condition monitoring (peer-reviewed review). A 2024 review in the journal Machines by V. Kannan, T. Zhang and H. Li surveys research on monitoring the health of rolling bearings (Kannan and others, 2024). Its abstract says that bearing damage contributes significantly to rotating machinery failures, that vibration monitoring is still the most common approach, and that other sensor types are being studied alongside accelerometers. It says systems that combine several sensors with information fusion show promise for accuracy and reliability, and that machine learning and deep learning have improved accuracy and robustness in various studies. A review summarises other studies; it is not a single test, and the abstract does not give a figure for how much earlier faults are caught in industrial use.

Sensor bearings and ball precision (company claims). Schaeffler says it now makes sensor bearings that measure speed, temperature or load, so that impending defects can be spotted before a machine breaks down and maintenance intervals can be reduced. It also says the balls it makes today reach accuracies in the single-digit micrometre range, and that rolling elements are now made of steel or ceramics (Schaeffler). These are the company's own statements, not independent test results.

What is speculative

These are possibilities, not findings from the sources above:

• That sensor-equipped bearings become standard across most industrial machinery rather than a specialist option.
• That new bearing materials or coatings will allow much higher speeds and temperatures in future electric vehicle drive units.

Neither source we used gives adoption rates, cost figures or field data that would support these forecasts.

Limits and open problems

Lubrication matters. Wikipedia notes that the journal bearings used on wheel axles before rolling bearings relied on a lubricant film, and failed from frictional heat when lubrication was inadequate (Tapered roller bearing). The sources we used give no life figures for rolling bearings under load.

Precision was a problem from the start. Early balls were hand-finished, and better grinding, by Schaeffler's account, was what made smooth running possible (Schaeffler).

Cost is not covered. None of the sources gives production cost or cost-benefit figures, either for ordinary bearings or for sensor bearings.

What to watch next

Useful evidence would include independent field trials of sensor bearings and condition monitoring in real plants, published life tests on ceramic and coated bearings, and cost data. Until such sources are available, claims about how widely and how well these newer approaches work should be read as reported rather than demonstrated.