A ball bearing works by rolling rather than sliding
What rolling contact changes, and its limits.
A ball bearing works by rolling rather than sliding because a rolling contact point is, in the ideal case, momentarily stationary relative to the surface it touches. That avoids almost all the surface shearing that makes sliding friction so large, in exchange for the much smaller resistance of the balls and races flexing slightly as they repeatedly meet and separate.
A ball and a block on the same ramp
A solid ball released at the top of a gentle ramp rolls down readily, needing only a small nudge to keep going, since almost nothing along its path drags across anything else. A block of the same weight on the same ramp needs a far firmer push to keep moving at the same pace, its whole underside dragging against the surface all the way down. Tilt the ramp steeper and the block eventually starts to slide on its own once gravity's pull along the slope overcomes the grip holding it. The ball has no such threshold and rolls away from a standstill on almost any slope, because rolling never has to wait for a grip to be overcome. A ball bearing inside a machine makes the same trade, swapping a dragging contact for a rolling one.
Where the remaining resistance comes from
Sliding friction comes from two surfaces dragging across each other, their microscopic high points continuously catching, deforming and releasing along the whole path. The point where a ball touches its race is momentarily at rest relative to it, the ball's rotation carrying a fresh point into contact at every instant instead of smearing one point along the surface. The resistance that remains comes mostly from the slight, repeated flexing of ball and race as each contact point is loaded and released.
A car tyre shows the same effect on a larger scale. The rubber flexes as it enters the contact patch and recovers as it leaves, losing a little energy as heat each time, which is why an underinflated tyre, flexing further on every turn, measurably raises fuel consumption even though the wheel is still rolling. A bicycle tyre pumped harder flexes less and rolls more easily, and a bearing designer exploits the same relationship by choosing harder, stiffer materials for balls and races wherever free running matters more than cushioning against shock.
The contact is also never quite a single point. It is a small patch, and different points across it would need to travel at slightly different speeds for pure rolling to hold everywhere at once, which a rigid ball cannot manage. A little genuine microscopic sliding therefore hides inside what is still called rolling contact. Designers of roller bearings address it by crowning the rollers, so that each bulges very slightly in the middle, concentrating contact where the speed mismatch is smallest and easing it at the edges where it would be worst.
What rolling costs, and when a plain bearing wins
Rolling contact is the right instinct wherever friction matters, and it has its own price. The balls and races still undergo repeated contact stress that eventually leads to the fatigue failure covered earlier in this era, and a rolling bearing demands tighter manufacturing tolerances, cleaner operating conditions and more careful alignment than a plain bearing needs. The lower friction is bought with precision, cleanliness and finite fatigue life.
In a dirty, high-shock or poorly aligned environment, a simple plain bearing can be the better choice despite its higher friction. It tolerates contamination, misalignment and sudden shock far more gracefully than precisely toleranced balls and races, and a rolling bearing degraded by dirt or misalignment can perform worse, and fail sooner, than the plain bearing it was meant to improve on. Rolling bearings are often supplied pre-packed with grease and sealed for this reason, since keeping dirt and moisture off those contact points matters as much to their life as anything about the balls and races themselves. Choosing between the two is a trade made against the real operating environment, and how that sensitivity to dirt plays out in a real bearing's failure is the subject of the next article in this set.