Inside the bearing: The difference between cages and spacers

Do you know why we use cages in some bearings to separate rollers and balls, in others spacers – and what the residual gap has to do with it?

In general, cages are more expensive than spacers. So if a bearing does not strictly require a cage, we use spacers (exceptions and a variant without any separators are explained at the end).

Cages keep balls or rollers in a precisely defined position relative to each other within the raceway. Spacers maintain the minimum distance to the next rolling element.

The basic function of both is to prevent rollers or balls from rubbing against each other during bearing rotation. Such friction leads to heat generation, material wear, and debris. This debris accumulates in the raceway, impairing bearing performance, smooth running, and even reducing the bearing’s service life.

^{But don’t the balls and rollers also rub against the cage or spacer?}

Yes, which is why these parts are made from low-friction materials. Ours are typically made of plastic, plastic‑coated steel, or – in the most expensive version – bronze. The friction behavior of these materials is far superior to that of steel, which the rolling elements are made of.

^{Ok, so both fulfill basically the same function. Why not always use spacers?}

Among other things, this depends on whether the “residual gap” matters.

Rolling elements and spacers are filled into bearings consisting of only two rings through a filling hole in one of the rings.

These two‑ring bearings are usually ball bearings. The filling process roughly works like this: One ring is fixed, the other is slowly rotated by a mechanical drive. The assembler adds the balls and spacers one by one.

Small gaps inevitably form between the elements. Manual filling through the filling hole does not allow for a 100% gap‑free fill, partly because the balls already move during assembly. In the end, these small gaps accumulate into a residual gap, which can amount to several centimeters in large bearings.

The balls and spacers can therefore shift several centimeters within the raceway even during operation. For the applications we manufacture these bearings for, that doesn’t matter. There are defined tolerances for this gap. For other bearings, however, it does matter.

By the way: In bearings with three or more rings, one ring is left out before filling, allowing the entire raceway to be filled at once and more densely, so the residual gap is much smaller.

^{Which applications must not have any residual gaps?}

Bearings with cages have no residual gap, which is essential for continuous rotation at relatively high speeds. For example, in main bearings of wind turbines. These rotate over 360 degrees, relatively fast, and continuously over 20 years or more. The fixed separation of the rolling elements ensured by a cage provides smooth running and ensures the load distribution in the bearing remains uniform, as the load‑carrying elements always maintain the same spacing. Thus, the same forces are always required to set the bearing in motion.

If we used spacers here, the residual gap would come into play: There would be regions in the bearing where the load is not evenly distributed. At the location of the gap, one ball would have to carry more load because its neighbor – who would normally share the load – is further away.

This effect becomes more critical in vertically installed bearings (with horizontal rotational axis), as is the case in wind turbine main bearings. Balls separated by spacers would always settle at the bottom of the bearing due to gravity. The residual gap would therefore always be in the same spot and thus constitute a permanent weak point. With cages, this does not happen because the rolling elements always stay where they are supposed to be.

^{And when does the residual gap not matter?}

Spacers with a residual gap are especially used when the bearing does not rotate beyond 360 degrees, does not rotate continuously, and is installed horizontally in the machine. For example, in a construction crane.

Here, the gap is irrelevant: because the balls do not slide downward, it is not always in the same location and not always the same size. Its effect is therefore random.

Additionally, the crane rotates slowly and not continuously, performing only swing cycles of less than 360 degrees. It moves back and forth. The influence of friction and slight unevenness in the distribution of the rolling elements is minimal, and perfectly smooth running is not critical.

^{Are there bearings where the rolling elements are not separated at all?}

Yes. This is called full complement. It is suitable when the bearing does not rotate often and usually only by a few degrees. This results in little friction even though the rolling elements contact each other, and there is no risk of the balls jamming. A typical application is truck fifth‑wheel couplings, the connection between the trailer and the tractor unit. Trucks usually drive straight, and when they turn, it is most of the time only by a small angle.

This is the most cost‑effective option, and full complement offers another advantage: the distance between rolling elements is minimal, meaning more balls or rollers can be inserted, enabling greater load capacity.

^{Are there cases where spacers are more expensive than cages?}

Yes. There are special cases such as high‑temperature applications. For example, in a ladle turret in a steel mill. It does not rotate continuously, only back and forth, and the bearing is mounted horizontally. Typically, a case for spacers. But not here.

Plastic is not suitable beyond a certain temperature, so steel must be used. However, steel spacers are more expensive than a steel cage, partly because elastic elements made of special spring steel would be required. Therefore, these bearings often simply use a steel cage.