Ceramic bearings

Thursday, March 29, 2012

Tags :  bearings

bearingsIn recent times the push behind developments in bearings containing a ball or rolling element has been that of improved, 'cleaner' steels. Because of its high hardness and therefore resistance to wear, chromium steel has been difficult to beat, and reducing the level of impurity - the number, type and size of any rogue inclusions - to improve the fatigue resistance has made them more so. In many cases when bearing selection is driven, at least in part by cost, mass-produced steel components of this type are perhaps the only solution. However, if cost is factored out of the decision and performance offered an enhanced role, ceramics in one form or another may offer a better solution.

Ceramic bearings fall into two distinct types - full ceramic units when outer race, inner race and balls are all made from a ceramic, and 'hybrid' units when for reasons of cost, only the inner ball/rolling elements are ceramic. With the latter, although the inner and outer races will be superfinished in steel, the full assembly still retains many of the advantages of the full ceramic. Furthermore, bearing ceramic materials can be classified into three common materials used: silicon nitride (Si3N4) - which seems to be the most favoured currently -silicon carbide (SiC) and zirconium dioxide (ZrO2), sometimes referred to as zirconia. The first two are black in colour while the third is white, but despite these differences, all have significant advantages over steel.

It has always bemused me that a ball or roller bearing had any significant friction because, at a theoretical level at least, only point contact with no relative sliding motion exists between the inner race and the ball or roller, or the ball/roller and the outer race. Think of it as an inner gear of an epicyclic gearbox surrounded by star wheels and an outer annulus, and you will understand that in theory, since there is no sliding motion, no friction should be present.


However, as well as the inevitable small amount of slip, energy will be needed to rotate the roller around first its own axis and then around the axis of rotation of the whole assembly. Ball or roller inertia therefore plays an important part in the 'friction' within a ball or rolling element bearing. So when the density of the ball or roller is around 40% that of steel, this constitutes a major reduction in the energy required to move them. Also, the forces in the bearing and its 50% greater stiffness over steel produce less distortion and generate less heat than its steel equivalent.

This reduced amount of heat, and the lack of steel-on-steel contact generating microwelding between the components, also means that adhesive-type wear will be substantially reduced. Ceramic bearings therefore run cooler and require much less in the way of lubrication than traditional materials.

Lighter weight, reduced friction, less wear and lower operating temperatures all add up to a longer operating life (sometimes by as much as ten times) or greatly improved performance; a sort of win-win-win-win situation, which is difficult to ignore. In the words of a recent UK TV advertisement, "Now you don't often see that do you?"

Fig. 1 - Ceramic bearings

Written by John Coxon

Leave your comment


Dear John,
I beg to differ just a wee bit on the reasons for friction in a ball- or roller bearing, where the roller's mass plays a part, but perhaps not xactly the way you describe it.

Roller bearing friction comes in two components:

A) Load-dependent hysteresis and microslip in the roller, inner race and outer race, from distorsion of interacting surfaces.

B) Not load-dependent viscous losses in the lubrication, viscosity-dependent.

Ceramic rollers are stiffer and lighter (think centrifugal force) than their steel counterparts why distorsion will be less, thus heat from A) is less, meaning that lubricant viscosity can be lower, which reduces heat from B).

Read more about it here;

my mind went more towards Hertzian stresses and uniformity of surface structures...
I stand corrected.
I agree with xpensive about rolling element bearing mechanical losses.  If the bearing is operated at low load and high DN, viscous losses will usually predominate.

There are also a couple other things to consider with ceramic bearings.  With hybrid bearings (ceramic rolling elements and steel races), fatigue life is usually not quite as good as all steel bearings.  Due to the much greater stiffness of the ceramic material, more strain is produced at the steel race surface for a given load.  100% ceramic bearings also do not fail quite as "gracefully" as steel bearings.  Ceramic materials tend to fail abruptly.  It is easy to monitor steel bearings for failure using lube oil magnetic chip detection, and steel bearings will usually begin to generate detectable amounts of debris long before they fail catastrophically.  

As an interesting side note,  it has recently been demonstrated that applying very hard surface coatings (like DLC) to the contact surfaces of steel bearings significantly increases fatigue life.
At the relative sppeds where I typically work, 1500 Rpm and 200-400 mm bore, viscous losses is actually dominant, where most of the lubricant doesn't pass the interface between roller and raceway anyway, why using a VG as low as possible is paramount.

This is perhaps where we can see a gain with hybrids, but so far not much enough to warrant the cost.
With rolling element bearings, as noted, viscous losses will almost always predominate, except at very low speeds.  But careful design of the bearings and mountings can minimize mechanical losses.

All rolling element bearings suffer from some degree of mechanical loss due to sliding, skewing, rubbing or skidding.  There is sliding contact between rollers and shoulders.  There is sliding contact between elements and retainers.  There is roller skidding due to shaft/housing deflections.   Since the oil flow requirement for most bearings is a function of cooling, optimizing bearing geometry (contact angle, roller crown, groove osculation, retainer design, preload, etc.) to minimize heat generation will reduce the cooling oil flow needed.  And less oil flow means less churning, windage, and viscous losses.