Home > EDITORIAL > Columnists > Motorhead Memo: Bearing Basics

Motorhead Memo: Bearing Basics

By Kip Woodring


Everywhere you look on a motorcycle there is a bearing. There are bearings in the engine, transmission, chassis, wheels, forks and more. As these bearings degrade, a motorcycle loses its “new” feeling. If they fail… you lose yours! This plethora of rotation-enabling is intended to locate and support stuff that spins; in other words, shafts! The choices a manufacturer makes in this regard obviously have a massive bearing (oops, just slipped out) on the overall design of your machine. Less obvious is the concomitant attention to lubrication. In theory, many bearing designs should be virtually frictionless. Modern micron-smooth, even ceramic-equipped, high-tech 21st century bearings might be close… but they still love lube. As long as it’s the right amount in the right places at the right time. This brings to mind both the major virtue and the Achilles heel of bearings.

Bearings are divided into two basic (but distinct) groups: plain bearings and rolling element bearings.

The first group is plain bearings. These bearings range from pivot points on control levers and simple bushings used on older machines to the complex hydrodynamic bearings used in most 4-stroke multi-cylinder engines. All plain bearings from high to low must support their shafts on some sort of “boundary” layer—mostly a hydraulic cushion of oil, but occasionally these days on PTFE/Teflon inserts, like those on the swingarm bearings on late model H-Ds. Swingarm bearings (to get this point out of the way) are not subject to great radial loads, do not spin, and, in fact, rarely rotate more than 25–30 degrees. Counterintuitively, the curvature in these big spheres is there primarily to handle any misalignments along the pivot shaft’s travels through swingarm and tranny cases. Otherwise a couple of bronze bushings would do just as well—and did for decades—when you come down to it.

At the other extreme, plain bearings that do spin at high speeds—like those on the cranks of sportbikes—need copious, reliable quantities of clean oil to prevent metal-to-metal scuffing and indeed to work at all! If they get it, the bearings work better and better, the faster and faster they spin. The Motor Company uses few of these in traditional V-Twins, but since the turn of this century, “all new” machines have plenty. Be apprised.

Our second group, rolling element bearings, use sophisticated inner and outer metal races with either balls or rollers in between. Most (but not all) rolling-element bearings also have “cages” of various types and materials to hold the rolling elements in position and reduce friction, and more. They are further defined using two major distinctions: ball and roller. Ball bearings have various sizes and numbers of balls between the inner and outer races, while roller bearings have cylindrical or conical shaped rollers to provide the rolling action between the inner and outer races. Roller bearings have a greater load capacity, and deflect less, but have greater startup friction than an equivalent-sized ball bearing. Both types suffer from a small but certain amount of scuffing friction—roller/ball against races/cages due to manufacturing inaccuracies—no matter how small, plus the inevitable shape-shifting from heat, and unintended side loadings. Compared to plain bearings, potential rolling speed is limited, but “iffy” lube is nowhere near as critical. Rolling element bearings require—in fact thrive on—just enough lubricant to keep their parts wet, whether by oil or grease.

Bearings are further defined by the loads they are designed to support, an important aspect to engineers too often overlooked by end users. In engineering speak, loads are defined as vectors. This is a fancy way of saying that loads have both a magnitude and direction, let’s call them axial and radial, or more likely some combination of the two. Radial loads push or pull in the same plane in which the bearing rotates and perpendicular to the supported shaft. Axial loads, often referred to as “thrust” (as are the bearings), radiate at angles to the rotation of the bearing and more parallel to the shaft, which can make proper selection and use a bit tricky. That’s the main reason why there are so many variations of rolling element bearings!

Aside from these basic ball or roller specs, there are either caged or full-complement versions of both types. A full-complement bearing is one in which the space between the inner and outer races is stuffed with as many rollers or balls as it can hold. The ultimate load carrier, but not good at high-speed spinning—no room! Since the load is only carried by a few elements at a time, it’s concentrated on the tiny points where those rolling elements contact the races. When that carrying capacity is exceeded, it causes permanent, damaging deformation of races and/or balls/rollers. Once this “brinelling” happens, the bearing starts to rumble and grumble, shudder and shake, causing the whole damn thing to disintegrate… usually at speed!

Actually, at speed is where the caged bearing comes into its own. This variety uses a thin, light, “floating” shell to surround the balls/rollers, keeping them separated and rolling independently, while minimizing potential scuffing as well as keeping centrifugal force from smashing the elements together. Caged bearings are used mostly in engines and transmissions, where moderate loads and high speeds are the norm.

That said, the most common bearing on a motorcycle is the deep-groove ball bearing. These bearings are used for wheel, crankshaft, and transmission applications galore, as well as places you wouldn’t normally think of, like a Sportster clutch throw-out. A breakthrough (and patented) design from about 1910, these bearings are capable of high speeds and will support substantial amounts of both radial and thrust/axial loads in both directions. (It’s important to note that all ball bearing have “channels” in the races for the balls to run in, but without those sophisticated deep grooves, they are unable able to deal with a variety of loads.)

As the name implies, bearings designed for loads primarily at angles to the shaft are so-called angular contact bearings. The inner and outer races are shaped in such a way that a line drawn from the contact points of the balls on the inner and outer race would be at the designed angle to best accommodate the direction the load comes from, typically with contact angles in standard increments of 15, 25, 30, and 40 degrees. The greater the angle, the more axial load and less radial load the bearing will support, but always more load than a deep-groove ball bearing. The bad news is that angular contact bearings can only resist axial loads in one direction, and must be used in pre-loaded opposing pairs.

Tapered roller bearings (the same idea but with rollers instead of balls) are common practice in Harley steering heads (and once upon a time in wheels and swingarms) offering much greater stiffness, load carrying ability, and convenience (but more friction) when compared with ball bearings. (In years past it was considered an upgrade to replace steering head ball bearings with tapered roller bearings in any motorcycle. Now, it seems sportbikes have reverted back to ball bearings, going for less friction, while the harsh world of dirt bikes and hogs calls for the more durable tapered roller bearing.)

“Needle” roller bearings are a special type of cylindrical roller bearings that have long slender rollers. These bearings can support large radial loads, but can’t support axial loads worth a damn! The big advantage of the needle roller bearing is the small outside diameter relative to the bearing inside diameter to support a given load. In other words, it’s ideal for use in applications where space is tight, such as suspension linkages, (some) swingarms, connecting rods and… (wait for it)… camshafts! Full complement needle bearings are known in H-D parlance as “Torrington” and caged needles bearings as “INA.” The debate rages over which of these two types is really most suitable for performance applications on Harleys. It’s a textbook example of the compromises in engineering—the reason so many bearing varieties exist—and the importance of choosing the right one for the task at hand!

Most bearings have a four-digit code that defines their intended use.

The first digit is “type”:

1 = Self-aligning ball bearing

2 = Barrel and Spherical roller bearing

3 = Tapered roller bearing

4 = Deep-groove double row ball bearing

5 = Axial (only) thrust ball bearing

6 = Deep-groove ball bearing

7 = Single-row angular contact bearing

8 = Axial cylindrical roller/needle bearing

R = Inch-sized (non-metric)

N = Cylindrical roller bearing

NN = Double-row cylindrical roller bearing

NA = Needle roller bearing

The second digit is the “grade” which determines both the (strange) relationship to bore size, outer diameter and race thickness—and the “duty” level of the bearing—in proper order:

9 = Very thin

0 = Extra light thrust

1 = Very light thrust

2 = Light thrust

3 = Medium thrust

4 = Heavy thrust

(Bearings are graded on two scales of increasing precision: ABEC and ISO class. Some people believe that high precision bearings produce less friction. Not! What high accuracy bearings offer is less radial and axial run out.)

The third and fourth digits indicate the bore size in millimeters, which is simply five times the third and fourth digits together.

(0 through 3, however, are different, like this:)

00 = 10mm

01 = 12mm

02 = 15mm

03 = 17mm

(The exception is when a “/” precedes the last two digits. For these bearings the last digits are the bore size. Example: “/22” has a 22mm bore.)

Letters (if any) that come after the numbers mean things like this:

Z = Single-shielded

ZZ = Double-shielded

RS = Single-rubber sealed

DRS = Double-rubber sealed

V = Single non-contact seal

VV = Double non-contact seal

DDU = Double-contact seals

NR = Snap ring and groove

M = Brass cage

A shield is a stamped metal part that is fixed to the outside diameter of the bearing and just clears the inner ring. The result is a very small clearance between the shield and the inner ring, thus greatly reducing the influx of foreign matter into the bearing. Because the shield does not actually contact the inner ring, there is no increase in friction. A bearing can have a shield on one or both sides, or neither.

Seals look very similar to the shield. Seals are constructed from synthetic rubber molded to a stamped metal part. Seals come in two basic flavors, non-contact and contact. Non-contact seals are just that. There is no contact between the seal and inner ring.

The advantage of a seal over a shield is that the flexibility of the rubber allows an arrangement called a “labyrinth seal” to be assembled. While not actually contacting anything rolling, this seal arrangement forces any material to travel through a maze before it can enter the bearing. This difficult path is made even more challenging by the barrier formed by grease blocking the way.

For ultimate protection of the bearing from dirt and moisture, a contact seal is used, much like the oil seals used on the output shaft of the transmission. This type of seal does have more friction and a lower maximum speed, but provides the most protection from the environment.

I’m not going to go through the drill of properly removing, installing and maintaining bearings, beyond these few generalities:

In a perfect world, no bearing should be installed by sheer brute force. To me that means (in descending order of preference) either heating the housing and freezing the bearing, if the bearing is to be inserted in a bore, then gently pushed into place. (Loctite actually makes an aerosol product that helps a lot.) If that won’t work, a proper press and installation tool should be employed and always/only used on the race that touches the bore. Going onto a shaft, the only race to touch is the inner. Either instance involves making damn sure the race is perpendicular, since nothing is worse than a crooked bearing.

Never hit a bearing with a hammer! If you simply must bang on the poor thing in an emergency, use a block of wood fer chrissakes!

So much for basics. Now, a few random observations and opinions.

Harley-Davidson wheel bearings (since the Timken tapered roller type were discontinued after four decades) are odd-sized. The factory switch to metric deep groove bearings, while still using inch-sized axles, means it is nearly impossible to find a suitable replacement anywhere but a dealer. This includes the bearings for 25mm axles, because the 21mm width is weird. To make matters worse, there’s many an online vendor who would have you type in the basic dimensions of the stock bearing, then offer up “alternatives” which they will be happy to sell you, quite often very cheaply. Beware! These alternatives are liable to be of a different type entirely and not suitable for wheels, let alone Harleys. Companies that specialize in motorcycle bearings (like All Balls) are the rare exception.

Another reason for caution involves load rating, a spec rarely checked let alone understood by the average rider. For instance, the old Timken wheel bearings had a load rating (based on that linear contact) of right at four tons! The newer deep-groove bearings are rated a lot less—sometimes half of that. While on the face of it, load ratings in multiples of anything you’d expect, even on a bagger burdened with payloads beyond the recommended GVWR, might seem more than adequate, it is this author’s opinion that this simple fact (coupled with poor installation practices) is the root cause of “issues” (perceived or real) with late-model bearing failures. Perhaps a price paid for stiffer chassis, increases in engine torque and, well, heavier motorcycles and lighter-duty bearings? Country of origin has little or nothing to do with it!

Leave a Reply

Your email address will not be published. Required fields are marked *