Tapered Roller Bearings

What is a Tapered Roller Bearing

A tapered roller bearing is a rolling element bearing whose rollers and both raceways are conical rather than cylindrical or spherical. The defining design rule is geometric: the projected lines of the inner raceway, the outer raceway, and the roller surfaces all converge at a common apex on the axis of the bearing. When this convergence condition is met, every point along a roller travels at a surface speed proportional to its radius, so the roller rolls cleanly across the raceway without the differential sliding (scrubbing) that would otherwise generate heat and wear. That single principle is the reason the bearing exists and the reason its rollers must be precisely ground cones rather than barrels or cylinders.

The conical contact also creates a line contact between roller and raceway, in contrast to the point contact of a ball bearing. Line contact spreads the load over a larger area, so for a given envelope a tapered roller bearing carries substantially more load than a ball bearing of the same bore. The tilt of that line contact means each roller reacts to load along a direction inclined to the bearing axis, which is why a tapered roller bearing naturally takes both a radial component and a one-directional axial (thrust) component at the same time. To take thrust in both directions, two bearings are mounted facing opposite ways, either back-to-back or face-to-face.

Physically the bearing has four parts. The cone is the inner ring; it carries the inner raceway and a guide rib (the large-end flange) against which the roller large ends slide. The rollers are the conical rolling elements. The cage retains and evenly spaces the rollers and keeps the cone assembly together as one non-separable unit. The cup is the outer ring, with the matching outer raceway machined inside it. Because the cup lifts off the cone, the two are handled and mounted separately, a property that distinguishes tapered roller bearings from most deep-groove ball bearings and which the rest of this guide returns to repeatedly.

The modern tapered roller bearing dates to 1898, when American carriage builder Henry Timken received a United States patent for applying tapered rollers to wagon wheels, replacing the plain journal bearings that depended on a fluid film of grease. The design quickly spread from carriages to early automobiles, then to railway axle-boxes, industrial gearing, and heavy machinery. More than a century later the same geometry dominates the same applications: the front and rear wheel hubs of cars and trucks, the differential pinion and ring-gear shafts, manual and automatic transmission shafts, agricultural and construction equipment, rolling mill rolls, and wind turbine main shafts and gearboxes.

Four engineering properties decide how well a tapered roller bearing fits its duty: the contact angle (how the load splits between radial and axial), the basic dynamic load rating C (which sets fatigue life), the precision and clearance class (which set running accuracy and noise), and the achievable mounted setting (preload or end play, which sets stiffness). These properties are not independent of cost, and a bearing that is cheaper to buy but runs hotter, drifts out of setting, or fails earlier usually costs more across a machine lifetime once downtime and replacement labour are counted.

Types and Row Configurations

Tapered roller bearings are classified first by the number of rows of rollers, and within each row count by the arrangement of the cups and cones. The row count determines the load capacity and the moment (overturning) stiffness, while the arrangement determines whether the bearing locates the shaft, floats, or is supplied as a pre-set assembly. Single-row bearings are by far the most common; multi-row designs appear where loads, moments, or compactness demand them. The table below summarizes the main configurations using the widely used Timken-style nomenclature, which the rest of the industry largely mirrors.

Single-row (TS) is the reference design. A single row takes radial load plus thrust in one direction only, so two single-row bearings are mounted opposed to handle reversing thrust and to be adjustable to a target setting. The two classic arrangements are back-to-back (designated DB, the wide effective spread that resists tilting moments well) and face-to-face (DF, easier to mount and tolerant of misalignment but with a narrower effective spread). The choice between DB and DF is one of the most consequential decisions in shaft design because it sets both the moment stiffness and how the bearings respond to shaft thermal growth.

Double-row designs (TDO and TDI) combine two rows into one component, eliminating the need to set two separate bearings in the field and providing thrust capacity in both directions out of a single part. The TDO (double cup) places the two contact-angle apexes outboard, giving a wider effective load spread and excellent resistance to overturning moments, which is why it is favoured for wheel ends and sheaves. The TDI (double cone) places the apexes inboard with a narrower spread; it is commonly used as the fixed (locating) bearing on rotating shafts and on rolling mill roll necks. Both are typically supplied as pre-set assemblies with a precision spacer that fixes the internal setting at the factory.

Four-row designs (TQO and TQI) pack four rows into a single cartridge that is delivered pre-set, ready to press into a housing. Their dominant home is the rolling mill, where work-roll and back-up-roll necks must carry enormous radial loads in a very compact axial space and where field setting of individual rows would be impractical. Beyond row count, tapered roller bearings are also offered in steep-angle thrust families and in special wide and high-capacity dimension series, all of which trade off radial versus axial capacity through the contact angle discussed in Chapter 4.

Materials, Cages, and Precision Grades

The fatigue life and the temperature limit of a tapered roller bearing are set largely by the steel of its rings and rollers, while running accuracy, noise, and speed are set by the precision class and the cage. Two steel routes dominate, and the choice between them is driven by whether the application is fatigue-limited or shock-and-fit-limited. The table below compares the two raceway steels and the three common cage materials.

Through-hardened AISI 52100 is the classic bearing steel, alloyed with roughly 1.0 percent carbon and 1.3 to 1.6 percent chromium and hardened throughout its section to about 58 to 64 HRC. It delivers the high, uniform hardness that rolling-contact fatigue resistance demands and is the default for fatigue-limited applications running below roughly 120 degrees Celsius. Above that, the steel must be given a special heat stabilization treatment (for example an S0 or S1 stabilization) to prevent dimensional change in service; without stabilization, continuous high temperature causes the rings to grow and the setting to drift.

Case-carburized AISI 8620 and similar low-carbon alloy steels are diffused with carbon at the surface and quenched to produce a hard wear-resistant case over a tough, ductile core. The combination tolerates shock loading and the heavy interference fits that tapered roller bearings often require, because the soft core resists the cracking that a fully through-hardened ring would suffer under a tight press fit. Carburized steel is therefore common in off-highway, rail, and heavy industrial bearings where impact and tight fits matter more than the last few percent of fatigue life.

Cage choice trades cost against speed and load. Pressed (stamped) steel cages are inexpensive and standard on most single-row catalog bearings. Machined brass cages tolerate higher speeds, higher loads, and elevated temperatures and are common on large and high-performance bearings. Polymer cages, typically glass-fibre-reinforced polyamide 6.6, are light, quiet, and low-friction, but their continuous temperature ceiling is lower (commonly around 120 degrees Celsius for PA66) and they can be attacked by certain aggressive lubricant additives, so the application temperature and lubricant must be checked.

Precision class is governed by ISO 492, which defines the tolerance classes from loosest to tightest as Normal (P0), Class 6 (P6), Class 5 (P5), Class 4 (P4), and Class 2 (P2). For tapered roller bearings, the North American ABMA inch system uses the parallel class designations Class 4 (the standard), Class 2, Class 3, Class 0, and Class 00 in order of increasing precision. Normal precision suits ordinary shafting; P5 and P4 suit machine tool spindles and high-speed gearing where runout and predictable preload are critical. Tighter classes cost more and are slower to deliver, so they should be specified only where the application truly needs them.

Dimensions, Designations, and Standards

Two dimension systems govern tapered roller bearings worldwide, and they are not interchangeable. The metric system follows ISO 355, which fixes the boundary dimensions (bore, outside diameter, width, and the cup and effective widths) and the series designations for metric tapered roller bearings. The inch (imperial) system follows the ANSI/ABMA inch standards that grew out of Timken practice in North America. The critical practical consequence is that a metric cup never mates with an inch cone, and a replacement must match the original system exactly.

In the traditional metric designation, the digits encode the dimension series and the bore. Take 30205 as a worked example: the first three digits 302 identify the dimension series, in this case a light single-row family with a standard (shallow) contact angle, and the last two digits 05 multiplied by 5 give the 25 mm bore. The principal metric dimension series, with their general character, are summarized below.

The contact angle is what splits the load between radial and axial. It is the angle between the line of contact force through the rollers and a plane perpendicular to the bearing axis. A larger angle directs more of the capacity into thrust. Standard-angle families such as 302xx and 303xx sit around 10 to 16 degrees and are radial-biased general-purpose bearings; the steep-angle 329xx family sits around 25 to 30 degrees and is selected when the load is dominated by axial thrust, such as on a hypoid pinion shaft. Choosing the angle band therefore precedes choosing the exact bore.

ISO 355 also defines a parallel alphanumeric scheme for newer metric bearings, of the form such as T2DB050. Read left to right, the T marks a tapered roller bearing, the next character is the angle series, the letter is the width series, the second letter is the diameter (outside) series, and the final digits give the bore in millimetres. A bearing whose cup width, cup small bore, and contact angle conform to ISO 355 may also be flagged with the suffix J. Catalog cross-reference tables map these ISO 355 codes back to the familiar 3xxxx numbers.

On the inch side, the cup and cone each carry an independent part number drawn from the ABMA inch series, and a complete bearing is quoted as a cone-plus-cup pair, for example cone 3782 with cup 3720. There is no embedded bore-times-five rule as in the metric system, so inch sizes must be looked up in dimension tables. Two further public standards govern the numbers a catalog prints: ISO 281 defines the basic dynamic load rating C and the L10 rating-life calculation, and ISO 76 defines the basic static load rating C0; the ANSI/ABMA 11 and 9 series give the parallel North American methods, now largely harmonized with the ISO formulas.

Key Specification Parameters

A tapered roller bearing datasheet lists boundary dimensions plus a block of performance numbers. Only a handful of those numbers actually drive selection. The eight that matter most are explained below: contact angle, basic dynamic load rating C, basic static load rating C0, the calculation factors e, X, and Y, limiting and reference speed, precision and clearance class, the setting (preload or end play), and the resulting L10 rating life.

Basic dynamic load rating C is the constant radial load under which a bearing achieves a basic rating life of one million revolutions, with C0 being the static equivalent that the bearing tolerates without harmful permanent deformation. C drives fatigue life through the ISO 281 expression L10 equals (C divided by P) to the power 10/3, where L10 is life in millions of revolutions at 90 percent reliability and the exponent 10/3 is the value standardized for roller bearings (it is 3 for ball bearings). Doubling the equivalent load P therefore cuts life by roughly a factor of ten, which is why correctly computing P matters more than chasing a marginally larger C.

The factors e, X, and Y turn the applied radial load Fr and axial load Fa into the single equivalent dynamic load P used in the life formula. Each tapered bearing has a limiting ratio e: when Fa divided by Fr is at or below e, the axial load barely changes the contact pattern and P is taken as Fr; when the ratio exceeds e, P is computed as X times Fr plus Y times Fa using the catalog X and Y for that bearing. Crucially, a single-row tapered bearing under radial load generates an induced axial force of about 0.5 times Fr divided by Y, which must be balanced by the opposing bearing in the pair; ignoring this induced thrust is a frequent cause of premature failure.

Setting is the axial play remaining in the mounted pair, and it is unique to adjustable bearings like these. A positive setting is end play (free axial movement); a negative setting is preload (axial interference with no free play). End play is common on wheel hubs and idlers, with automotive practice often targeting roughly 0.025 to 0.130 mm (0.001 to 0.005 inch). Preload is used where stiffness and running accuracy matter, such as pinion shafts and machine spindles, but it raises friction and operating temperature, so the optimum setting is usually a narrow band rather than simply tighter is better.

Speed ratings, precision, and clearance complete the picture. The catalog limiting speed and reference speed bound how fast the bearing may run under grease or oil lubrication before heat or cage stress become limiting. Precision class (ISO 492 Normal/P6/P5/P4) sets runout and the predictability of the mounted preload. Radial internal clearance is set indirectly through the mounted setting and the interference fits, which is why the as-installed clearance differs from the as-manufactured clearance. The remaining datasheet entries to verify are listed below.

• Contact angle (alpha): determines the radial-to-axial split; confirm it matches the load case before anything else.
• Calculation factor Y: needed both for equivalent load and to size the induced axial thrust in a mounted pair.
• Cup and cone widths (T, B, C): tapered bearings have several width dimensions, not one; confirm which the housing shoulder uses.
• Fit recommendations: shaft and housing tolerance bands (for example a rotating-inner-ring fit such as k5 or m6 on the shaft) that determine mounted clearance.
• Lubrication and seal type: open, sealed, or pre-greased; grease type and quantity govern the practical speed and maintenance interval.

Selection Decision Factors

To turn the preceding chapters into a specific part number, follow the decision sequence below. Most selection errors come not from a single wrong number but from deciding the wrong thing first, for example fixing a bore before confirming the dimension system or the contact angle. These eight steps double as an RFQ template.

1. Dimension system and envelope: First confirm metric (ISO 355) versus inch (ABMA), then fix the available bore, outside diameter, and width envelope. A metric cup will never mate an inch cone, so this gate prevents the most expensive mistake.
2. Load case and contact angle: Resolve the radial load Fr and axial load Fa, compute Fa divided by Fr, and pick the angle band: standard angle (302xx / 303xx) for radial-biased duty, steep angle (329xx) for thrust-dominant duty such as pinion shafts.
3. Row configuration and arrangement: Single row (TS) opposed pairs for adjustable general duty; back-to-back (DB) for high moment stiffness, face-to-face (DF) for misalignment tolerance; TDO/TDI/TQO where compact multi-row capacity or factory pre-set is needed.
4. Load rating and L10 life: Compute the equivalent load P from Fr, Fa, e, X, and Y, then size C so that L10 equals (C divided by P) to the 10/3 meets the target life. Add the induced axial thrust of the opposing row before sizing.
5. Precision and clearance class: Normal (P0) for ordinary shafting; P6, P5, or P4 for high-speed or high-accuracy spindles and gearing per ISO 492. Specify tighter classes only where runout and preload repeatability truly demand it.
6. Setting strategy: Decide preload (stiffness, accuracy, higher heat) versus end play (simplicity, thermal-growth tolerance) and state the target setting band and the adjustment method (shim, spacer, nut torque, or collapsible spacer).
7. Material, cage, and temperature: Through-hardened 52100 for fatigue-limited duty below about 120 degrees Celsius (stabilized steel above that); carburized 8620 for shock and tight fits; pressed steel, brass, or polymer cage per speed, load, and lubricant compatibility.
8. Total cost of ownership (TCO): Purchase price plus mounting and setting labour plus lubrication and re-setting intervals plus failure-downtime cost. A cheaper bearing that drifts out of setting or runs hot can erase its price advantage within one maintenance cycle.

One last commonly overlooked dimension is serviceability and sourcing: confirm local stock and lead time, verify cross-reference availability between brands and between the metric and inch systems, and require that the supplier publish traceable C, C0, e, X, and Y values rather than copied catalog numbers. Tapered roller bearings are made by Timken (the originator of the design and a benchmark for inch and rail products), SKF, Schaeffler (FAG and INA), NSK, NTN, JTEKT (Koyo), and NACHI for the global metric range, with C&U, ZWZ, LYC, and HRB serving high-volume metric demand at lower cost. Matching brand to criticality, and confirming field calibration and spare availability, determines repair response over a machine that may run for decades.