Over the years, engineers have estimated the life of ball and roller bearings with the aid of industry standards and guidelines. These include the ANSI/AFBMA standards for predictin g the life of ball bearings (1950) and roller bearings (1953), which were updated in 1978 and 1990, plus the ASME rolling-element bearing life factors (1971). These standards were published by the American National Standards Institute (ANSI) and the Anti-Friction Bearing Manufacturers Association (AFBMA), which is now the American Bearing Manufacturers Association (ABMA). The American Society of Mechanical Engineers (ASME) was the first to publish life factors to compliment these standards.

Recently, the Society of Tribologist and Lubrication Engineers (STLE), with the aid of rolling-element bearing experts, took these standards and guidelines one step further by developing new bearing life factors that further reflect improvements in steel processing, manufacturing, design analysis, and lubrication over the last 50 years. These improvements have gradually increased bearing life to a level that is much higher than that predicted by both the original industry standards of the 1950s and the 1971 ASME life factors. These new bearing life factors, along with the manufacturing advances, are described in the book STLE Life Factors for Rolling Bearings, 1992.

Better steel

Research in steel metallurgy and processing has significantly improved bearing life over that obtained in the 1940s and reflected in standards of the 1950s. As summarized in Figure 1, here are the major manufacturing advances that contributed to improved life.

Heat treatment. Developments in bearing steel manufacturing began in the early 1940s with heat-treating equipment that incorporated improved temperature controls. Plus, the use of neutral atmospheres during heat treatment eliminated, for practical purposes, surface decarburization.

Bearing material research at the NASA Lewis Research Center beginning in the late 1950s culminated in the discovery of the differential hardness principle, which shows that differences between rolling element and race hardness significantly affect bearing fatigue life. For AISI 52100 steel, as an example, optimum life is achieved when the ball or roller hardness is 1 to 2 points Rockwell C higher than the races.

Melting. Major advances in melting practice evolved from 1952 to the early 1970s. Vacuum degassing and vacuum melting processes started in the late 1950s. One such process, vacuum-arc remelting (VAR), releases entrapped gases and alters the type of inclusions and trace elements in the steel so they are less likely to form failure initiation sites. At the same time, Pratt & Whitney Aircraft Div. of United Technologies started using AISI M-50 steel, which has higher operating temperature capabilities for aircraft engine bearings.

In the 1960s, argon atmosphere protection of the molten steel during teeming (pouring) substantially improved micro and macroscopic homogeneity and cleanliness, thereby reducing fatigue failures.

Testing. With the advent of vacuum processes, non-destructive evaluation (NDE) testing methods (eddy current and ultrasonic) were applied to billets, bars, and tubing to ensure the cleanliness of bearing steel.

Finishing. Before the 1950s, workers hand-polished as-ground bearing races to improve finish and appearance. But overly aggressive hand polishing created a thin layer of plastically displaced or smeared material that was softer and more prone to fatigue failure. In 1964, manufacturers replaced this manual process with mechanized honing, which finishes parts more uniformly.

Metalworking. As the need for bearing steel increased, manufacturers installed electric arc furnaces that produced larger billets. These billets had to be mechanically worked to reduce them to the size of smaller cross-section tubing or cylindrical forgings. This working process refines the grain and grain carbide and reduces the size of inclusions and segregates in the material.

A NASA Lewis Research Center study in 1958 led to the introduction of forged races with controlled fiber orientation in 1963, extending the life of angular-contact ball bearings.

Continue on page 2

Better lubrication

Advances in lubrication technology also contributed to improved bearing life. In the early 1960s, researchers found that bearing fatigue life and wear depend on the elastohydrodynamic (EHD) lubricant film thickness between bearing components, and on the surface finishes of these components. Hence, lubricant selection, improved bearing surface finishes, and control of operating conditions (temperature, speed, and load) that affect EHD film thickness enabled rollingelement bearings to operate at higher temperatures and for longer times.

Putting it all together

By the late 1960s, NASA Lewis combined manufacturing and lubrication advances to boost rolling element bearing life by approximately 13 times the amount predicted by the 1950 standards. These advances included improved surface finishes (obtained by honing), lubricants that provided larger EHD film thickness, controlled fiber and hardness, VAR AISI M-50 steel, and improved nondestructive (NDE) testing of steel billets.

In 1973, NASA’s research culminated in the first use of vacuum-inductionmelted, vacuum-arc-remelted (VIM-VAR) AISI M-50 angular-contact ball bearings with a lubrication system incorporating improved 3-mm oil filtration. The VIMVAR process uses better process controls to improve part consistency, and the improved filtration reduces damage to bearing surfaces from wear debris and contaminants. These ball bearings demonstrated lives of over 100 times the amount predicted by the 1950 standard at speeds to 3 million DN (D = bore diameter, mm, and N = rotational speed, rpm).

Lastly, in 1983, the General Electric Co. developed an improved AISI M-50 steel called M-50NiL. This material can be case hardened, and it exhibits more than twice the life of through-hardened VIM-VAR AISI M-50 steel. Developers attribute the life improvement to compressive residual stresses (not present in conventional AISI M-50 steel), and a finer carbide structure.

Updated life factors

Figure 1 also shows how bearing life has steadily increased over the last 50 years and how that increase was generally reflected in the ANSI/AFBMA standards and the life factors published by the ASME and STLE.

The original ANSI/AFBMA standards (1950s) contained bearing geometry and material coefficients, but no life adjustment factors, which were introduced later. Therefore, for purposes of comparison, assume a hypothetical life factor of 1. In 1971, ASME introduced design and environmental life adjustment factors of approximately 15 for both ball and roller bearings, reflecting the state-of-the art improvements since 1940.

In 1978, the AFBMA introduced the concept of separate life factors for reliability, manufacturing, and operation. The 1990 version of the ANSI/AFBMA standards incorporated life factors of 2.2 and 1.0 for ball and roller bearings respectively. These life factors are more conservative than the earlier ASME factors, but are intended to be used in combination with additional life adjustment factors obtained from the bearing manufacturers.

The 1992 STLE life adjustment factors, which are based on the manufacturing and lubrication improvements described earlier, reflect a life increase of over 60 times, compared to the 1950 standards. Though life predictions based on the STLE factors are much more accurate for today’s bearings, they are still conservative — representing only about one-third of the potential life attainable with current technology. Why? The bearing life curve in Figure 1 is based on full-scale bearing tests, whereas the STLE life factors are mostly based on bench rig rolling-element component tests, which nearly always produce conservative life estimates.

Using the original ANSI/AFBMA standards without life factors gives very conservative life estimates for angular-contact ball bearings, Figure 2. With no life factors, the life can be underpredicted by a factor of 100 or more. By contrast, the new STLE factors give much more accurate, though still conservative, results. This method generally underpredicts bearing life by a factor of 2 or 3 when used with the standards.

Continue on page 3

General calculation method

The ANSI/AFBMA and ISO (International Organization for Standardization) standards combine three life factors in calculating the life of a bearing:

Lna = a1a2a3L10 (1)

where:
Lna = Adjusted life, hr
a1 = Adjustment factor for reliability
a2 = Adjustment factor for materials and processing
a3 = Adjustment factor for operating conditions
L10 = Basic rating life, hr (obtained from manufacturer’s catalog or ANSI/AFBMA standards)

The STLE calculation method uses the same basic formula, but incorporates a more detailed set of life factors for manufacturing and operating variables. Space doesn’t permit a full explanation of the calculation procedure, which is described in the STLE book. But the following example illustrates the basic steps:

An angular-contact ball bearing, size 1905, is to operate with a load of 158 lb at 3,600 rpm and 210 F. Inner races, outer races, and balls are made of AISI 52100 steel, hardened to 63 Rockwell C. Using the standard, the L10 life is 5,253 hr. The bearing is lubricated by an SAE 20 mineral oil in a closed lubrication system with a 3-mm filter. Determine the adjusted bearing life, Lna.

• First list the bearing design parameters and operating conditions, which includes bearing type and size, component dimensions, materials, melting process, hardness, load, operating speed, temperature, L10 life, and type of lubricant.

• Then determine the three life factors a1, a2, and a3 for these parameters. The first factor, a1, is easily determined because it depends only upon the probability of survival that you select. For example, if you want to determine life at a 90% probability of survival, the ANSI/AFBMA life factor a1 is 1.

Determining a2 and a3 is more involved. For a2, you need to obtain individual life factors for manufacturing (materials and processing) variables from various tables and equations in the STLE book, then combine them into one value of a2. Life factor a3, which applies to the operating conditions, is determined in a similar manner. Some of these factors may also be found in bearing manufacturer’s catalogs or incorporated in the lives of bearings listed in these catalogs.

Table 1 lists all of the manufacturing and operating variables that may apply to a given problem. It also gives specific a2 and a3 values for the angular-contact ball bearing in our example.

• Finally, enter the three life factors, along with the L10 value into equation (1) to obtain the predicted bearing life:

Lna = 1 × 4.08 × 0.55 × 5,253
= 11,788 hr

By using this approach, you can determine the probable life of a bearing at the required reliability level with a more reasonable certainty than with previous methods.

For information on the book STLE Life Factors for Rolling Bearings, contact STLE at 840 Busse Highway, Park Ridge, Ill., 60068-2376.

Erwin V. Zaretsky is chief engineer for structures, NASA Lewis Research Center, Cleveland.

Related Articles

High-performance bearings boost professional cycling speeds
Bearing breakdown