Every rolling-element bearing catalog quotes a basic dynamic load rating “C” and a basic rating life “L10.” Most engineers learn the L10 formula in school, use it once, and never revisit it. That works until you specify a bearing for a gearbox running 18 hours a day in a sealed enclosure at 80°C and find out two years later that “10% failure at one million revolutions” was not the answer to the question you should have asked.
This article walks through the real difference between L10 and the adjusted rating life Lna (also written Lnm), why ISO 281 added the adjustment factor aISO, and how to use it without pretending you have data you do not have. If you want to skip the math, the Bearing Life Calculator implements both forms.
The Basic L10 Formula
For ball bearings:
L10 = (C / P)3 (millions of revolutions)
For roller bearings the exponent is 10/3 instead of 3. The variables are simple:
- C — basic dynamic load rating from the catalog (kN)
- P — equivalent dynamic bearing load (kN), computed from radial and axial components per the bearing-specific formulas
- L10 — life that 90% of an identical bearing population will reach or exceed
That last bullet is the part everyone glosses over. L10 is not “average life.” It is the 10th percentile of a Weibull distribution. The median life (L50) is roughly five times L10. If you size for L10 = 30,000 hours, you are saying you accept a 10% probability of fatigue failure within that period.
Why L10 Was Not Enough
The original L10 formula was calibrated against bearing test data from the 1940s. Steel quality has improved by an order of magnitude since then. Lubrication science is unrecognizable. Surface finish, manufacturing tolerances, and cleanliness of operating environments all affect fatigue life independently of load.
By the 1980s, field data was diverging from L10 predictions in both directions. A clean, well-lubricated bearing in a precision spindle would run ten times its L10. A bearing operating with contaminated grease and inadequate film thickness might fail at one-tenth of L10. The formula was not wrong — it was incomplete.
ISO 281:2007 codified the response: the modified rating life Lnm (also called Lna in older texts).
The Adjusted Life Formula
Lnm = a1 · aISO · L10
Two new factors. Here is what they mean:
a1 — Life Adjustment for Reliability
If you want a survival probability other than 90%, scale L10 by a1:
| Reliability | a1 |
|---|---|
| 90% (L10) | 1.00 |
| 95% (L5) | 0.62 |
| 96% (L4) | 0.53 |
| 97% (L3) | 0.44 |
| 98% (L2) | 0.33 |
| 99% (L1) | 0.21 |
If your application is a critical aircraft accessory and the mission profile requires 99% reliability, your usable life is 21% of L10. That is not a tweak. That is the whole reason aerospace bearings are specified with such margin.
aISO — Life Modification Factor
This one is the big lever. aISO is a function of three things:
- eC — the contamination factor (0.0 to 1.0). Sealed bearing with clean grease: 0.8 to 1.0. Open bearing in a dirty environment: 0.05 or worse.
- Cu — fatigue load limit of the bearing (catalog value).
- κ — viscosity ratio = ν / ν1, the ratio of actual operating viscosity to the rated viscosity. Need κ ≥ 1 for full EHL film; below 0.1 you are in boundary lubrication and aISO collapses toward zero.
The product eC · (Cu / P) feeds into a chart family in ISO 281 (one chart per viscosity ratio κ). At high κ and low contamination, aISO can reach 50. Below the threshold where the chart asymptotes to zero, fatigue life is dominated by lubricant breakdown and the bearing will not reach L10 at all.
A Worked Example
Consider a 6308 deep groove ball bearing in a vertical-shaft pump:
- C = 41 kN, Cu = 1.34 kN (catalog)
- Radial load = 5 kN, axial = 0.8 kN ⇒ P = 5.4 kN (computed per bearing geometry)
- Speed = 1750 rpm
- Oil viscosity at operating temp = 32 cSt; rated ν1 for size and speed = 16 cSt ⇒ κ = 2.0
- Sealed unit, filtered oil: eC = 0.7
- Required reliability = 95% ⇒ a1 = 0.62
Basic L10:
L10 = (41 / 5.4)3 = 437 million revolutions = ~4,160 hours at 1750 rpm
For the aISO term: eC · (Cu/P) = 0.7 · (1.34/5.4) = 0.174, at κ = 2.0 the ISO 281 chart gives aISO ≈ 2.4.
Lnm = 0.62 · 2.4 · 4160 = 6,190 hours
The adjusted life is 50% longer than basic L10 — because the clean, well-lubricated environment more than compensates for the tighter 95% reliability requirement. Without the ISO 281 method you would not have seen that.
When Each Method Applies
- Use basic L10 for quick catalog selection, preliminary sizing, and when lubrication and contamination conditions are unknown.
- Use Lnm when you have real lubricant viscosity data at operating temperature, know your filtration class, and need to size for a specific reliability target.
- Always cross-check against bearing speed limits using a DN-value calculator — fatigue life is meaningless if you exceed the cage speed limit.
- For mounted bearings, verify the fit with the Bearing Fit Calculator so internal clearance is not eaten by interference.
Common Pitfalls
- Plugging in nameplate horsepower as P. Equivalent dynamic load is a vector calculation, not just radial force.
- Forgetting that grease viscosity matters too — for greased bearings, use the base oil viscosity of the grease, not the apparent grease viscosity.
- Assuming a1 = 1 for everything. If the failure mode is critical, 90% reliability is not enough.
- Quoting Lnm without stating what reliability, contamination, and κ assumptions were used. The number is meaningless without those.
Bearing selection is one of the most decision-rich parts of mechanical design. Get the load path right, the lubrication right, and the cleanliness right, and bearings rarely fail. Get any of those wrong, and no amount of catalog margin will save you.
For a guided walkthrough including a1, aISO, and the full set of bearing-specific equivalent load formulas, use our Bearing Life Calculator. The Pro tier includes a contamination-factor wizard and a complete aISO chart implementation that matches ISO 281:2007 within 2%.