Bearing life L10 explained: from load rating to operating hours

What L10 life actually says

L10 is the basic rating life of a rolling bearing. It is the life that 90 percent of a large group of identical bearings are expected to reach or exceed under ideal conditions. Ten percent are expected to fail earlier. It is therefore a statistical figure for a population, not a promise for one bearing, and it is not a guaranteed minimum.

L10 is stated in millions of revolutions. Dividing by speed converts it to operating hours, written L10h. The bearing life calculator returns both values plus the hours expressed as years of continuous operation. The basic calculation covers fatigue of the rolling surfaces under load only. Lubrication quality, contamination, misalignment, temperature and a reliability target above 90 percent are outside it.

The formulas and their inputs

The calculator uses these relationships:

• L₁₀ (million revolutions) = (C ÷ P)^p
• L_10h (hours) = L₁₀ × 1,000,000 ÷ (60 × RPM)

The exponent p is 3 for ball bearings and 10/3 for roller bearings. C is the basic dynamic load rating from the bearing data sheet. P is the equivalent dynamic bearing load, a single load figure that combines radial and axial forces using the factors the bearing supplier publishes. C and P must be in the same unit; the calculator labels both in kN.

Two consequences follow from the exponent. Life falls steeply as load rises, and life is inversely proportional to speed. Doubling P cuts a ball bearing life to one eighth (2³ = 8) and a roller bearing life to about one tenth (2^10/3 ≈ 10.08). Doubling the speed halves the hours.

Worked example 1: a ball bearing on a motor shaft

A deep-groove ball bearing has C = 32.5 kN. The equivalent load worked out for the duty is P = 6.2 kN at 1,480 RPM.

1. C ÷ P = 32.5 ÷ 6.2 = 5.2419
2. L₁₀ = 5.2419³ = 144.0373 million revolutions
3. L_10h = 144.0373 × 1,000,000 ÷ (60 × 1,480) = 1,622.0418 hours
4. Years of continuous running = 1,622.0418 ÷ 8,760 = 0.1852

The calculator gives the same figures. About 1,622 hours is roughly 68 days of 24-hour running, so a result like this would prompt a check of the load, the bearing choice or the speed rather than a replacement date. Run at half the speed (740 RPM) with the same load and the hours double to 3,244.0837.

Worked example 2: a roller bearing and the effect of load

A roller bearing has C = 75 kN and carries P = 14 kN at 750 RPM.

1. C ÷ P = 75 ÷ 14 = 5.3571
2. L₁₀ = 5.3571^10/3 = 269.0156 million revolutions
3. L_10h = 269.0156 × 1,000,000 ÷ (60 × 750) = 5,978.1232 hours, or 0.6824 years of continuous running

The next table repeats example 1 (C = 32.5 kN, ball, 1,480 RPM) for different equivalent loads. It shows how sensitive the result is to P.

A load 20 percent higher (7.44 kN instead of 6.2 kN) gives 938.7 hours instead of 1,622.0, because 1.2³ = 1.728. Treat load as the input that deserves the most care.

Where the load figure comes from

P is not a quantity you read off a nameplate. Start from the torque the shaft carries, then find the forces it creates. Steady shaft torque is T = P_kW × 1000 ÷ (2π × RPM ÷ 60), which the shaft torque calculator evaluates.

Example: 7.5 kW of shaft power at 1,450 RPM gives T = 7,500 ÷ (2π × 1,450 ÷ 60) = 49.3929 N·m. On a pulley of 0.08 m pitch radius the tangential force is 49.3929 ÷ 0.08 = 617.4 N. That is the net driving force only. Belt tension, the weight of the rotor, gear forces and any axial thrust still have to be resolved into bearing reactions and combined into P with the supplier factors. The torque result excludes starting torque, transient loads and gearbox losses, so a shock or start-heavy duty needs more than a steady value.

Common mistakes

• Entering the shaft power or the applied force in the P field. P is the equivalent dynamic bearing load, not a raw force.
• Mixing units, for example C in kN and P in N. The ratio then becomes meaningless.
• Using the ball exponent for a roller bearing, or the reverse. Check the bearing type selection first.
• Reading L10h as a guaranteed life. One bearing in ten is expected to fail before it.
• Using the average speed when the duty has long slow and short fast periods. Different duty points need separate treatment with the supplier method.
• Ignoring poor lubrication, contamination, high temperature or misalignment because the calculation does not mention them.
• Rounding P down to make a number look acceptable. Because of the exponent, small load changes move the answer a lot.

Checks before you trust the result

• Confirm C on the current data sheet for the exact bearing designation, including its basis and unit.
• Confirm how P was built: loads, bearing factors and the operating case it represents.
• Compare the result with an independent check or the supplier selection software, which can apply modified rating life for lubrication, cleanliness and reliability.
• Do a sanity test: halve the load and the hours should rise by about 8 times for a ball bearing.
• Compare the predicted life with the actual replacement history. A large gap points to a load, lubrication or installation issue rather than a calculation error.

The output is a screening estimate. It does not replace the manufacturer method or site procedures.

Frequently asked questions

Is L10h the time to replace the bearing?

No. It is the time by which about 10 percent of a population is expected to have failed under ideal conditions. A replacement interval also depends on consequences of failure, condition monitoring and the operating environment.

Why is a roller bearing exponent 10/3?

The basic life equation uses a different load-life exponent for line contact than for point contact. The calculator applies 3 for ball bearings and 10/3 for roller bearings.

Can I use the result for 99 percent reliability?

Not with this tool. It gives the 90 percent basic rating life only. Higher reliability needs an adjustment factor from the bearing supplier.

What if the load varies?

Use an equivalent load that represents the whole duty cycle, built with the supplier method, or evaluate each duty point separately. A single average value can mislead.

Recording data for bearing decisions

A life estimate is only as useful as the history behind it. For each critical bearing, record the designation, C, the P used and how it was derived, the speed, the date installed, the lubricant and the date and reason of each removal. When a bearing is replaced, note running hours since installation next to the predicted L10h. After a few events you can see whether predictions are close or consistently high. Routine checks on temperature, noise, vibration and lubricant condition fit the inspection checklist template, with a Fail result creating an owned action. Use the same record to attach failure causes, which feed an RCA and a reliability analysis such as Weibull and FMEA.