EN

Bearing life calculator (L10 / L10h)

A rolling bearing does not wear out like a rubbing part: it eventually spalls through metal fatigue, after millions of over-rollings by the balls or rollers. ISO 281 quantifies this endurance with the basic rating life L10 — the life reached by 90 % of a batch of bearings before the first signs of fatigue appear.

Life is governed above all by the C/P ratio between the bearing's load capacity and the actual load: because it is raised to the power 3 (ball bearings), doubling that ratio multiplies life by 8. A slightly oversized bearing therefore lasts a very long time.

Bearing, load and speed

Bearing

Sets the exponent p: 3 for ball bearings, 10/3 (≈ 3.33) for roller bearings.

From the bearing catalogue (usually written C).

The load applied (see the equivalent load calculator).

Rotational speed of the shaft carried by the bearing.

The first asterisk on a term jumps to the glossary.

What L10 actually means

A rolling bearing does not wear down the way a plain bush does. Each pass of a ball or a roller puts a Hertzian stress into the raceway. After enough of those passes a crack starts under the surface, and a flake of steel lifts away. That flake is spalling.

On a bearing mounted straight and kept lubricated, spalling is the ordinary end of the part. Fatigue. Not a scandal in the steel.

ISO 281 calls the result of this calculation the basic rating life L10*. Counted in millions of revolutions, it is the life at which 90 % of a batch remain free of the first spall. Reliability* at that mark is 90 %: a 10 % chance that fatigue has already shown. Part of the batch runs far past L10. The unlucky 10 % never reaches it.

Not a guarantee for the single bearing in a housing. Not a mean life either.

L10h is the same life in hours, at one stated speed. Specifications, tenders and motor datasheets are written in hours. The damage accumulates in revolutions. At 3,000 rpm a shaft spends a given L10 twice as fast as at 1,500 rpm. Hours with no speed beside them do not describe a life.

C, P and n — the three inputs

Three numbers, and the formula is short. The metallurgy already sits inside them.

  • C — basic dynamic load rating*: the catalogue figure, in the same units as the load. By definition, the constant equivalent load that produces a life of one million revolutions. Not a “safe working load”. A bearing can run with P above C; L10 drops below 1 million revolutions, and the result block on this page states that.
  • P — equivalent dynamic load*: the load the bearing sees. Purely radial, P = Fr. Radial and axial together, P = X·Fr + Y·Fa, from the equivalent load calculator. Feeding Fr in as P while a helical gear pushes Fa is the usual route to a short bearing.
  • n — rotational speed in rpm. It leaves L10 in revolutions unchanged. It turns that stock of revolutions into hours. The faster the shaft, the sooner each million-revolution block is used up.

L10 = (C / P)p  ·  L10h = 106 / (60 × n) × (C / P)p

C and P belong in one unit. The form converts a mixed entry; a spreadsheet that pairs kN with N does not. n is the speed of the shaft in the bearing, not the motor nameplate when a belt or a gearbox sits between them. A duty that alternates between idle and a peak needs a cubic-mean P on a ball bearing, or a 10/3-mean P on a roller bearing. One “typical” load is a wish. It is not that mean.

Why the exponent p is so brutal

Life runs as the power p of the ratio C/P. p = 3 for ball bearings, where the contact is nearly a point. p = 10/3 ≈ 3.33 for roller bearings, where the contact is a line and the stress is spread a little further along the roller. That exponent is what makes bearing sizing sharp.

Halve P, or double C/P, and a ball bearing’s life is multiplied by 23 = 8. A roller bearing’s life is multiplied by about 10. The other way is harsher. A 25 % overload — P up by a quarter — cuts ball life to 0.83 ≈ 0.51 — half of it, gone.

A millimetre of extra section, or a slightly quieter load path, outweighs a heroic steel.

The form opens on values meant as a feel for the scale: C = 30 kN, P = 5 kN, n = 1,500 rpm, ball bearing.

C/P = 6  →  L10 = 63 = 216 million rev

L10h = 216 × 106 / (60 × 1,500) = 2,400 h

Two thousand four hundred hours is a household-appliance life, not a mill-motor life. The ratio 6, cubed, is those 216 million revolutions; at 1,500 rpm they come to 2,400 h. At the same speed, a target of 40,000 h on a ball bearing needs (C/P)3 ≈ 3,600, which is C/P ≈ 15. A 30 kN bearing calls for P near 2 kN, not 5. Below about C/P = 2 this page flags a low margin. A C/P of 6 to 10 is a common aim where a long life is the point, and plant that runs 24/7 often sits higher.

Hours people size to

L10h is set against a target, not against a moral absolute. The target follows from the way the machine is used, from how painful a change-out is, and from whether a person is standing beside it when the bearing lets go.

  • Household appliances, power tools — a few thousand hours are enough. The product is obsolete before the raceway is.
  • Industrial electric motors — 20,000 to 40,000 h is the band commonly aimed at. Two to five years at one shift, or a year-plus of non-stop running.
  • Machines running 24/7 (fans, pumps, ID fans on a kiln) — 50,000 to 100,000 h, and beyond. What is being bought is the crane visit, not the bearing.
  • Gearboxes, machine tools — sized to the service life expected of them, often 30,000 h, unless a spindle specification says otherwise.

Those hours are running hours at the load that was entered. A pump that works 3,000 h a year on duty and sits for the rest of the year is not the case the formula treats as 8,760 h/year. Calendar years have to become running hours before a claim of twenty years means anything. A weekend machine with a brutal peak is sized on that peak, through P, not on a gentle annual average.

ISO 281 carries no shutdown calendar. Hours per year belong on the calculation sheet. The figure this page returns is not a calendar.

What basic L10 leaves out

Basic rating life accounts for load, speed and the catalogue capacity. That is the whole of it. The calculation assumes a clean mounting, a lubricant film that keeps the surfaces apart, ordinary workshop cleanliness, and a moderate temperature.

On a real shaft, one of those assumptions usually fails early.

ISO 281 therefore defines a modified rating life, using the life adjustment factor a1* and the factor aISO*.

Lnm = a1 · aISO · L10

a1 carries reliability. aISO rolls lubrication, contamination and the fatigue load limit into one multiplier. Real life can sit well above L10 when the film is thick and clean, or well below it — grit, a dry start, a housing bored out of true. The load sum is the part that can be finished at a desk. Lubrication and sealing decide a bearing far more often than a 10 % miss on P.

A handsome L10h does not excuse a dry assembly, a lip seal running on rust, or a C0 too small for the standing load. The static safety factor s0 belongs on the peak at rest. Fatigue and indentation are two failures, two formulae, and two catalogue columns, C and C0.

a1 and aISO: the modified rating life in ISO 281

Basic L10 is the 90 % survival figure. A requirement written as 99 % reliability is a request for L1. ISO 281 supplies a1, and that factor scales L10 down to the stricter reliability.

  • 90 % (L10) — a1 = 1. The default, and the figure this calculator prints.
  • 95 % (L5) — a1 = 0.64.
  • 96 / 97 / 98 % — a1 = 0.55 / 0.47 / 0.37.
  • 99 % (L1) — a1 = 0.25. Three quarters of the basic hours are gone, on reliability alone.

aISO is the other lever, and it can sit on either side of 1. It is read from ISO 281, or from a manufacturer’s reading of the same standard — SKF’s aSKF, Schaeffler’s eC* diagrams — as a function of the viscosity ratio κ*, a contamination factor, and the ratio of the fatigue load limit Pu to P. Pu is a catalogue value, smaller than C: the load below which, in the model, subsurface fatigue never starts. A clean oil film on a lightly loaded bearing can push aISO to 2, to 3, even toward the cap of 50 on the diagrams in the standard. Dirt and a starved film push it toward 0.1, and a basic 40,000 h becomes 4,000 h.

This page stops at L10 on purpose. A modified life needs the operating temperature, an oil grade, a cleanliness class, and Pu. Without those inputs, aISO is a guess dressed up as a factor. Applying a1 by hand is plain arithmetic: L5 ≈ 0.64 L10. Setting aISO = 2 because a salesman called the grease “premium” is not.

Lubrication and the viscosity ratio κ

The oil film is a few tenths of a micrometre thick. It is also what separates a bearing that outlives the shaft from one that runs rough within six months. ISO 281 captures that film as a viscosity ratio: kinematic viscosity at the operating temperature, divided by the viscosity required for an adequate film at that speed and that pitch diameter.

κ = ν / ν1

  • κ < 1 — film too thin. Boundary or mixed lubrication. aISO falls, wear can join fatigue, and a dry start marks the race.
  • κ ≈ 1 — about the reference film that basic L10 assumed. Neither a failure nor a gift.
  • κ ≈ 2 to 4 — full separation on most rolling contacts. This is where aISO earns its place, and service life pulls ahead of L10.
  • κ > 4 — extra viscosity stops adding life and starts churning. Heat, drag, grease thrown off the track. High-speed spindles do not want that regime.

ν1 falls as speed rises. A fast bearing builds a film with less viscosity, which is why a 3,000 rpm motor can run a thinner grease than a 50 rpm kiln trunnion. Temperature moves ν the other way. An ISO VG 220 that looks heavy in the can is a VG 32 once the housing sits at 80 °C. The grade is chosen at the temperature the bearing will reach, not at the 40 °C printed on the lubricant sheet.

Grease is an oil held in a thickener. The bearing lives on the oil that bleeds out, not on the soap.

A line that reads “lithium EP2” does not fix κ until the base-oil viscosity and the running temperature are known. A sealed-for-life 6204 in a clean fan is a different lubrication problem from an open spherical roller in a paper-mill dryer. Same life formula. A different refill story.

Contamination, sealing, and why clean oil outlives a bigger C

Grit in the contact is a local Hertzian spike the life formula never included. ISO 281 folds that spike into a contamination factor — eC in many catalogues — and that factor feeds aISO.

The scale is steep.

  • Laboratory cleanliness — filtered circulating oil, eC near 1. Rare on a real shop floor.
  • Good circulating oil, with filters — eC around 0.5 to 0.8. Gearboxes that are maintained.
  • Normal cleanliness — grease, decent seals, eC around 0.2 to 0.5.
  • Contaminated — coarse seals, a dusty plant, eC from 0.1 down to 0.01. aISO wrecks Lnm no matter how large C is.

Doubling C in order to buy life is an eightfold gain in L10 on a ball bearing. A ruined eC can take more than that back. A better seal, a flinger, a housing vent that breathes through a filter, a grease that stays in place: those change life. They are not accessories. On the shop floor, the bearing that dies at 10 % of its L10h is usually dirty or dry, not undersized.

Water is contamination in the same sense. A few hundred ppm in the oil breaks the film and starts etching.

Outdoor plummer blocks need a cover that covers, and a grease that water will not wash into an emulsion. A labyrinth does not save a housing that holds standing water.

Grease versus oil when the hours start to matter

Most industrial bearings run on grease. It is cheap, it seals in its own right, and it needs no tank and no pump.

Oil comes in when speed, heat, or the need to carry debris away make grease a poor fit.

  • Grease — electric motors, fans, ordinary gearboxes, agricultural discs. The relubrication interval shrinks quickly as speed and temperature rise (the rules of thumb in any SKF or Schaeffler handbook plot hours against n·dm). Over-greasing cooks the bearing from the inside. Under-greasing starves it. Either failure looks, afterwards, like a wrong life calculation.
  • Oil bath, splash — medium-speed gearboxes. The oil level sits at about the centre of the lowest rolling element, not halfway up the shaft. Too high, and the oil churns. Too low, and the top row runs dry.
  • Circulating oil, oil jet, oil-air — high speed, high heat, large spherical rollers on a vibrating screen. The oil is a coolant as much as a lubricant. κ is a design input at that point, not a hope.

A sealed-for-life deep-groove bearing (2Z, 2RS) carries the grease it was filled with. An L10h of 80,000 h is a paper figure if the grease life at that temperature is 15,000 h. Grease life, not fatigue, sets the change-out. A high-temperature grease — polyurea, or some calcium sulfonate complexes — buys hours back. It does not change C.

Oil is no quality upgrade on a 1,500 rpm 6205 that will never see 70 °C. The swap adds a leak and a level to check.

Lubrication has to match n, dm and temperature, and the way the machine will be serviced. The factor aISO in ISO 281 follows that choice. It will not forgive a poor match.

Glossary

Basic rating life L10
Life of a rolling bearing in millions of revolutions, to ISO 281. At that life, 90 % of a large batch remain free of the first fatigue spall; 10 % have already spalled. Not a guarantee for one bearing, and not a mean. L10h is the same life in hours at a stated speed.
Basic dynamic load rating C
Catalogue load capacity. The constant equivalent load that gives a life of one million revolutions. Not a safe working load: with P above C, L10 falls below one million revolutions. Distinct from the static capacity C0.
Equivalent load P
The load the life formula treats the bearing as seeing. Purely radial, P = Fr. With an axial component, P = X·Fr + Y·Fa. C and P enter (C/P)p in the same unit.
Life adjustment factor a1
Multiplier in the modified rating life Lnm = a1 · aISO · L10 that accounts for reliability. It equals 1 for L10 (90 %). It is 0.64 at 95 % (L5), 0.55 / 0.47 / 0.37 at 96 / 97 / 98 %, and 0.25 at 99 % (L1). It shortens the stated life. It does not change the bearing.
Reliability
The share of a batch free of first spalling at the life being quoted. Basic rating life is the 90 % figure. A 99 % target is a shorter life, L10 multiplied by a1, not a different formula.
Factor aISO
Multiplier in the modified rating life for lubrication, contamination and the fatigue load limit Pu. It can lie above 1 or well below. Read from ISO 281, or from a manufacturer’s diagram such as SKF’s aSKF. A clean thick film on a light load can push it to 2, to 3, or toward the cap of 50. Dirt and a starved film push it toward 0.1.
Viscosity ratio κ
κ = ν / ν1: kinematic viscosity at operating temperature, divided by the viscosity needed for an adequate film at the speed and pitch diameter in question. Near 1 is the reference film behind basic L10. Below 1 the film is thin. From about 2 to 4, most rolling contacts run fully separated. Above 4, extra viscosity churns more than it adds life.
Contamination factor eC
Catalogue factor for hard particles in the contact. It feeds aISO, and the same symbol labels Schaeffler diagrams of the modified-life adjustment. Near 1 is laboratory cleanliness with filtered oil. Around 0.5 to 0.8 is filtered circulating oil in good condition; around 0.2 to 0.5 is ordinary greased duty. From 0.1 down to 0.01 the machine is dusty or coarsely sealed, and Lnm falls with aISO.

Related calculators