Bearing Analysis Calculator — L10 Life, Modified Rating Life & Static Safety Factor (ISO 281)

The primary engine follows ISO 281:2007 — the international standard for rolling bearing dynamic load ratings and rating life. The L10 basic life and modified rating life Lnmh use the ISO 281 framework including the analytic aISO factor (§9.3.3), the reference viscosity from ISO 281:2007 Annex B, and the reliability factor a₁ from Table 2. The static safety factor follows ISO 76. Two secondary checks (heat generation by a constant-coefficient friction estimate with a published friction coefficient per bearing family, and dmn speed-limit thresholds) are engineering estimates rather than ISO-codified limits, and are labelled accordingly in the app.

L10 is the basic ISO 281 life at 90 % reliability — it depends only on the load ratio C/P and bearing type, and does not account for lubrication or contamination. Lnmh (the modified rating life) multiplies L10 by the reliability factor a₁ and the ISO 281 life modification factor aISO, which captures how well the bearing is lubricated (via the viscosity ratio κ) and how clean the environment is (via the contamination factor ηc and the fatigue load limit Cu). In well-lubricated, clean conditions Lnmh can be several times larger than L10; in poorly lubricated or contaminated conditions it can be much shorter.

The viscosity ratio κ = ν/ν₁ compares the actual kinematic viscosity of the oil at operating temperature to the reference viscosity the bearing needs at its speed and size to form a full elastohydrodynamic film. When κ ≥ 1 the film is complete and aISO is at its maximum for the given contamination level. When κ < 1 the film is incomplete (mixed lubrication), aISO drops sharply and predicted life falls — often dramatically. κ < 0.5 represents severe boundary lubrication, which the calculator flags as a warning and where the ISO 281 model reaches its lower validity limit.

Some can, and yes — but as a separate check, because the thrust does not go through the rolling contact. A cylindrical roller bearing carries axial load on the sliding contact between the roller ends and the ring flanges, so the equivalent load P stays equal to the radial load Fr whether or not there is thrust. That is what the standards give for a bearing of this kind, and they then say the axial capacity depends on the design and must come from the bearing's maker. Whether the bearing can carry any at all is decided by its design: an NU, N or NUB has no flange on one ring and carries none in any direction, an NJ or NF carries it in one, an NUP, NP or NH in both — state which you have and the calculator refuses the combination that cannot work rather than reporting a long life for it. Two screens are then applied: a continuous axial load is held to 0.4 times the radial load, and a grease-lubricated bearing is taken to have no axial capacity at all above dm·n = 200 000 mm/min. Enter the permissible axial load from the data sheet and it is checked too; without it the result says the flange's own capacity has not been checked, rather than implying it is fine.

Yes, and by a lot. The contact angle decides when the axial load starts counting towards the equivalent load and how much it counts for, and the three standard angles are far apart on both. Take the calculator's own default bearing, a 7210 (typical C = 32.5 kN) at Fr 2 kN, Fa 2 kN and 1500 rpm: P comes out 3.82 kN at 15°, 2.56 kN at 25° and 2.00 kN at 40°, and the rating life about 6 800, 22 700 and 47 700 hours — seven times longer at 40° than at 15°, on identical loads. Pick the angle from the bearing's data sheet (the designation suffix usually names it: C is 15°, AC is 25°, B is 40°); the calculator will not take it from the designation for you, and it refuses an angle between the three rather than rounding to the nearest, because the step between them is large enough to change the answer several-fold.

Yes, when you give it the opposing bearing's radial load. A tapered roller or angular-contact bearing carries radial load on an angled contact, so it pushes an axial load along the shaft — about 0.5·Fr/Y for a tapered roller, and for an angular-contact bearing a ratio set by its own contact angle — 1.14·Fr at 40°, 0.70·Fr at 25°, 0.40·Fr at 15° — and in the usual opposed (back-to-back or face-to-face) mounting each bearing drives that load into the other. Enter the opposing bearing's radial load in the Applied Loads panel and this bearing is calculated at Fa = max(Fa + induced by the opposing bearing, its own induced load) + preload, the standard adjusted-pair analysis. Leave the field empty and nothing is added, so the axial load you enter is taken as the total for that bearing. The opposing bearing is taken to be the same kind of bearing, and for a tapered roller its induced load is divided by ITS OWN Y: state its three-digit series, or the angle-class suffix on a five-digit metric designation (C or B for a medium-angle variant, D for a steep one — these are far steeper than the plain series and drive much more load; an inch-series bearing is none of them), and leave it unknown to have the steepest single-row tapered roller found in either catalogue assumed, which gives the largest induced load and so never reads a longer life than the pair will deliver.

The calculator covers all five main rolling-element types: deep-groove ball bearings (ISO 281 interpolated X/Y table), angular-contact ball bearings (the published 15°, 25° or 40° factors for the angle you state), cylindrical roller bearings (the axial thrust is carried by the flanges, not the rolling contact, so it is not part of P — the flanges are checked separately), tapered roller bearings (the e/Y factors of the bearing's own series, or its catalogue factors when entered) and spherical roller bearings (Y₁/Y₂ axial-load factors, likewise per series). A built-in catalog spanning the 6000, 6200, 6300, 7200, 7300, NU, 22200, 22300 and 302xx–322xx designation series lets you select a size directly and auto-populate typical conservative C, C0 and dm — computed by MechanixCalc from the boundary dimensions, so replace them with your bearing's data-sheet ratings for final design.

ISO 281 treats the bearing as one spring with a catalogue number on it: you give it an equivalent load P, it gives you a life. It cannot see inside. The internal load distribution solves the bearing's own equilibrium instead — every rolling element is a Hertzian contact, and the rings are displaced radially, axially and in tilt until the summed element loads balance what you applied. That is the load-distribution method of ISO/TS 16281, and it answers questions the rating life cannot: which element carries the worst load and how much, how many of them are carrying anything at all, how far the contact angle opens under thrust, and how much the number moves as the cage rotates. It also produces the bearing's radial, axial and tilting stiffness, which is the one quantity a shaft-deflection or rotordynamic model needs and no catalogue prints. The panel computes the load distribution only; it does not compute the ISO/TS 16281 modified reference rating life, and the life figures on this page remain ISO 281.

Whichever you give it. Ball count, ball diameter, pitch diameter, raceway osculation, free contact angle and diametral clearance are all inputs. If you only know the bore and outside diameter, it fills them with MechanixCalc's typical proportions for a bearing of that envelope and marks the result as derived from an estimate — and it keeps checking: edit any of those numbers by hand and the badge retires itself, because a badge that describes the wrong numbers is worse than no badge. Internal geometry is not usually published, so the estimate is there to make the panel usable, not to replace a data sheet. The ball count in particular is deliberately taken at the low end of the plausible range, which over-states the load per element rather than under-stating it. If the envelope you enter is not a bearing the solver can make sense of, it refuses rather than returning a number, and a refused result never reaches the PDF or the workbook.

Run the internal load distribution at the operating load and read the stiffnesses off the panel, then enter them in the Rotor Dynamics calculator — radial stiffness in N/µm, tilting stiffness in kN·m/rad. One caution that matters more than any other here: the panel reports ONE ROW. A single spindle-size row is worth only a few tens of kN·m/rad in tilt, while a back-to-back or face-to-face PAIR is worth far more, because its axial stiffness acts on the lever arm between the two contact lines. Nothing in either tool converts between them, and neither knows how many rows your support has. Moment restraint can only raise a predicted nose stiffness and critical speed, so an over-stated figure shows more speed margin than you have — which is why the Rotor Dynamics tilting-stiffness inputs are optional and default to zero, the pinned-support assumption, and why leaving them blank is the conservative choice.

You can run it during a free 30-minute preview with no sign-up required. A free 14-day account trial unlocks every calculator on the platform with no credit card. The branded PDF engineering report and saved calculations are included in the free 14-day trial and in every paid plan.