Khash was reading below NKE catalog after my company got the chance to represent NKE in Oman, I decided to read cover to cover and in details.
Beyond the Bearing Number: What My Cover-to-Cover Study of the NKE General Catalogue Reveals
From load spectra and contact geometry to operating clearance, lubrication and installation control.
By Khash, MLE, CLS, MLA III, MLT II, VIM, VPR
A bearing can match the required bore, outside diameter and width—and still be unsuitable for the application.
It can have a high dynamic load rating and still operate below its required minimum load. It can be supplied with increased internal clearance and still lose too much clearance after mounting. It can pass a fatigue-life calculation and still suffer from sliding, contamination or inadequate lubrication.
These are the distinctions I concentrated on when reading the NKE General Catalogue cover to cover.
I did not approach it simply as a product directory. I approached it as an interconnected engineering reference: how the bearing is constructed, how loads pass through it, how its operating condition changes after installation, and which assumptions must remain valid for its calculated performance to mean anything.
My central conclusion is this:
A technically sound bearing recommendation must define more than the component. It must define the conditions under which that component is expected to function.
1. Dimensional interchangeability is only the first level of comparison
Boundary dimensions establish whether a bearing can occupy the available space. They do not fully define its internal geometry, contact angle, cage guidance, clearance, sealing arrangement or suitability for a particular load distribution.
The NKE angular contact range illustrates this clearly. Its reinforced BE design uses larger balls to increase load ratings while retaining the same overall dimensions. An unchanged dimensional envelope therefore does not imply an unchanged internal design.
The designation system requires the same attention. In single-row angular contact bearings, B identifies a 40° contact angle. In the cylindrical roller bearing chapter, B identifies a special crowned inner-ring raceway. In the insert-bearing range, B identifies a spherical outer ring.
A letter is not a complete specification. Its meaning depends on the bearing family and the complete designation.
For a replacement assessment, I would therefore compare the actual design features—not merely translate a string of suffixes. The comparison must establish whether the proposed bearing preserves the required load-carrying function, guidance, clearance and installation conditions.
Matching the dimensions answers, “Will it fit?” Application engineering must also answer, “Will it function as required?”
2. Contact geometry explains both capability and limitations
The distinction between ball and roller bearings is not simply the shape of the rolling element.
Under load, the nominal point contact of a ball becomes a small contact area. A roller develops an elongated contact area. The NKE discussion connects these differences to load capacity, friction, rigidity and sensitivity to misalignment.
The elongated contact of a roller is useful for supporting high loads, but it also makes the distribution of contact pressure along the roller important. Misalignment can concentrate loading towards an edge. Roller profiling and modified raceway geometry are intended to reduce those concentrations; they do not create unlimited tolerance for shaft deflection or mounting error.
The same reasoning applies to full-complement designs. Removing the cage permits more rolling elements within the available section, increasing load-carrying capability. However, adjacent rolling elements then contact one another, introducing additional friction and limiting speed capability.
My reading is that bearing selection is a balance between contact geometry, load distribution, friction and kinematics.
A design feature that improves one aspect of performance can introduce a qualification elsewhere. Higher load capacity, higher speed and greater misalignment accommodation cannot simply be assumed to increase together.
3. Static capacity, fatigue life and service life answer different questions
The basic static load rating, C₀, is not a fracture load or a promise of zero permanent deformation. Its reference condition is associated with a defined small permanent deformation at the most heavily loaded rolling-element/raceway contact.
The catalogue relates this reference deformation to approximately 0.0001 times the rolling-element diameter. For an illustrative 20 mm rolling element, that corresponds to approximately 2 µm.
The static assessment uses:
S₀ = C₀/P₀
P₀ is the maximum equivalent static load, determined using the appropriate bearing-specific relationship. The acceptable S₀ must reflect the bearing type, running-accuracy requirement and any additional product-specific restrictions.
Dynamic rating life addresses fatigue under repeated loading:
L₁₀ = (C/P)ᵖ
L₁₀h = [10⁶/(60n)] × (C/P)ᵖ
Here, C and P use the same force units, n is rotational speed in rpm, L₁₀ is in millions of revolutions, and L₁₀h is in operating hours. The exponent is 3 for ball bearings and 10/3 for roller bearings.
L₁₀ corresponds to a 90% fatigue-survival level for a population of identical bearings operating under the stated conditions. It is not a guaranteed life for an individual bearing.
Actual service life is a broader outcome. Lubricant deterioration, unsuitable clearance, mounting damage, contamination and misalignment can become limiting before the calculated fatigue life is reached.
I therefore treat a static check, a fatigue-life calculation and a service-life assessment as related—but distinct—engineering tasks.
4. The load spectrum matters more than an arithmetic average
The exponent in the fatigue equation makes bearing life highly sensitive to the load used in the calculation.
As an illustrative sensitivity check, increasing the equivalent load by 20%, with all other inputs unchanged, gives:
Ball-bearing life ratio = (1/1.2)³ ≈ 0.579
Roller-bearing life ratio = (1/1.2)^(10/3) ≈ 0.545
The calculated life falls by approximately 42% and 46%, respectively.
This is why I paid close attention to NKE’s treatment of variable operating conditions. For a stepped duty cycle, its mean-load relationship is:
Fₘ = [Σ(Fᵢᵖ × nᵢ × tᵢ) / Σ(nᵢ × tᵢ)]^(1/p)
Fᵢ is the bearing load during interval i, nᵢ is the corresponding speed, and tᵢ is the duration. Combined loading must first be represented appropriately for the selected bearing.
Consider an illustrative ball-bearing duty cycle at constant speed: 80% of the time at 10 kN and 20% at 30 kN.
The arithmetic average is 14 kN. The fatigue-equivalent mean load is:
Fₘ = [0.8 × 10³ + 0.2 × 30³]^(1/3) ≈ 18.37 kN
Using 14 kN instead would overestimate basic rating life by approximately 2.26 times. Although the high-load interval occupies only 20% of the operating time, it contributes approximately 87% of the load-power sum in this calculation.
This is not a field case; it is an illustration of the equation’s sensitivity.
For me, it explains why understanding the duty cycle is not optional. A short period of high loading can dominate the fatigue assessment.
5. Minimum load is an independent requirement—not a footnote to maximum capacity
A lightly loaded bearing can produce an impressive fatigue-life result while failing to meet its kinematic requirements.
NKE explains that sufficient loading is needed to promote proper rolling-element motion. Underloading can permit excessive sliding, leading to wear, smearing, increased noise, vibration and temperature. The basic fatigue-life calculation does not account for this failure mechanism.
The catalogue’s general reference values are 1% of Cr for caged radial ball bearings, 2% for caged radial roller bearings, and 4% for full-complement radial designs, unless the specific product instructions state otherwise.
For an illustrative caged radial roller bearing with Cr = 300 kN, the 2% reference corresponds to 6 kN. Selecting that bearing for an application providing only 2 kN at a relevant operating condition would therefore require further investigation, despite the apparently generous capacity-to-load ratio.
Thrust bearings demand their own checks. The NKE spherical roller thrust bearing instructions, for example, specify a minimum axial load of 5% of the dynamic load rating.
My selection review must therefore examine both ends of the operating range: the maximum loading that governs capacity and the minimum loading needed for correct rolling behaviour.
Oversizing is not automatically a reliability improvement.
6. Modified life connects the load calculation to lubrication and cleanliness
The modified rating-life relationship is:
Lₙₘ = a₁ × aISO × L₁₀
The reliability factor a₁ and the operating-condition factor aISO serve different purposes.
In the reviewed catalogue, a₁ changes from 1.00 at 90% reliability to 0.25 at 99% reliability. With the remaining inputs unchanged, the calculated life associated with the higher survival probability is one-quarter of the corresponding 90%-reliability value.
The aISO assessment considers lubrication, contamination and bearing material through the catalogue’s prescribed method. It requires more than a favourable C/P ratio.
I find this important because it connects design assumptions to maintenance conditions. The cleanliness and lubrication entered into an assessment must correspond to conditions that the actual arrangement can achieve.
NKE describes how particles passing through the contact can produce residual indentations. These act as local stress raisers, affecting fatigue performance. The significance of contamination is therefore not confined to visibly abrasive wear; it also changes the condition of the surfaces carrying repeated contact stresses.
My engineering interpretation is that a modified-life result should be accompanied by its assumptions. A numerical improvement is not meaningful unless the lubrication and cleanliness conditions behind it are also understood.
7. Operating viscosity—not the label on the grease container—determines the viscosity ratio
The catalogue’s worked example considers a 6210 bearing, with a 50 mm bore and 90 mm outside diameter, operating at 1,000 rpm and 70°C.
Its mean diameter is:
dₘ = (d + D)/2 = 70 mm
The required reference viscosity is approximately 16 mm²/s. For the illustrated grease base oil, with a nominal viscosity of 68 mm²/s at 40°C, the estimated operating viscosity at 70°C is approximately 18 mm²/s.
Therefore:
κ = ν/ν₁ = 18/16 = 1.125
The relevant comparison is 18 against 16—not 68 against 16.
NKE regards this as a suitable selection in terms of viscosity for the stated conditions. That conclusion does not independently establish correct lubricant delivery, contamination control or the suitability of every other bearing feature.
The distinction between viscosity and grease consistency is equally important. NLGI grade describes consistency; it does not define base-oil viscosity at operating temperature.
I would therefore assess the base oil, its temperature-dependent viscosity, the grease consistency, the operating speed and the lubrication method together. The speed characteristic n × dₘ is also useful, but the catalogue explicitly treats its general lubricant speed-capability values as guidance rather than unconditional product approval.
8. Bearing fits must follow the load relative to each ring
“Rotating shaft” is not, on its own, a complete basis for selecting a fit.
The relevant distinction is how the load moves relative to the bearing ring. A ring under circumferential loading has successive portions of its raceway loaded and tends to creep on its seating unless adequately secured. A ring under point loading has a different seating requirement.
A stationary outer ring can therefore require an interference fit when the load direction rotates relative to it. The load relationship—not merely whether the ring visibly rotates—governs the assessment.
NKE also makes clear that axial clamping is not a substitute for the required radial fit.
The numerical tolerance stack matters. Its 75 mm j5 shaft example combines shaft deviations of +6/−7 µm with a normal-tolerance bearing bore of 0/−15 µm. The resulting fit can range from 21 µm interference to 7 µm clearance.
This illustrates why a tolerance designation should be translated into its possible assembled conditions.
Diameter is only one part of the seating specification. Roundness, cylindricity, shoulder perpendicularity and surface finish also require control. The catalogue explains that thin bearing rings respond to the form of their supporting seats and that rough seating surfaces are susceptible to settling and fretting-related damage.
My practical conclusion is to verify the seat—not simply assume that a nominal fit designation describes its actual condition.
9. Operating clearance is a calculation and verification problem
CN, C3 and C4 describe manufactured clearance groups. They do not directly state the clearance that will exist after mounting and heating.
An interference fit can expand the inner ring or contract the outer ring. A hotter inner ring can produce a further clearance reduction. NKE’s approximate fit-related estimates are:
Inner-ring contribution ≈ 0.80 × probable shaft interference
Outer-ring contribution ≈ 0.75 × probable housing interference
For the stated steel/cast-iron conditions, its thermal approximation is:
ΔCₜ = 1,000 × α × dₘ × ΔT
ΔCₜ is in micrometres, dₘ is in millimetres, α is approximately 12 × 10⁻⁶ K⁻¹ for steel, and ΔT is the inner-to-outer-ring temperature difference.
For an illustrative dₘ of 100 mm and an inner ring 20°C hotter than the outer ring:
ΔCₜ ≈ 24 µm
Now suppose the measured initial clearance is 65 µm, the assumed probable shaft interference is 25 µm, and the housing interference is 20 µm. Combining these catalogue approximations gives:
Estimated remaining clearance ≈ 65 − 20 − 15 − 24 = 6 µm
That is a screening calculation, not an acceptance criterion. It illustrates how a substantial initial clearance can become small after installation and heating. The appropriate analysis must respect the actual construction and the limitations of the approximations.
The catalogue also warns that light-metal housings require particular care because their thermal expansion differs from steel.
Selecting C3 does not complete the clearance assessment. It establishes an initial condition that still has to be evaluated.
10. The non-locating function must exist physically
A locating bearing establishes the shaft’s axial position. A non-locating position must permit the required change in length without unintentionally constraining the arrangement.
NKE distinguishes two fundamentally different ways of achieving this.
With N- or NU-type cylindrical roller bearings, axial displacement is accommodated within the bearing. This can permit interference fits on both rings where the application requires them.
With the illustrated arrangement using two spherical roller bearings, the non-locating bearing instead requires axial movement relative to its housing seat. The housing fit and surrounding construction must permit that movement.
Self-alignment does not answer the axial-displacement question. A bearing’s ability to accommodate angular misalignment is different from an arrangement’s ability to accommodate a change in shaft length.
When reviewing a drawing, I would therefore identify the actual displacement path. I would not accept “floating end” as a complete explanation without understanding which surfaces move relative to one another.
The catalogue also distinguishes cross-locating and preloaded arrangements from the conventional locating/non-locating configuration. The arrangement must be assessed as designed, rather than forcing every machine into one simplified layout.
11. Angular contact pairs require coordinated geometry, matching and load calculation
The angular contact range provides a clear example of why two bearings cannot be treated as an arbitrary pair.
For the stated single-row designs, NKE identifies contact angles of 30° with A, 40° with B, 15° with C and 25° with E. The BE execution is separately described as a reinforced internal design.
Single-bearing B/BE designs must not be assumed to have the controlled face relationships of universally matched bearings intended for adjacent pairs or sets.
The universal designs distinguish axial-clearance classes CA, CB and CC from preload classes GA, GB and GC. For the specified universal 72 B/BE and 73 B/BE bearings with bores over 50 mm up to 80 mm, CB corresponds to 26–38 µm axial clearance under the catalogue’s stated unmounted-pair conditions.
That is not an automatic prediction of hot operating clearance.
Arrangement also matters. Back-to-back pairs provide a larger effective support spread and greater moment rigidity. Face-to-face arrangements have a smaller support spread; the catalogue specifically warns that shaft heating can reduce their axial clearance or increase preload. Tandem bearings share thrust in one direction and require an opposing arrangement for guidance in the other direction.
For the specified side-by-side universal pairs:
Cr,pair = 1.62 × Cr,single
C₀r,pair = 2 × C₀r,single
NKE also applies a 20% reduction in the stated speed ratings relative to single bearings.
For the applicable 40° universal pairs arranged back-to-back or face-to-face, the dynamic equivalent load is:
P = Fr + 0.55Fa, when Fa/Fr ≤ 1.14
P = 0.57Fr + 0.93Fa, when Fa/Fr > 1.14
Here, Fr and Fa are the forces acting on the pair. These equations must be used with the corresponding pair rating—not mixed with a single-bearing rating or a different arrangement’s factors.
This coordination between physical arrangement and calculation is one of the most valuable lessons in the catalogue.
12. Cylindrical roller bearing axial capacity is constrained by flange contact
N and NU designs provide the non-locating function without carrying axial load. NJ accommodates axial load in one direction. NUP, and the appropriate NJ/HJ combination, provide axial guidance in both directions.
However, axial guidance does not mean unrestricted thrust capacity.
In the designs that carry thrust, the axial force creates sliding contact between roller end faces and guiding flanges. Lubrication, heat dissipation and adequate flange support therefore become central to the permissible axial load.
NKE specifies that both of the following limits must be respected for the applicable designs:
Fa,max ≤ 0.4Fr
Fa,max ≤ 0.1Cr
These limits also require the stated conditions of optimum lubrication, no shock loading, optimum heat dissipation and adequate axial flange support.
For an illustrative Cr of 300 kN and Fr of 50 kN, the two upper bounds are 30 kN and 20 kN. The lower value governs: 20 kN, subject to those conditions and the remaining application checks.
Two further details connect design to workshop practice. NKE warns against interchanging components of its C1-clearance cylindrical roller bearings. It also describes roller drop during assembly: unsupported rollers can hang within their cage pockets, making their end faces vulnerable during shaft insertion.
A guide sleeve can address that assembly risk. A separable bearing still requires a controlled assembly method.
13. Spherical roller bearing execution includes cage guidance—not just cage material
“Brass cage” is an incomplete description when the application depends on the way the cage is guided.
NKE’s MB spherical roller bearing design includes a fixed centre flange and an inner-ring-guided machined brass cage. The SQ34 vibration execution instead includes an outer-ring-land-guided solid brass cage, reduced outside-diameter tolerances and reduced bore tolerances for cylindrical-bore versions.
The SQ34 designation therefore represents a coordinated set of design requirements. It should not be reduced to a statement about cage material alone.
The lubrication features deserve similar precision. NKE describes the W33 outer-ring groove and lubrication holes as standard in the discussed spherical roller range except series 213.
Taper geometry is another part of the execution. The usual K taper is 1:12, while the discussed 240 and 241 series use the shallower K30, 1:30 taper.
These distinctions lead me to ask for the complete specification when reviewing a demanding application: cage design and guidance, bore form, clearance range, lubrication features and the applicable special execution.
I would not assume that retaining the same nominal bearing size preserves all of those characteristics.
14. Tapered-bore mounting must satisfy two clearance conditions simultaneously
Driving a tapered-bore bearing onto its seat expands the inner ring and reduces radial clearance. Installation therefore requires control of the mounting displacement and the resulting clearance—not merely tightening until the assembly feels secure.
The NKE guidance specifies both a permitted clearance reduction and a minimum remaining mounted clearance.
For the stated spherical roller bearing bore band over 100 mm up to 120 mm, the recommended reduction is 50–70 µm, while the minimum mounted clearance for CN is 50 µm. The tabulated method includes a solid-steel-shaft qualification.
Consider an illustrative bearing with a measured initial clearance of 110 µm. Reducing it by 70 µm leaves:
R₂ = 110 − 70 = 40 µm
The reduction lies within the stated reduction range, but the remaining clearance is below the stated minimum.
That is why the two criteria cannot be checked independently or selectively.
The catalogue also requires a final clearance recheck after securing the locknut. An intermediate reading is not the final installed condition.
For me, this is an important example of translating catalogue information into acceptance criteria: record the starting clearance, control the mounting operation and verify the final condition.
15. Speed and temperature limits belong to the complete arrangement
NKE distinguishes thermal reference speed from kinematic limiting speed.
The thermal reference speed belongs to defined heat-balance conditions, including a 70°C bearing reference temperature and a 20°C ambient reference. It is an input to evaluating thermal speed capability—not an unconditional permission to operate at that speed in every installation.
The kinematic limit considers additional constraints such as running behaviour, mechanical strength, sealing and centrifugal effects. Improved cooling does not, by itself, permit that limit to be exceeded.
The reviewed catalogue instructs a 25% reduction in the relevant tabulated limiting speeds for grease lubrication, with a 60% reduction for cylindrical roller thrust bearings. It also notes that reductions have already been included for the applicable factory-sealed, pregreased bearings. These qualifications must be read together to avoid either missing or duplicating a correction.
The IKOS integral tapered roller bearing illustrates the same principle for temperature. Its rings and rollers are stabilised to 150°C, but the catalogue limits the complete bearing’s continuous operating temperature to 120°C because the seals, polyamide cage and grease impose additional constraints.
IKOS also has controlled ring-face geometry for axial adjustment, while still requiring compliance with maximum clamping-force and snap-ring-joint limits.
A component rating does not replace the assessment of the assembled bearing and its surrounding construction.
16. Lubrication quantity, interval and discharge path must work together
A lubrication recommendation is incomplete when it specifies only a grease name.
The NKE discussion separates initial filling, housing-cavity filling, relubrication quantity, lubricant service life and the discharge of used grease. These describe different parts of the lubrication process.
For new applications without established experience, the catalogue recommends initial relubrication intervals of approximately 50–60% of calculated lubricant service life, followed by careful observation and adjustment.
Its stated mineral-oil-based interval guidance calls for halving the calculated interval for each 15°C increase above 70°C. As an illustrative application of that rule, a 4,000-hour reference interval becomes 2,000 hours at 85°C and 1,000 hours at 100°C—before other applicable adjustments.
That illustration is not a universal grease-life law or a maintenance instruction for an unspecified machine. Moisture, contamination, vibration, lubricant formulation and operating experience remain relevant to the final decision.
The grease must also have an appropriate path through the arrangement. NKE discusses relief passages, used-grease discharge and the avoidance of excessive backpressure. Adding fresh grease without understanding where displaced grease goes does not establish effective relubrication.
Grease conversion requires similar control. The catalogue warns that even apparently similar greases can interact unexpectedly and calls for attention to residual grease in housings, pipes and grooves during a changeover.
My conclusion is to specify the lubricant, delivery method, quantity, interval and discharge arrangement together—and to verify the result during operation.
17. Product-specific restrictions can govern before the headline load rating does
Several less obvious sections demonstrate why the product instructions remain essential.
For spherical roller thrust bearings, the catalogue permits combined loading within the stated condition:
Fr ≤ 0.55Fa
Within that scope, it gives:
P = Fa + 1.2Fr
P₀ = Fa + 2.7Fr
It also distinguishes a static safety requirement of S₀ ≥ 8 for the specified adjacent-part arrangement from S₀ ≥ 4 when both washers have full axial support across their width and the housing washer has satisfactory radial support. The lower requirement is conditional on the support arrangement; it is not a universal substitution.
For cam rollers, the outer ring contacts an external track over a limited area. NKE therefore gives permissible dynamic and static radial loads that account for outer-ring deformation, in addition to conventional bearing ratings. Both aspects need checking.
For self-aligning ball bearings, adding seals can reduce both speed capability and permissible angular misalignment. The discussed open 13 and 23 series permit up to 3° under their stated conditions, whereas the stated sealed designs are limited to 1.5°.
For bearing units, the spherical outside surface provides accommodation of static misalignment. AY and GAY inserts are also not equivalent in maintenance capability: GAY includes relubrication holes, while AY does not.
For adapter sleeves, the bore reference identifies the associated bearing bore rather than automatically identifying the shaft. NKE’s H320 example combines a 100 mm bearing bore with a 90 mm shaft.
These details are not peripheral. They can determine whether an otherwise plausible selection is actually usable.
18. Reliability must survive the journey from storage to operation
The engineering assessment does not end when the correct designation is ordered.
NKE’s storage guidance calls for original packaging, a clean environment, controlled temperature and humidity, and protection against vibration, shocks and condensation. Factory preservation does not eliminate the need to maintain suitable storage conditions. Grease-filled bearings also require attention to lubricant changes during prolonged storage.
During mounting, forces must be transmitted through the ring being fitted—not through the rolling elements. If both rings are fitted simultaneously, the tooling must support the appropriate ring faces without loading projecting cages or rolling elements.
Thermal mounting requires the same discipline. The catalogue states that 90–110°C is usually sufficient, depending on size and fit, while its general instructions prohibit heating standard bearings above 120°C. The actual bearing execution must still be considered. Induction-heater requirements include temperature control and automatic demagnetisation; open-flame heating is explicitly rejected.
Before commissioning a circulating-oil arrangement, the lubrication supply must be established before shaft rotation. Temperature, noise and running behaviour then require observation rather than an assumption that successful assembly proves successful operation.
Dismounting also deserves planning. NKE warns that hydraulically released bearings can move suddenly. Appropriate retention, equipment instructions and precautions are integral to the method.
I would translate these requirements into an installation and commissioning record: the verified bearing execution, seating condition, initial and final clearance where applicable, mounting method, lubricant details and observed operating behaviour.
That is how the engineering assumptions become verifiable work.
My conclusion: the real value is in connecting the requirements
The most valuable outcome of reading the NKE General Catalogue cover to cover was not learning more designations. It was following the connections between its engineering requirements.
Contact geometry affects load distribution. Load distribution affects fatigue calculations. Fits affect ring geometry and clearance. Temperature affects both clearance and lubricant viscosity. The bearing arrangement determines how axial movement is accommodated. Cage and flange designs introduce their own lubrication and operating requirements. Mounting and commissioning determine whether the intended conditions are actually achieved.
My standard for a bearing recommendation is therefore more demanding than a dimensional cross-reference.
I want the complete execution identified, the relevant loading understood, the calculation assumptions stated, the mounting conditions verified and the lubrication arrangement defined. Where a product-specific qualification applies, I want it carried into the recommendation rather than lost during procurement or installation.
The bearing number identifies a component. Application engineering establishes whether that component belongs in the machine.
And that is the difference between reading a catalogue and using it to support reliability.
Khash, MLE, CLS, MLA III, MLT II, VIM, VPR
The numerical guidance discussed here applies to the NKE designs and conditions described in the reviewed catalogue. Illustrative calculations are not reported field results. Final application decisions require confirmation of the exact bearing specification, machine requirements and applicable manufacturer instructions.
