E-scooter rolling-element bearing engineering: ISO 281 L₁₀ rating life, ISO 76 C₀, ABEC/ISO 492 precision, NLGI greases, types, and failure modes
In the articles on frame and fork engineering and motor and controller engineering we mentioned rolling-element bearings as a critical part of the load-bearing structure — angular contact in the headset, deep-groove ball in wheel hubs, 6900-series double-row in the hub-motor stator-rotor interface, and 6900-2RS sealed deep-groove on the motor shaft when diagnosing whining noise. In the article on IP protection we showed that a rotating sealed shaft in a hub motor caps the IP rating at IPX5 — a direct consequence of bearing and lip-seal physics. In tire engineering — how rolling resistance decomposes into tire hysteresis losses and hub bearing friction (typically 0.1–0.3 % of rolling resistance). Bearings are transparently present everywhere — and never described as a standalone engineering-axis discipline.
This is the thirteenth engineering-axis deep-dive in the guide series (after helmet, battery, brakes, motor and controller, suspension, tires, lighting, frame and fork, display and HMI, charger, connectors and wiring, IP protection) — adding the tribology axis as the integrator of all rotational loads: everything the motor produces, the frame holds, and the tire transfers to the road, passes through bearing rolling elements. The owner cannot change the frame’s σ_y nor the tire compound after purchase, but can replace the bearing in half an hour and for a few dollars — making bearing engineering the most accessible engineering axis for DIY maintenance.
This article addresses: why 6001-2RS in the Xiaomi M365 (12×28×8 mm) is not a random number but an ISO-standard designation with specific geometry and C = 5.4 kN dynamic load rating; why L₁₀ = (C / P)^p × 10⁶ revolutions gives ~3700 hours at P = 1 kN for a ball bearing (p = 3) and only ~370 hours at P = 2 kN (cubic drop); why NLGI 2 with a lithium-complex thickener is the universal default for scooter hubs, and why NLGI 0 in the hub-motor shaft is the first thing to evaporate after 5 years; why ABEC 7 in skate ads is mostly marketing noise for scooters spinning at 1500–2500 RPM (not 30 000 RPM like a high-speed spindle); and why false brinelling appears on bearings after winter storage in a garage next to a washing machine — not from mileage.
Prerequisite — understanding frame structure and headset bearing, hub-motor architecture, maintenance practice, and the pre-ride safety check.
1. Why bearings are their own engineering discipline
Geometrically, a rolling-element bearing is a 5-component mechanism that replaces the μ_dry ≈ 0.4–0.8 of two sliding metal surfaces with the μ_roll ≈ 0.001–0.003 of rolling motion (the elements roll between the rings), reducing friction losses by a factor of 200–500. The economic impact is hard to overstate — every axis turning faster than one revolution per second in any 20th–21st-century machine depends on it.
But this saving is not free. The contact patch — a point (ball bearing) or a line (roller bearing) — funnels all radial load through a microscopically small area where local pressure reaches p_max ≈ 1.5–4.0 GPa — 5–15× higher than the structural-steel yield strength (~300 MPa). This works only because the steel at the contact point is in a state of 3-D compression with low deviatoric stress: the hydrostatic component doesn’t drive plastic yielding, and the Hertzian contact ellipse develops a ≈10× effective contact area due to elastic deformation of the bodies themselves. This is the classical Hertz problem (Heinrich Hertz, 1881) — the foundation of all bearing theory.
A bearing that survives ≥10⁶ revolutions without visible damage under such pressure is built from AISI 52100 steel (DIN 100Cr6, ISO 683-17 / GB/T 18254) — high-carbon chromium steel with vacuum-induction-melt + vacuum-arc-remelt (VIM-VAR) processing, achieving an oxide-inclusion volume fraction ≤0.003 % per DIN 50602 K0, with hardness HRC 60–65 after quench-and-temper at 160 °C. Steel cleanliness is the key parameter: a single non-metallic inclusion of 20 μm in the Hertzian stress zone can be the nucleus of a fatigue crack that propagates to spalling in 10⁵–10⁶ cycles instead of the design 10⁹. This is why a budget bearing without vacuum remelt lives 100× less than its high-quality counterpart — and why “6900-2RS SKF” means one lifetime while “6900-2RS NoName” means another.
2. Anatomy of a rolling-element bearing — 5 components
A standard rolling-element bearing has five functional elements:
- Inner ring — annular part with a raceway groove for the rolling elements. Sits on the shaft with an interference fit per ISO 286 (k5/k6/n6 — see § 8). Rotates with the shaft in the classical configuration.
- Outer ring — the larger ring, sits in the housing with clearance fit H7/J7 or light interference N7. Stationary in the classical configuration.
- Rolling elements — ball, cylindrical roller, taper roller, spherical (barrel) roller, needle roller.
- Cage (retainer) — keeps the rolling elements at equal angular spacing, preventing them from bunching into one point (where they would create a local overload). Material: stamped steel, bronze, polyamide (PA66 + 25 % glass fiber), PEEK for high temperatures.
- Seal or shield — protects the internal volume from contamination and retains grease. Can be a non-contact metal shield (Z/ZZ) or a contact rubber lip seal (RS/2RS).
Decoding a typical scooter 6001-2RS designation:
6— single-row deep-groove ball bearing (Conrad-style).0— light-load dimensional series (thinner outer ring, larger internal volume).01— bore code: 12 mm (see § 4).2RS— two contact rubber lip seals (one on each side).
The mass of such a bearing is ≈30 g, axial width 8 mm, OD 28 mm. Four of the six wheel-hub bearings in a typical scooter are this 6001-2RS or the nearby 6201-2RS (12×32×10 mm with a thicker outer ring).
3. Bearing types by rolling-element geometry
Not all bearings are equal — each type is optimized for a particular load combination. Scooter applications use mainly five of the eight basic types:
Type 1. Deep-groove ball. The most widespread type on the planet — ≈70 % of global sales by value. Rolling elements are balls, raceways are deep (depth ≈ 30 % of bore-to-OD radius), assembled in Conrad style (cage-less assembly through eccentric ring offset was the standard until 1937). Carries full nominal radial load + moderate axial load ~25 % of C in either direction. Speed class is the highest: nDm ≈ 800 000 mm·min⁻¹ for grease lubrication. This is the workhorse of every scooter wheel hub.
Type 2. Angular contact ball. Raceway groove is offset such that the line of action passes through the balls at angle α to the radial plane: 15° (B), 25° (AC), 30° (E), 40° (A/C). A single angular-contact bearing handles axial load in one direction only — so they are used in pairs:
- DB (back-to-back) — contact-angle lines diverge outward; the virtual support center is wider than the physical pair width → very stiff against tilting moments, ideal for a headset.
- DF (face-to-face) — contact-angle lines converge inward; the virtual center is in the middle → tolerant to axle bending, used in some mid-drive motor configurations.
- DT (tandem) — both pointing the same way; doubled axial capacity in one direction → spindle applications.
The standard headset configuration in scooters is a DB back-to-back angular-contact pair at 36° or 45° (e.g., FSA Orbit, Cane Creek IS-42/52). Cane Creek IS-42 (41.8 mm diameter) and IS-52 (51.8 mm) are the two most common headset formats in NAMI, Apollo, Dualtron scooters.
Type 3. Cylindrical roller. Cylindrical rollers give line contact instead of point contact → 2–5× higher radial load at the same envelope. But zero axial capacity (without flanges) or limited capacity (with NJ/NU flange geometry). Not used directly in scooters but present in the gearbox of geared hub motors (NU 203 / NJ 204 of the typical 1:6 planetary step-down).
Type 4. Taper roller. Conical rollers and raceways → combined axial + radial capacity in one package. Classical automotive wheel hub. Rare in scooters — mainly in high-end NAMI / Apollo with adjustable preload.
Type 5. Needle roller. Needle rollers — very small diameter (1–5 mm), large quantity (20–60), long length (5–25 mm). Very compact axial width → used in transmission cams where radial space is tight. In scooters — mostly in the freewheel / one-way clutch of geared hub motors (Bafang G310, MXUS XF40).
Type 6. Thrust bearing. Rolling elements between two flat washers perpendicular to the axis → axial capacity only, zero radial. In scooters — mostly as the bottom-cup support of the folding-stem pivot.
4. Designation system — ISO 15:2017
ISO 15:2017 (“Rolling bearings — Radial bearings — Boundary dimensions, general plan”) defines the unified designation system all major manufacturers follow (SKF, NSK, FAG/Schaeffler, Timken, NTN, NACHI, Koyo, ZWZ). Key:
P Y XX [seals/shields/cage] [precision class]
Where:
-
P(first digit) — bearing type:1— self-aligning ball2— spherical roller3— taper roller4— double-row deep-groove5— thrust ball6— single-row deep-groove ball (most common)7— single-row angular contactN/NU/NJ/NF/NUP— cylindrical roller (letter, not digit)BK— needle roller bushing
-
Y(second digit) — dimensional series (OD and width):8— extra-light (thin ring, small OD) — 6800 series9— extra-extra-light (ultra-thin) — 6900 series0— light — 6000 series2— medium-light — 6200 series3— medium — 6300 series4— heavy — 6400 series
-
XX(third and fourth digits) — bore code:00→ bore = 10 mm01→ bore = 12 mm02→ bore = 15 mm03→ bore = 17 mm04→ bore = 20 mm≥04→ bore =XX × 5 mm(so 05 → 25 mm, 06 → 30 mm, 10 → 50 mm, 20 → 100 mm)
Examples for typical scooter applications (values of C dynamic load and C₀ static load — from the SKF General Catalogue):
| Designation | Bore × OD × Width, mm | Series | C, kN (dyn) | C₀, kN (stat) | Typical application |
|---|---|---|---|---|---|
608-2RS | 8 × 22 × 7 | 600 | 3.45 | 1.37 | skate wheel, kids’ scooter front fork |
6000-2RS | 10 × 26 × 8 | 6000 light | 4.55 | 1.96 | kids’ scooter front hub |
6001-2RS | 12 × 28 × 8 | 6000 light | 5.40 | 2.36 | Xiaomi M365 / Pro 2 / 4 Pro front + rear, Ninebot ES1/ES2 |
6002-2RS | 15 × 32 × 9 | 6000 light | 5.85 | 2.85 | Ninebot Max G30 / F40 / G2 |
6201-2RS | 12 × 32 × 10 | 6200 medium | 7.02 | 3.10 | Apollo City Pro, Inokim Light/Quick |
6202-2RS | 15 × 35 × 11 | 6200 medium | 7.80 | 3.75 | Dualtron Mini / Kaabo Mantis 8 |
6900-2RS | 10 × 22 × 6 | 6900 thin-section | 2.70 | 1.27 | hub-motor stator-rotor interface, display pivots |
6901-2RS | 12 × 24 × 6 | 6900 thin-section | 2.70 | 1.37 | motor shaft, low-load applications |
6902-2RS | 15 × 28 × 7 | 6900 thin-section | 4.03 | 2.32 | high-end hub motor (NAMI Burn-E) |
The bore code is easy to memorize: 00 = 10, 01 = 12, 02 = 15, 03 = 17, then × 5. Hence the engineering mnemonic: “00, 01, 02, 03 — diameters 10, 12, 15, 17; then just multiply.”
5. ISO 281:2007 — L₁₀ rating life
What does “this bearing will last N hours” mean? Bearings do not “wear” linearly — they fail by Hertzian-contact fatigue: at the ball↔raceway contact the cyclic p_max ≈ 2–4 GPa compression accumulates microscopic cracks in the subsurface layer (the Hertzian stress peak lies at a depth of ≈0.3a from the surface, where a is the contact-ellipse semi-minor axis — not at the surface itself). After some number of cycles the crack reaches the surface → a flake spalls off → spalling pit → noise, vibration, accelerated degradation.
Lundberg and Palmgren (1947 / 1952) observed that the number of cycles to spalling is statistically scattered — bearings from a single production lot of identical material can differ by 2–10× in actual revolutions to failure. So the standard talks not about “mean life” but about L₁₀ — the number of revolutions at which 10 % of the lot has failed (90 % still working). Per ISO 281:2007:
L₁₀ = (C / P)^p × 10⁶ revolutions
Where:
L₁₀— basic rating life in millions of revolutions.C— basic dynamic load rating in kN, from the manufacturer’s catalog. This is the load at which 90 % of bearings will run exactly 10⁶ revolutions. Neither the maximum nor the working load.P— equivalent dynamic load on the bearing in kN. For pure radial:P = F_r. For combined radial + axial:P = X·F_r + Y·F_awith coefficients X, Y depending onF_a/F_rratio and bearing geometry.p— bearing-type exponent:p = 3for ball bearings (point contact).p = 10/3 ≈ 3.33for roller bearings (line contact).
Converting revolutions to hours:
L₁₀_h = L₁₀ × 10⁶ / (60 × n)
Where n is rotational frequency in revolutions per minute.
Worked example — Xiaomi M365 front wheel.
- Bearing:
6001-2RS,C = 5.4 kN. - Wheel ⌀8.5“, speed 25 km/h →
n = 25 000 / (60 × π × 0.216) ≈ 614 RPM. - Rider 75 kg + scooter 12.5 kg = 87.5 kg → static load on the wheel
F = 87.5 × 9.81 / 2 ≈ 430 N = 0.43 kN. - Dynamic bump factor ≈ 2× →
P = 0.86 kN. L₁₀ = (5.4 / 0.86)^3 × 10⁶ = 6.28^3 × 10⁶ = 247 × 10⁶revolutions.L₁₀_h = 247 × 10⁶ / (60 × 614) = 6700 hours.
That’s 25 km/h × 6700 = 167 500 km of riding — effectively unlimited. In practice, the M365 front bearing dies from contamination through a worn-out seal at 2000–5000 km, not from fatigue. Key insight: a scooter bearing almost never fails from ISO 281 L₁₀ fatigue. It fails from loss of lubrication, dirt through a broken seal, or false brinelling during winter storage.
Why p = 3 for ball and p = 10/3 for roller? Lundberg-Palmgren’s empirical data on ~5000 bearings showed life falls faster than the square of load: doubling P shortens L by 2^p = 8× for balls. This follows from p_Hertz ~ F^(1/3) for point contact (Hertz 1881 → p_max ∝ F^(1/3)) → S-N fatigue curve slope b ≈ −1/9 → inversion gives p = 3. For line contact p_Hertz ~ F^(1/2) → theoretical p ≈ 4, but empirical data gave 10/3.
Ioannides-Harris modification (1985 / ISO 281:2000+). Classical Lundberg-Palmgren assumes that any stress accumulates in the subsurface layer — i.e., life falls at any P > 0. Real-world tests showed that low loads accumulate no damage below a certain threshold — fatigue limit ≈ 0.15 × C for premium bearings. Ioannides-Harris introduced:
L_nm = a₁ × a_ISO × L₁₀
Where:
a₁— reliability factor (L₁₀ → 1, L₅ → 0.64, L₁ → 0.21).a_ISO— operating-conditions factor: lubrication (κ-ratio), contamination (η_c), fatigue limit (C_u/P).
If P < C_u ≈ 0.15·C, theoretical life → ∞ per Ioannides-Harris (the bearing accumulates no fatigue damage). In the M365 front-wheel example P = 0.86 kN < C_u = 0.81 kN (at the limit), so theoretically — fatigue life ≈ ∞.
6. ISO 76:2006 — static load rating C₀ and true brinelling
ISO 281 describes dynamic life — the bearing rotates. But another criticality exists: static overload, when the bearing stands still and takes a shock through the rolling elements onto motionless raceways. If contact pressure exceeds 4 GPa, plastic deformation of the raceway under the rolling element occurs — permanent indentations called true brinelling (after the English engineer Johan August Brinell, inventor of the 1900 hardness tester).
ISO 76:2006 defines C₀ — basic static load rating — as the load at which total permanent deformation of ball + raceway at the point of maximum stress = 0.0001 × ⌀_ball. This is the threshold below which the bearing “doesn’t remember” static load.
A practical rule from the SKF General Catalogue: avoid static loads P > C₀ / 4 for premium bearings; P > C₀ / 2 guarantees true brinelling.
Worked example. A 25-kg scooter dropped from a 30 cm curb → impact load on the wheel ≈ 6× static weight = 150 kg = 1.47 kN. The 6001-2RS has C₀ = 2.36 kN. Ratio P / C₀ = 0.62 > 1/4 = 0.25 → guaranteed brinelling. On the wheel you’ll get 8 indentations (one per ball) on the outer-ring raceway, showing as “clicks” at low speed and uneven rolling resistance.
True-brinelling diagnostic: spin the wheel slowly by hand. If you hear 1–8 point “clicks” per revolution — that’s true brinelling. If a continuous rustle — that’s contamination or wear.
7. ABEC vs ISO 492 vs DIN 620 — precision classes
The manufacturing precision of a bearing (ring concentricity, raceway ovality, runout deviation during rotation) is described by precision classes. There are three parallel standards all saying the same thing:
| ABEC (USA, ANSI/ABMA) | ISO 492 | DIN 620 | JIS B1514 | Radial runout K_ir, μm (⌀≤18 mm) |
|---|---|---|---|---|
| ABEC 1 | Normal Class 6X | P0 | Class 0 | 10 |
| ABEC 3 | Class 6 | P6 | Class 6 | 7 |
| ABEC 5 | Class 5 | P5 | Class 5 | 4 |
| ABEC 7 | Class 4 | P4 | Class 4 | 2.5 |
| ABEC 9 | Class 2 | P2 | Class 2 | 1.5 |
Critical: ABEC scale is inverted relative to ISO/DIN: ABEC 1 (low number) ↔ P0 (high alphanumeric, lowest quality); ABEC 9 ↔ P2 (low ISO/DIN number, highest precision). A perpetual source of catalog confusion.
Why is ABEC 7+ almost always redundant for scooters? The ABEC scale does not specify:
- Load capacity (
C). - Steel cleanliness and material quality.
- Hardness (HRC 60–65 — larger impact on life than runout).
- Seal and grease quality.
- Noise and vibration levels.
For a high-RPM spindle (CNC ⌀ 50 000 RPM, dental drill 300 000 RPM) runout is critical — at 50 000 RPM even 4 μm produces centrifugal load that destroys the bearing. But a scooter wheel spins at 600–2500 RPM (mid-drive motor at 1500–3000 RPM pre-gearbox). At these speeds the actual-life difference between ABEC 1 and ABEC 7 is zero. Marketing paradox: skate and scooter forums often push “ABEC 9 for speed” — that’s actively harmful, because ABEC 9 costs 5–10× more, its seals are made thinner (reduces friction at high RPM but speeds up contamination), and the actual material quality is often lower than a good ABEC 3 SKF/NSK.
Bottom line: for all typical scooter applications, ABEC 3 (P6) from a quality manufacturer (SKF, NSK, NTN, NACHI, FAG/Schaeffler) is enough. ABEC 7 is marketing.
8. ISO 286 fits — shaft and housing tolerances
A bearing is not “bolted” to the shaft — it is pressed on with an interference fit on the rotating shaft and with a clearance fit in the stationary housing. This is fundamental: if both rings were interference-fit, then heating of the shaft during operation (thermal expansion α_Fe ≈ 12 × 10⁻⁶ /°C) would crush the bearing radially and destroy it in hours.
General rule (SKF Engineering Reference):
- Ring that rotates with the shaft — interference. Shaft tolerance
k5,k6,m5,n6(i.e.,+5…+30 μmabove nominal of the matching H7 hole) → the inner ring cannot spin on the shaft under load reversal. - Ring stationary in a stationary housing — clearance. Housing bore
H7,J7,K7(i.e.,+10…+25 μmabove nominal) → the outer ring can slightly turn in the housing under thermal expansion, without being radially crushed.
If the situation is reversed (rotating outer ring, stationary shaft — rare in scooters but possible in geared hub motors), the rule inverts: interference in housing, clearance on shaft.
Worked example for a typical scooter front wheel:
- Shaft ⌀ 12 mm,
12k6→ tolerance+1…+12 μm→ actual ⌀ 12.001–12.012 mm. - Hub bore ⌀ 28 H7 → tolerance
+0…+21 μm→ actual 28.000–28.021 mm. - Bearing
6001-2RSbore tolerance ABEC 1:−10…0 μm→ actual 11.990–12.000 mm. - Bearing OD tolerance ABEC 1:
−13…0 μm→ actual 27.987–28.000 mm.
Net interference inner ring → shaft: 12.001–12.012 (shaft) vs 11.990–12.000 (bore) → interference +1…+22 μm → light interference, but guaranteed. In the OD direction: 28.000–28.021 (housing) vs 27.987–28.000 (OD) → clearance 0…+34 μm → outer ring sits freely and can rotate slightly with thermal expansion.
Why does this matter for DIY maintenance? If you remove an old bearing from the shaft and it comes off by hand — the shaft has worn (12k6 interference has degraded to 12h6 clearance). A new bearing will then spin on the shaft, accumulate wear and fretting, and live 5–10× less. Solution: either replace the shaft (in a hub-motor it means full motor replacement) or apply Loctite 638 retainer compound (anaerobic adhesive that fills gaps up to 0.15 mm and cures to 7000 psi shear strength).
9. Seals — Z, ZZ, RS, 2RS — and IP correlation
Sealing-suffix marks contamination protection:
| Marking | Type | Contact | Friction | IP rating | Speed |
|---|---|---|---|---|---|
| no suffix | Open | — | Lowest | None | Highest (requires an external seal or labyrinth in housing) |
Z | Metal shield (1 side) | Non-contact (~0.1 mm gap) | Low | IP3X-IP4X | High |
ZZ (2Z) | Metal shield (2 sides) | Non-contact | Low | IP4X-IP5X | High |
RS | Contact rubber lip (1 side) | Lip touches inner ring | Moderate | IP54-IP65 | Moderate (~15–25 % derate) |
2RS (DDU / LLU) | Contact rubber lip (2 sides) | Lip touches inner ring | Moderate | IP54-IP65 | Moderate |
RSL/LLB | Low-friction contact | Light lip-ledge on inner | Low-moderate | IP4X-IP54 | High |
Lip-material chemistry:
- NBR (Nitrile Butadiene Rubber, ARP 568 standard) — default. Operating range −30…+110 °C, good resistance to mineral oils, poor to ozone and UV (surface cracking in 2–3 years).
- HNBR (Hydrogenated NBR) — −40…+150 °C, better ozone/UV resistance, costlier.
- FKM (Viton/Fluorelastomer) — −20…+200 °C, only for high-temperature applications (high-power hub motors).
Why is 2RS the scooter standard? A 2RS lip blocks >99 % of road grit, dust, and rainwater at the IPX4 level (the bearing in the hub is usually further protected by a housing labyrinth — see IP protection). The 15–25 % speed penalty is irrelevant at 600–2500 RPM. 2RS is the universal default for all scooter wheel-hub and headset bearings.
10. Lubrication — NLGI, base oil, thickener, EP additives
If ISO 281 describes fatigue as the critical failure mode, then in reality 80 % of scooter bearings die from lubrication loss/degradation (not from steel fatigue). Lubrication is its own science.
10.1 NLGI grades — consistency
The National Lubricating Grease Institute (USA) classifies grease by the worked-penetration test per ASTM D217: a standard cone (A_top = 21.4 mm², ρ_total = 102.5 g) falls into the grease for 5 seconds, with penetration measured in tenths of a millimeter after 60 working strokes of the plunger at 25 °C. Classification:
| NLGI | Worked penetration, 0.1 mm | Consistency | Analogy | Typical use |
|---|---|---|---|---|
| 000 | 445–475 | Fluid | Cooking oil | Open gear, automatic central lubrication |
| 00 | 400–430 | Semi-fluid | Apple sauce | Gear oil-grease, low-temp |
| 0 | 355–385 | Very soft | Brown mustard | Subzero applications, central lub |
| 1 | 310–340 | Soft | Tomato paste | Bearings, low-temp |
| 2 | 265–295 | “Normal” | Peanut butter | Universal default — ≥90 % of ball bearings |
| 3 | 220–250 | Firm | Vegetable shortening | High-temp, high-vibration |
| 4 | 175–205 | Very firm | Frozen yogurt | Special applications |
| 5 | 130–160 | Hard | Smooth pâté | — |
| 6 | 85–115 | Very hard | Cheddar cheese | — |
For scooter wheel hubs and headsets — NLGI 2 lithium-complex with ISO VG 100–220 base oil — covers 95 % of factory pre-greased bearings.
10.2 Thickener
Grease is base oil + thickener + additives in roughly 80 % + 10–15 % + 5–10 % proportion. The thickener is a sponge-like matrix that holds base oil capillarily and releases it under pressure (squeeze film). After release it reabsorbs. This gives grease its “sleep-when-not-used” property — in the absence of motion it doesn’t flow out of the bearing, unlike oil.
Four primary thickener systems:
| Thickener | Operating range | Dropping point | Water resistance | Cross-compatibility | Typical use |
|---|---|---|---|---|---|
| Lithium 12-hydroxystearate (plain Li-soap) | −30…+120 °C | 190–210 °C | Moderate | Compatible with most metallic soaps | Generic Li-grease (95 % of budget bearings) |
| Lithium complex | −40…+150 °C | 260–280 °C | Good | Similar to Li-soap | Premium default, SKF LGMT 2 |
| Polyurea | −40…+170 °C | 260–280 °C (deg.) | Very good | NOT compatible with Li or Ca soap | High-temp electric-motor bearings |
| Calcium sulfonate complex | −40…+180 °C | >300 °C | Excellent | Compatible with Li-complex | Marine/wet environment |
Grease incompatibility — a classic DIY mistake. If you add polyurea to a Li-greased hub (e.g., topping up from another tube), the two thickeners mutually break down and the grease “collapses” into liquid oil over weeks. Result — a bearing without lubricant in a month. Rule: when changing grease, fully wash out the old grease with solvent (white spirit, IPA, mineral spirit), then apply new.
10.3 Base oil — ISO VG ranges
Base oil — ≈80 % of the grease — has a viscosity that determines the lubrication regime at operating temperature:
- ISO VG 32–46 (mineral): low-temp, high-speed spindles.
- ISO VG 100–150 (mineral/PAO): ball-bearing standard,
nDm ≤ 500 000. - ISO VG 220–460 (mineral/PAO/ester): high-load, low-speed roller bearings.
- ISO VG 680–1000 (synthetic): worm gears, gearboxes.
For scooter wheel hubs the typical range is ISO VG 100–220 (higher viscosity at low operating speed and high contact pressure). For high-speed hub motors — ISO VG 68–100 (lower viscosity reduces churning losses).
PAO (polyalphaolefin) synthetic vs mineral — PAO has a 2–3× wider operating range and better oxidative stability (lives ≈2× longer at high temperatures). Premium scooters (NAMI, Dualtron Storm) ship with PAO-based lubricants.
10.4 EP/AW additives — ZDDP, MoS₂
Under boundary lubrication (see § 10.5) metal-on-metal contact causes microscopic adhesive welding. EP (extreme pressure) and AW (anti-wear) additives prevent this by forming a thin protective film on the surface.
ZDDP (zinc dialkyldithiophosphate) — introduced in the 1940s as an AW/EP additive. Mechanism: under boundary contact the Zn-O-P-S-O molecule thermo-catalytically decomposes and forms a Zn-phosphate-glass tribofilm of 50–150 nm thickness on the steel surface (Watson et al. 1945; classic review by Spikes 2004). This tribofilm:
- Has lower hardness
≈2–3 GPathan the steel itself — absorbs shock loads plastically. - Has directional roughness — oriented along sliding (orientation effect).
- Wears out — over time ZDDP concentration drops, the tribofilm thins → wear accelerates.
MoS₂ (molybdenum disulfide) — twin-layer solid lubricant. MoS₂ sheets slide over each other with μ ≈ 0.03–0.06. Added to grease at 2–5 % wt. Works in vacuum (unlike graphite which needs moisture). Cannot replace ZDDP — it’s a solid lubricant, not an AW additive.
Sulfur-phosphorus (S-P) packages — generic EP additive for gearboxes, not for bearings.
10.5 Stribeck curve — λ-ratio and the three regimes
Lubrication has three regimes depending on the ratio of oil-film thickness h₀ to composite surface roughness R_q = √(R_q1² + R_q2²):
λ = h₀ / R_q
| Regime | λ | What happens | Friction μ |
|---|---|---|---|
| Boundary | λ < 1 | Plain metal-on-metal through R_q peaks; EP-additive tribofilm critical | 0.08–0.15 |
| Mixed | 1 ≤ λ ≤ 3 | Partial oil film + partial metal contact at asperities | 0.02–0.08 |
| Full-film EHL (elasto-hydrodynamic) | λ > 3 | Full oil film, metals don’t touch | 0.001–0.005 |
h₀ is calculated from the Hamrock-Dowson formula (1981) for EHL contact:
h₀ / R_x = 2.69 × (η₀·u / E'·R_x)^0.67 × (α·E')^0.53 × (W/E'·R_x²)^(−0.067) × (1 − 0.61·e^(−0.73k))
Where η₀ — base-oil viscosity, u — entrainment velocity, E' — composite Young’s modulus, R_x — composite radius in the rolling direction, α — pressure-viscosity coefficient, W — contact load.
A scooter bearing at normal speed (n = 600–2500 RPM) operates in full-film EHL with λ ≈ 3–8 — essentially zero friction, zero wear. But:
- Startup (n → 0) →
λ → 0(starts in boundary) → wear accumulates in the first seconds after start. - Low speed + high load (climbing a hill with a 75-kg rider) → λ ≈ 1.5 → mixed regime → significant wear.
- Grease loss (seal lip cracked at 3 years, grease evaporated) → λ → 0 → boundary → catastrophic wear in weeks.
11. Failure modes — how bearings die
The classification ISO 15243:2017 (Rolling bearings — Damage and failures — Terms, characteristics and causes) identifies 6 primary mechanisms:
1. Subsurface-initiated fatigue (spalling). The classical ISO 281 mechanism. The Hertzian stress peak at depth ≈0.3a (where a is the contact-ellipse semi-axis) accumulates micro-cracks over 10⁹–10¹⁰ cycles. The crack reaches the surface → a fragment spalls off → spalling pit. Sound: quiet rumble, growing to “grumbling” at full RPM. Location: on the outer-ring raceway, evenly around the circumference. Forecast: bearing survives 100–500 hours after spalling begins.
2. Surface-initiated fatigue (peeling, micropitting). Surface micro-cracks from λ < 1 boundary regime. Tiny pits <20 μm. Propagates faster than spalling. Cause: contamination, bad grease, overload.
3. True brinelling. Static overload P > C₀/4 → plastic deformation of the raceway under the rolling elements → 8–10 indentations (one per ball) on the raceway. Sound: “click-click-click” per revolution at low speed. Cause: scooter drop, overload at parking (riding off a curb with 100 kg payload), impact from height.
4. False brinelling / fretting corrosion. The sneakiest mechanism for scooters. Occurs when the bearing stands still under static load and is subjected to external vibration. Micro-oscillations (< 1° rotation) squeeze grease out of the contact zone and prevent it from returning (no circumferential motion). Boundary contact → adhesive wear + oxidation → Fe₂O₃ (hematite) third-body abrasive → indentations that look like brinelling but are caused by vibration, not static force.
Classic false-brinelling scenarios in scooters:
- Winter storage in a garage next to a washing machine / gas boiler (50–100 Hz vibration).
- Truck transport (10–25 Hz road vibration, weeks at a time) without rotating the wheels.
- Permanent curbside parking near a busy road (10–60 Hz traffic-induced vibration).
Prevention: during long storage, every month rotate wheels by hand through a full turn (restores lubrication film) or lift the scooter on a stand (remove static load).
5. Fluting (electrical erosion). When metal-on-metal contact is unstable through a thin oil film, a potential difference >1 V between rings drives electrical arcs — micro-arc discharges that vaporize metal at the breakdown point. Characteristic fluted patterns (washboard-like grooves) appear on the raceway.
Fluting in scooters is rare (low voltages 36–60 V), but possible in two scenarios:
- Hub motor with damaged winding insulation → leakage current through the bearing → fluting on the motor bearing.
- ESD (electrostatic discharge) from the rider’s body after walking on dry/cold/synthetic surfaces.
Prevention: ensure the motor frame is electrically bonded to the deck/frame (grounding bond).
6. Wear from contamination. The most common mode in scooters — 60–70 % of all failures. A cracked 2RS seal admits road dirt, sand, water. Hard particles (silica, ⌀10–100 μm) act as third-body abrasive between ball and raceway → linear wear, gradual runout growth, noise, vibration. Looks like clouding and scratches on the raceways.
Prevention: annual seal inspection; with damage — replace the bearing, not “flush and re-grease” (an NBR lip seal cannot be regenerated).
12. Bearing-diagnostic symptom matrix
| Symptom | Likely failure mode | Root cause | Action |
|---|---|---|---|
| Quiet hum at constant speed | Early spalling | ISO 281 fatigue or surface contamination | Audio-monitor the next 50 km; if it grows — replace |
| Grumbling at full RPM | Mature spalling | Subsurface fatigue completing | Immediate replacement (<200 km to catastrophic failure) |
| “Click-click” per revolution at low speed | True brinelling | Drop from height / overload | Replace |
| Rustle + gradual rolling-resistance growth | Contamination wear | Cracked 2RS seal | Replace |
| Fluted feel rotating by hand (wavy points of resistance) | Fluting (electrical erosion) | Hub-motor leakage / ESD | Check motor grounding, replace bearing |
Wheel side-play (axial play >0.5 mm) | Inner/outer ring + housing wear | Interference fit lost | Inspect housing OD wear; apply Loctite 638 if marginal; otherwise replace shaft |
| Rusted ball surface on disassembly | Contamination + lubrication loss | Old grease + cracked seal | Replace (cannot be regenerated) |
| Whistle/whine during hub-motor acceleration | Dry bearing or resonance | Old grease + spent EP additive | Replace 6900-2RS standard |
13. Bearings in scooter subsystems
Front wheel hub. Two deep-groove ball bearings — 6001-2RS (M365) or 6002-2RS (Ninebot Max). Interference inner ring on a ⌀12 mm or ⌀15 mm shaft. Clearance outer ring in the hub. Replacement tool: 3-claw internal bearing puller + arbor press for the new one. DIY time ≈45 min for the pair.
Rear hub / hub motor. Internal architecture is typically 6001-2RS motor side + 6201-2RS axle side (boost models step up to 6002 / 6202). In a brushless hub motor the stator-rotor interface has the outer ring rotating around a fixed shaft → rotating outer ring, inverting ISO 286 fits (interference in housing, clearance on shaft). Replacement is harder, requiring hub-motor disassembly (8–12 side-cover bolts, careful housing split). DIY time ≈3 hours for the pair. Available in high-end mod shops for OEM hub motors (Bafang G310, MXUS XF40).
Headset (steering column). Standard formats:
- Threadless 1-1/8“ (28.6 mm steerer) — the most common in scooters. Headset cups IS-42 (semi-integrated, 41.8 mm diameter) or IS-52 (51.8 mm). Conical angular-contact bearings 36°/45° in a back-to-back DB pair.
- Threaded 1“ (25.4 mm steerer) — older standard; occasionally on budget or kids’ models.
Replacement tools: slide hammer for extraction of old cups (press-fit into the frame head tube), press tool for the new ones. Bearings are cartridge-type 1-1/8“ × 36/45° (FSA Orbit, Cane Creek, Neco). DIY time ≈90 min.
Freewheel / one-way clutch (geared hub motors). Bafang G310, MXUS XF40, Dapu DMHC09 use a 1:6 planetary gearbox + one-way clutch (sprag clutch). The sprag is an eccentric cam that locks when rotating in one direction and freewheels in reverse → allows coasting (no motor-drag when off-throttle). The sprag bearing is a needle roller in the inner race. Failure mode: sprag wear → wheel “slips forward” without drive (clicking sound) → gearbox-assembly replacement.
Throttle / brake-lever pivots, kickstand. Simple plain bushings (no rolling elements), mostly nylon-polymer with NLGI 1 grease. Don’t need attention < 10 000 km.
14. Recap — 8 key takeaways
- L₁₀ formula:
L₁₀ = (C/P)^p × 10⁶ revolutions(ISO 281:2007). p = 3 for ball, 10/3 for roller. Conversion:L₁₀_h = L₁₀ × 10⁶ / (60n). - C and C₀ are two different ratings: dynamic (ISO 281) describes fatigue under rotation, static (ISO 76) describes brinelling at standstill. Avoid
P > C₀ / 4. - 6xxx-series — the ISO 15:2017 system: first digit — bearing type (6 = deep-groove ball); second — dimensional series (0/2 = light, 8/9 = thin-section); last two — bore code (00=10, 01=12, 02=15, 03=17, ≥04 = ×5).
- ABEC ≡ ISO 492 (inverted): ABEC 1 ↔ P0; ABEC 9 ↔ P2. For scooters, ABEC 3 (P6) from a quality OEM is enough; ABEC 7+ is marketing.
- ISO 286 fits: rotating inner → shaft
k5/k6/n6interference; stationary outer → housingH7/J7/K7clearance. Reversed for a rotating outer ring (hub motors). - 2RS rubber lip — universal default for scooter bearings. NBR is the standard material; HNBR/FKM for high temperatures.
- NLGI 2 lithium-complex grease — 95 % of factory pre-greased bearings. Thickener incompatibility (Li-soap ⊕ polyurea) is a classic DIY mistake.
- Failure modes: 60–70 % of scooter bearings die from contamination through a broken seal, not from ISO 281 fatigue. Second cause — false brinelling from winter storage vibration. ZDDP tribofilm is critical for startup boundary regime.
Conclusion
A bearing is the cheapest mechanical part on a scooter ($5–20 per pair) and the most critical for ride quality. It integrates all rotational loads of the system through a microscopic contact point where p_max ≈ 2–4 GPa is sustained only because the steel is locally under near-hydrostatic compression. No part of the ISO 281 formula is determined by a marketing label “ABEC 9” — actual life is determined by steel quality (AISI 52100 VIM-VAR), fit geometry and tightness (ISO 286 k6 / H7), seal (2RS NBR rubber), grease (NLGI 2 lithium-complex, ISO VG 100–220 base oil, ZDDP), and operating regime (full-film EHL with λ > 3).
The owner cannot change the C rating of their bearing, but can: (1) avoid drops > 30 cm onto the wheel side (don’t exceed C₀/4); (2) winter-store the scooter with wheels lifted on a stand, or monthly rotate wheels by hand (false-brinelling prevention); (3) annually inspect seals and replace the bearing at first signs of contamination wear (don’t “flush” — 2RS lip seal doesn't regenerate); (4) on replacement, use only 2RS from OEM-tier manufacturers (SKF, NSK, NTN, NACHI) and completely flush old grease with solvent before applying a new compatible grease. This approach yields real-life 5000–10 000 km between replacements instead of the factory 1500–3000 km of budget bearings without vacuum remelt.
Further reading
- Frame and fork engineering — angular-contact bearings in the headset, with ISO 286 fits context.
- Motor and controller engineering — hub-motor bearing architecture and whine diagnosis.
- Suspension engineering — bearings in pivot points (Hiley Tiger, NAMI).
- IP-protection engineering — why a rotating shaft seal caps hub motors at IPX5.
- Maintenance and storage — practice of bearing inspection.
- Pre-ride safety check — a quick audio test for bearing wear.
- Used-scooter pre-purchase inspection — how to find hidden bearing failures.
Sources
- ISO 281:2007 — Rolling bearings — Dynamic load ratings and rating life. International Organization for Standardization, 2007 (canonical L₁₀ formula).
- ISO 76:2006 — Rolling bearings — Static load ratings. International Organization for Standardization, 2006 (canonical C₀ rating).
- ISO 492:2014 — Rolling bearings — Radial bearings — Geometrical product specifications (GPS) and tolerance values. International Organization for Standardization, 2014.
- ISO 15:2017 — Rolling bearings — Radial bearings — Boundary dimensions, general plan. International Organization for Standardization, 2017.
- ISO 286-1:2010 — Geometrical product specifications (GPS) — ISO code system for tolerances on linear sizes — Part 1: Basis of tolerances, deviations and fits.
- ISO 15243:2017 — Rolling bearings — Damage and failures — Terms, characteristics and causes.
- ASTM D217-21a — Standard Test Methods for Cone Penetration of Lubricating Grease. ASTM International.
- ASTM D2266-91(2015) — Standard Test Method for Wear Preventive Characteristics of Lubricating Grease (Four-Ball Method). ASTM International.
- ABMA Standard 20 — Radial Bearings of Ball, Cylindrical Roller and Spherical Roller Types (ABEC-equivalent table).
- Lundberg, G., & Palmgren, A. (1947). Dynamic Capacity of Rolling Bearings. Acta Polytechnica, Mechanical Engineering Series, Vol. 1, No. 3 (canonical fatigue-life theory).
- Ioannides, E., & Harris, T. A. (1985). A New Fatigue Life Model for Rolling Bearings. ASME Journal of Tribology, 107(3), 367–377 (modern fatigue-limit modification).
- Hertz, H. (1881). Über die Berührung fester elastischer Körper (On the contact of elastic bodies). J. reine angew. Math., 92, 156–171.
- Hamrock, B. J., & Dowson, D. (1981). Ball Bearing Lubrication: The Elastohydrodynamics of Elliptical Contacts. Wiley-Interscience (canonical EHL formula).
- Spikes, H. (2004). The History and Mechanisms of ZDDP. Tribology Letters, 17(3), 469–489.
- SKF Group (2024). Rolling Bearings — General Catalogue. PUB BU/P1 17000/1 EN.
- NSK Corporation (2023). Technical Report — Bearing Doctor Diagnostic Guide (CAT. No. E728g).
- NLGI (National Lubricating Grease Institute). Grease Glossary and Grade Classification System (canonical NLGI consistency-number reference).
- Wikipedia: NLGI consistency number (ASTM D217 worked-penetration table summary).
- Wikipedia: ABEC scale (ABEC↔ISO 492↔DIN 620↔JIS B1514 cross-reference).
- Wikipedia: Rolling-element bearing (Lundberg-Palmgren formula and bearing types).
- Wikipedia: False brinelling and Fretting corrosion (storage/vibration damage mechanisms).
- ScienceDirect: Experimental study on ZDDP tribofilm formation in grease lubricated rolling/sliding contacts. Tribology International, 2025.
- ONYX Insight: Fretting Corrosion Bearing Failures — Failure Atlas (wind-turbine application reference, transferable mechanism).
- Watson, R. W., McTurk, T. M., & Roselin, M. (1945). The Use of Zinc Dialkyldithiophosphates as Anti-Oxidants and Anti-Wear Additives in Lubricating Oils. SAE Technical Paper (canonical ZDDP introduction).