Planetary Gearbox Service Life — How 20,000 hr and 30,000 hr Are Defined, Calculated, and Verified

Korea Ever-Power test centre, Ansan-si. Endurance test rigs run EP-series units at rated torque and rated speed under controlled temperature to validate the 20,000 hr and 30,000 hr L10 design life. Test data generates the bearing dynamic load rating inputs used in Korea Ever-Power’s life calculation model.

Throughout Korea Ever-Power’s application guides, product pages, and technical specifications, two service life figures appear consistently: 20,000 hours and 30,000 hours. A third figure appears with an asterisk — 30,000 hr* — on the FAL and FALR series. These numbers answer a specific engineering question: under what conditions, and for how long, can the gearbox be expected to reach the end of its design life without bearing fatigue failure or gear surface degradation below specification?

The answer depends on three parameters that the specification number alone does not state: the duty cycle (IEC 60034-1 S-class), the applied torque as a fraction of rated torque, and the operating speed. The 20,000 hr and 30,000 hr figures are calculated and validated at rated output torque, at rated input speed, under the duty cycle indicated in the specification. Operating at a fraction of rated torque, at lower speed, or under an intermittent rather than continuous duty cycle extends the achievable life — in some cases dramatically. Operating above rated torque or above rated speed reduces it equally dramatically.

Understanding the duty cycle definition and the L10 bearing life calculation behind these numbers gives the machine design engineer two practical tools: the ability to confirm that the specified EP-series and frame size is correctly sized for their specific application duty profile, and the ability to calculate the expected service interval for their maintenance schedule. Neither calculation requires specialist bearing knowledge — the procedure is a straightforward application of the standard ISO 281 bearing life formula with Korea Ever-Power’s published dynamic load ratings.

Load Profile Over Time

S1 — Continuous Running

Rated torque → 100% of time

S5 — Intermittent Periodic

■ Rated load
■ Braking
■ Rest (motor off)

Fatigue consequence: S5 accumulates bearing fatigue cycles only during the loaded phase. Rest periods do not add fatigue, so a 30%-on S5 duty accumulates fatigue at 30% the rate of S1 — L10 life in calendar hours is 3.3× longer.

The IEC 60034-1 Duty Cycle Classes Relevant to EP-Series

IEC 60034-1 defines ten duty cycle classes (S1 through S10) that standardise the description of motor and drive loading conditions for equipment rating and life calculation. Of these, two are relevant to EP-series planetary gearbox specifications: S1 (continuous running) and S5 (intermittent periodic duty with electric braking).

S1 — Continuous Running Duty: The motor and gearbox operate at constant rated load indefinitely, with no rest periods. Thermal equilibrium is reached and maintained at the rated operating temperature. S1 is the most demanding condition for both motor and gearbox thermal rating. For bearing life calculation, S1 means the bearing accumulates fatigue cycles continuously at the maximum rate. The EP-FAD 30,000 hr and EP-FAB 20,000 hr ratings are calculated at S1 — they represent the life at the maximum possible fatigue accumulation rate.

S5 — Intermittent Periodic Duty with Electric Braking: The motor and gearbox operate in a repeating cycle: a period at rated load followed by a braking phase, then a rest period with the motor de-energised. The key characteristic of S5 (vs S3 intermittent without braking) is that the load cycle includes a regenerative or resistive braking event before the rest period — which is the normal operating profile of a servo-driven axis that accelerates, reaches a target, decelerates under servo control, stops, and waits for the next command. S5 is the correct duty class for the overwhelming majority of servo automation axes: robot joints, CNC linear axes, packaging machine drives, and carriage belt axes are all S5 in operation, not S1.

Why S5 produces longer gearbox life than S1: Bearing rolling element fatigue obeys the same physics as any metal fatigue mechanism — it accumulates with cycles, not with calendar time. A bearing that operates for 8 hours at rated load followed by 16 hours off accumulates the same fatigue in each operating hour as a bearing running continuously, but only accumulates it for 8 hours per day. After 3,750 days (approximately 10.3 years), the 8-hour-per-day bearing has accumulated the same total fatigue cycles as a continuously running bearing would accumulate in 30,000 hours. This is the mathematical basis for the S5 life extension: the same 30,000 hours of fatigue equivalent running time takes proportionally longer in calendar time when the actual running time per day is a fraction of 24 hours.

A concrete illustration: an EP-FAD unit on a robot J3 axis in an automotive welding line operates in a move-and-weld cycle: 0.4 seconds of motion, 2.1 seconds stationary for welding, 0.4 seconds return, 2.1 seconds stationary for fixture release. Total cycle: 5.0 seconds. On-time fraction: 0.8 seconds motion ÷ 5.0 seconds cycle = 16% on-time, 84% rest. Over a three-shift production day (21 hours running): the gearbox accumulates 21 × 0.16 = 3.36 hours of fatigue-equivalent running per day. In calendar hours, the gearbox reaches its 30,000 hr L10 fatigue equivalent after 30,000 ÷ 3.36 × 24 = 214,286 calendar hours — approximately 24.5 years of production. Even with conservative maintenance planning at 60% of L10 life, the planned replacement interval is 12,860 calendar hours — approximately 14.7 years. This is why robot joint gearboxes are rarely the life-limiting component in an automotive welding cell: the weld tips, cable harnesses, and encoder bearings require service far sooner than the planetary gearbox bearing reaches its fatigue limit.

The contrast with true S1 applications is instructive. An offset press plate cylinder drive at 8,000 sheets/hour runs continuously for 16 hours per shift with the gearbox under rated torque at all times — an on-time fraction of essentially 100% for the running period. Over a 250-day production year at 16 hours/day, the gearbox accumulates 4,000 hours of rated fatigue per year. At this rate, the EP-FAD 30,000 hr S1 rating represents 7.5 years of production before the L10 is reached. With maintenance planning at 70% of L10, the replacement interval is approximately 5.25 years — which is why printing press OEMs typically specify a 5-year gearbox replacement interval on high-speed commercial offset presses.

ISO 281 L10 Bearing Life Formula

L₁₀ = (C/P)³ × 10⁶ ÷ (60 × n)

L₁₀ = hours (90% survival probability)
C = dynamic load rating (N) — from bearing catalogue
P = equivalent dynamic load (N) — from application
3 = exponent for ball bearings (10/3 for roller bearings)
n = rotational speed of bearing (rpm)

Example: C=25,000N, P=12,500N (50% rated), n=3,000 rpm
L₁₀ = (25,000/12,500)³ × 10⁶ / (60×3,000) = 8 × 10⁶ / 180,000 = 44.4 hr… wait, multiply by factor: = (2)³ × 10⁶ / 180,000 = 44.4… (note: this is millions of revolutions divided by rpm×60, gives hours correctly)

The ISO 281 L10 bearing life formula is the universal standard method for calculating the expected service life of rolling element bearings — and therefore of gearboxes whose life is bearing-limited. L10 life is defined as the number of operating hours at which 10% of a large population of nominally identical bearings operating under the same conditions would be expected to have experienced a fatigue spalling failure. Equivalently: 90% of bearings operating under these conditions are expected to survive beyond the L10 life.

The most important insight from the L10 formula is the cubic relationship between load ratio (C/P) and life. Doubling the load (halving C/P) reduces the L10 life by a factor of 8. Halving the load (doubling C/P) increases L10 life by a factor of 8. This cubic relationship means that even a modest operating torque reduction — say, operating at 80% of rated torque rather than 100% — extends L10 by a factor of (1/0.8)³ = 1.95, nearly doubling the expected life. This is why Korea Ever-Power’s application engineering team asks for the typical operating torque in addition to the rated torque when reviewing life requirements: an EP-FAD unit rated at 100 N·m and operated at 60 N·m typical may have an effective L10 life three times the rated 30,000 hr figure.

What the equivalent dynamic load P includes: For EP-series output bearings, P is calculated from the vector sum of the radial load (from the reaction force of the gear mesh and any external overhung load on the output shaft) and the axial load (from helical gear thrust for EP-FAD/FAB, or zero for spur-gear configurations). For FAL/FALR, P at the input bearing additionally includes the radial component from the belt tension — which is the reason the FAL/FALR input bearing is more heavily loaded than the output bearing and is the life-limiting component in continuous belt operation.

What limits EP-series life in practice: For EP-FAD and EP-FAB at rated load and speed, the output bearing L10 life sets the design life figure. The gear train (tooth flank surface fatigue, root bending fatigue) has a calculated life significantly longer than the bearing life at rated conditions, because DIN Class 5 profile-ground gears produce lower tooth contact stress than the gear geometry alone would suggest — the precise tooth form distributes the Hertzian contact stress more evenly across the tooth face width, reducing peak stress and extending fatigue life. The practical implication is that at rated load, the gearbox will reach bearing fatigue before gear surface fatigue — and at below-rated load, both lifetimes extend substantially.

This worked example shows the general procedure. Korea Ever-Power’s application engineering team performs this calculation as part of the gearbox selection review for any application where service life is a specification requirement — send your typical operating torque (not just rated), typical speed, and duty cycle to [email protected] for a confirmed life calculation.

Life vs Operating Torque

100% rated torque
1.0× rated

80% rated torque
1.95× rated

63% rated torque
4× rated

50% rated torque
8× rated

40% rated torque
15.6× rated

L10 scales as (100/T%)³. Cubic relationship — modest torque reduction gives large life multiplier.

Four Conditions That Shorten Life Below Rating

1. 01
   Operating above rated torque

   Exceeding rated output torque increases the bearing equivalent dynamic load above the value used in the L10 calculation. At 120% rated torque, L10 is reduced by (120/100)³ = 1.73× — the gearbox reaches its bearing fatigue life 42% sooner than the rated figure. The rated torque on Korea Ever-Power’s data sheets is the maximum continuous torque at which the 30,000 hr or 20,000 hr life is achievable. The peak torque (typically 2–3× rated, listed separately) is the instantaneous maximum that the gearbox can transmit without tooth breakage or shaft yielding — it does not produce rated life if applied continuously.

2. 02
   Housing temperature above 40°C ambient

   The rated life assumes an ambient temperature of 40°C maximum at the gearbox housing. Above 40°C, the NYOGEL 792D grease viscosity is lower, the elastomer seals age faster, and the bearing steel’s fatigue properties may be slightly reduced. For installations near heat sources (stenter frames, drying ovens, furnace zones), confirm the expected housing temperature and contact Korea Ever-Power for a life adjustment. NYOGEL 792D retains adequate lubricating film up to +125°C housing temperature, but the bearing L10 adjustment factor applies above 40°C ambient — the life reduction is a bearing specification effect, not a lubricant effect.

3. 03
   Excessive overhung load on the output shaft

   An overhung load — a radial force applied to the output shaft at a point along its length rather than at the shaft end flange — adds to the bearing equivalent dynamic load in addition to the gear mesh reaction force. Korea Ever-Power’s data sheets specify the maximum permissible overhung load at a given distance from the housing face. Exceeding this value increases P in the L10 formula, reducing bearing life. Common sources of overhung load: sprockets, spur gears, or pulleys mounted directly on the output shaft rather than supported by a separate bearing in the machine frame.

4. 04
   High-frequency shock loading or vibration at input

   Repetitive shock loads at the gearbox input — from emergency stops, abrupt acceleration profiles, or mechanical resonance — produce instantaneous bearing loads significantly above the steady-state equivalent dynamic load. If these shock events occur frequently, the cumulative fatigue damage may be substantially higher than the L10 formula predicts for the steady-state load. For applications with frequent emergency stops or aggressive acceleration/deceleration profiles, Korea Ever-Power recommends specifying the peak torque and the frequency of peak events when requesting a life calculation review.

Five Conditions That Extend Life Above Rating

1. +1
   Operating below rated torque — cubic relationship means even modest torque reduction multiplies life significantly. At 80% rated torque: 1.95× life; at 50% rated: 8× life. Most servo robot joints spend the majority of their cycle below 60% rated torque.
2. +2
   S5 intermittent duty — rest periods between moves reduce fatigue accumulation rate in proportion to the on-time fraction. A 30% on-time S5 duty theoretically gives 3.3× the calendar life of S1 at the same hourly loading. Most servo automation is S5, not S1.
3. +3
   Lower operating speed — at lower rpm, the bearing accumulates fewer fatigue cycles per hour. An EP-FAD running at 1,000 rpm input accumulates fatigue at one-third the rate of the same unit at 3,000 rpm. The L10 formula’s 1/(60×n) factor captures this directly.
4. +4
   Ambient temperature below 40°C — lower temperatures maintain grease viscosity above its optimal operating range for longer, improve bearing steel fatigue properties slightly, and extend elastomer seal life. Cold-climate applications may benefit from longer seal and lubricant life, partially offsetting the cold-start viscosity issue addressed in the lubricant guide.
5. +5
   DIN Class 5 gears reducing bearing load — DIN Class 5 profile-ground gears produce lower gear mesh vibration force (lower dynamic load increment) than DIN Class 6–7 gears at the same rated torque. The dynamic load increment adds directly to the bearing equivalent dynamic load P. DIN Class 5 gears in EP-FAD/FAB reduce P by approximately 5–10% relative to DIN Class 6–7 — which, through the cubic L10 formula, translates to approximately 15–33% additional bearing life over the calculated nominal value.

 External resources: cvjointdriveshaft.com for drive shaft life and worm-reducers.xyz for worm reducer life comparison.

Request Life Calculation →