Designing for 500,000-cycle gearmotor life

Half a million cycles is often under fifty hours of running. The wear driver is starts, reversals and grease age, and the test that proves it takes a month and a half of real time.

Wedge
Obsolescence, second source, and reshoring
Catalog reviewed
2026-08-18
Configurations screened
32 published

Any mechanism specified in cycles rather than hours: a cam that indexes, a door that opens and closes, a carousel that steps, a valve that strokes, a dispenser that meters. The distinguishing feature is that each cycle is a short move followed by a long dwell. This is the single most common life specification in OEM motion, and it is also the one most often carried across from an incumbent supplier without anyone having established what it means physically.

Convert the cycle specification into the quantities that actually wear things, and do it before selecting anything. Total output revolutions is cycles times revolutions per cycle. Total motor revolutions is that times the ratio. Total running time is cycles times the moving time per cycle. Number of starts is cycles times the starts per cycle, which is two for anything that reverses. Then the thermal quantity: mean-square current is the integral of i²dt over one period divided by the period, and the starting transient contributes to it out of all proportion to its duration because the starting current is an order of magnitude above the running current.

What normally controls the selection

Which of four independent mechanisms runs out first: brush and commutator wear, which tracks commutation events; gear surface fatigue, which tracks tooth contact cycles and is a hundred times worse at the first stage than at the last; grease condition, which tracks calendar time and temperature more than it tracks motion; and bearing life, which tracks revolutions. These do not scale together. A cycle count alone determines none of them, which is why the first task is to convert the specification into all four quantities and see which one is close to a limit.

Four conditions, and they are usually decided by four different numbers.

Sizing on one of them is how a mechanism gets a part that works on the bench and fails in the field.

Row Continuous Sizes the thermal case Peak Sizes the torque case Starting Sizes the transient Abnormal Sizes the protection
What the mechanism is doing There is no continuous condition. The RMS current over a full period, dwell included, is the number that stands in for it, and on a cycling mechanism it is usually well under the rated current.The acceleration phase of each move. It is short and it repeats a million times, which makes it a fatigue input rather than a thermal one.This is the condition the specification is really about. Every cycle is one or two starts, and the starting current is 5 to 20 times the running current on the published configurations. The commutation interface sees that current at a nearly stationary rotor.A jam somewhere in half a million cycles is a certainty rather than a possibility. The life specification has to account for the drive's stall behaviour, because on 29 of the 32 published configurations a stall exceeds the gearhead's rating.

Failure modes and misleading specifications

What follows is what actually fails, and which published number sends people the wrong way.

Common failure modes
Failure mode Physical cause What it looks like in the field
Commutator wear concentrated at a few segmentsThe mechanism always starts and stops at nearly the same rotor position, so the same segment pair carries the starting current on every cycle.Localized commutator erosion or brush transfer at regular angular intervals, on a motor whose total running hours are far below any brush-life figure.
First-stage pinion wear with an undamaged output stageContact cycles scale with the ratio. At 100:1 the first-stage pinion sees a hundred times the tooth contacts of the output stage.Pitting or polishing on the motor pinion and the first wheel, with the output stage still showing its original tooth flank finish.
Grease failure before any wear limit is reachedThe mechanism runs for minutes a day over ten years. The grease ages by calendar and temperature, not by motion, and the first stage throws it out of the mesh.Rising current and rising noise on units with very low accumulated running time. Dry or separated grease at teardown with intact tooth flanks.
Backlash growth at the reversal pointA million reversals, each traversing lost motion and impacting the trailing flank under acceleration.Increasing reversal error and increasing reversal noise, with flank damage confined to one side of the teeth.
Specifications that mislead here
Specification Why it misleads in this application
Rated life in hoursIt describes continuous running and it is normally set by brush wear. A cycling mechanism accumulates hours very slowly and starts very quickly, and the hours figure says nothing about starts.
A cycle count with no cycle definitionHalf a million cycles of a 5 degree nudge and half a million cycles of a full revolution differ by a factor of 72 in every quantity that matters. The definition is the specification.
Duty cycle expressed as a percentageA 3 percent duty implies a thermal margin, and the thermal margin is real. It also implies that the remaining 97 percent is free, which it is not: grease ages and mechanisms sit still under load.
MTBFIt is a constant-hazard figure derived from a population running continuously. It does not describe wear-out, and every mechanism on this page fails by wear-out.

The tradeoffs that matter here

Spur or planetary reduction

Row Spur Planetary
When it is right Lower torque, and where the simpler train makes the wear inspection at teardown unambiguous. Fewer meshes means fewer contact-cycle populations to account for.Where the jam has to be survived by the drivetrain, or where the output torque needs it.
What it costs Lower output rating, so the inevitable jam over half a million cycles is more likely to exceed it.More meshes, more grease reservoirs to age, and more lost motion to grow at the reversals.

Brushed coreless or brushless

Row Coreless brushed Brushless
When it is right The converted numbers show that starts and running hours are both comfortably inside what the commutation interface can take, and that has been demonstrated rather than assumed.High starts per hour, a long calendar life, or a program that cannot carry a brush-wear argument through its own reliability review.
What it costs One of the four wear mechanisms is a consumable that no calculation can bound without test evidence. On a start-dominated duty it is the one most likely to surprise.A driver and its own reliability case. Bearings and grease still age, so brushless removes one wear mechanism out of four rather than solving life.

How life is demonstrated

Row Real-time test at the real duty Compressed duty, shortened dwell Elevated temperature or load
When it is right The duty is fast enough that the full count fits in the program schedule. In the worked example this is 46 days.The schedule cannot take real time and the failure mechanism of interest is motion-driven.A specific mechanism is being accelerated with a known model, such as grease life against temperature.
What it costs Calendar time, and test stands occupied for the duration. It is the only method that ages the grease and the mechanism together at the right rate.It over-tests the motion and under-tests the calendar. Grease age and thermal cycling are not accelerated by removing the dwell, they are removed from the test.It requires the model to be right. Without one, an accelerated test proves only that the part survived a different test.

Convert a 500 000-cycle specification into the quantities that wear things, for an indexing cam drive.

Cycle definition
90° forward, 90° reverse, then dwell
Cycle count
500 000
Cycle period
8 s
Output torque
6 mN·m running, 12 mN·m peak
Configuration
MM-C1226-S01002-120A, 100.22:1 spur, 12 V
Calculation
Step Expression Result
Move time per cycle0.25 rev ÷ (100 rpm ÷ 60) × 2 moves0.30 s
Total running time500 000 × 0.30 s150 000 s, or 41.7 hours
Total output revolutions500 000 × 0.5 rev250 000 output revolutions
Total motor revolutions250 000 × 100.2225 055 000 motor revolutions
Contact cycles, first stage against last25 055 000 against 250 000The first-stage pinion sees 100 times the tooth contacts of the output stage
Total starts500 000 × 21 000 000 starts
Running current6 mN·m ÷ (100.22 × 0.59) ÷ 10.997 mN·m/A + 2.9 mA12.1 mA, 71 percent of the 17 mA rated current
Electrical loading from running(0.0121 A)² × 150 000 s22.1 A²·s
Electrical loading from starting(0.6 × 0.301 A)² × 0.020 s × 1 000 000652 A²·s, 30 times the running contribution
RMS current over the whole program√[(652 + 22.1) ÷ 4 000 000 s]13.0 mA, inside the 17 mA rated current
Real time to run the full count500 000 × 8 s4 000 000 s, or 46.3 days

Result

Half a million cycles is 41.7 hours of running. Almost no brushed motor fails from 42 hours, which means brush wear is not the risk here and any life argument built on a rated-hours figure is answering the wrong question. What the conversion exposes is that the mechanism performs a million starts, and that those starts contribute 30 times more electrical loading than all the running combined, concentrated at a nearly stationary rotor. It also exposes that the first-stage pinion accumulates 25 million tooth contacts against the output stage's 250 000, so the wear inspection belongs at the motor end of the train and not at the output. Thermally there is nothing to worry about: the RMS current over the full program is 13.0 mA against a 17 mA rating. The four risks in rank order are commutation wear from the starts, first-stage contact fatigue, grease age over the calendar life, and reversal-driven backlash growth. Testing the specification honestly takes 46.3 days at real duty, and that is a program schedule input.

What would change it

The cycle definition dominates everything. Changing the move from 90 degrees to a full revolution multiplies every revolution and contact-cycle figure by four while leaving the start count alone. Shortening the cycle period from 8 s to 2 s quadruples the RMS current, which at 26 mA would exceed the 17 mA rating and turn this into a thermal problem. Removing the reversal halves the starts and removes the backlash-growth mechanism entirely. And the assumed starting transient is the weakest number in the calculation: a measured current trace through a real start would change the 30-to-1 ratio in either direction and should be the first measurement taken.

A short qualification plan

Each step names the measurement, not the intention.

  1. Convert the specification before selecting anything

    Cycles converted to output revolutions, motor revolutions, running hours, starts and RMS current, with the cycle definition written down.

  2. Measure the starting transient

    Current trace through a real start on the assembled mechanism, at both temperature extremes, integrated to give the actual per-start loading.

  3. Run at real duty, not compressed

    Full-count cycling at the real period, so the grease ages and the mechanism thermal-cycles at the rate it will in service.

  4. Tear down at intervals, not only at the end

    Units removed at 10, 25 and 50 percent of the count and inspected. First-stage pinion flanks, commutator surface, grease condition and reversal backlash, each measured against its own baseline.

  5. Include the jam

    Stall events distributed through the life test at the rate the mechanism will really see them, with the torque limit active.

Send the duty definition and a cycled unit.

Send the cycle definition, the period and a unit that has accumulated real cycles. You get back the converted quantities, a teardown against the four wear mechanisms, and a test plan sized to the schedule.

Send the duty definition and a cycled unit