Planetary Gear Motor Backlash Specification & Sourcing Guide | Planetary Gear Motor Manufacturer

Q: Does backlash increase over the service life of the machine?

Wear can increase play, but there is no universal arcmin per hour rate or guaranteed stable period. Preloaded designs can behave differently. Agree an initial limit and, if needed, an end of test limit under specified torque, speed, lubrication, and temperature. Do not convert a generic service life figure into a backlash guarantee.

To specify backlash in planetary gear motors, start with the allowable output position error, convert it to arcminutes, and reserve part of that budget for torsional deflection and the rest of the drivetrain. Then require a maximum backlash value for the exact gearbox ratio, a defined measurement torque, and a minimum torsional stiffness over the working torque range.

A lower catalog backlash value alone does not establish axis accuracy. A machine can still miss its target because of elastic windup, coupling play, mounting compliance, or feedback limitations. Conversely, a continuously loaded conveyor may gain little from a precision gearhead if the torque never changes sign.

This guide helps OEM motion engineers and procurement teams compare suppliers and write an acceptance specification. Use the supplier qualification checklist to prepare your RFQ.

Scope, date, and engineering limits: Published and reviewed on October 27, 2026 for global OEM buyers of planetary gear motors. The calculations below are reproducible screening examples, not hardware measurements or customer case studies. Backlash bands are descriptive purchasing ranges, not ISO, AGMA, or DIN classes. Gear tooth quality grades do not establish an assembled gearbox's angular backlash. Validate the selected unit under its actual load, temperature, duty cycle, and agreed inspection method before production approval.

1. What Is Backlash in a Planetary Gearbox? (The Arcminute Baseline)

Gear tooth backlash is the clearance between mating tooth flanks, commonly measured tangentially at the pitch circle. Catalog angular backlash is the corresponding rotational play measured at the gearbox output with its input constrained. These quantities have different units: in this guide, j_t denotes tangential tooth clearance and b denotes output angular backlash in arcminutes.

Conventional gear meshes need clearance to accommodate manufacturing variation, lubrication, and dimensional changes. Preloaded designs can reduce observable output play while retaining internal working clearances; they still have elastic compliance and application limits. maxon's gear technology overview describes an output-stage preload approach for its reduced-backlash planetary gearheads.

Visualizing Tooth Clearance and Output Backlash

The following unwrapped mesh sketch separates tooth clearance from assembled output angular play. It is conceptual and does not represent an involute tooth profile or a measured gearbox.

Original conceptual sketch, not to scale. The red gap is tangential tooth clearance; output backlash b is the assembled gearbox's angular play.

Defining the Arcminute (Arcmin)

In catalog specifications for rotary planetary gearboxes, precision gearbox backlash is commonly expressed in arcminutes (symbol: ' or arcmin):

1 revolution = 360° (degrees)
1° = 60 arcminutes = 3,600 arcseconds
1 arcminute = ≤ft((1)/(60))° ≈ 0.01667° ≈ 0.00029088 radians

Converting Arcminutes to Linear Positioning Error

To understand what an arcminute rating means on a physical machine, motion control engineers must project the angular play out to the working load radius (R).

The formula for the theoretical maximum linear play (Δ x) at the working tool tip, belt pulley, or robotic end effector is:

Δ x = R · θ_rad = R · ≤ft( Backlash in arcmin · (π)/(180 × 60) )

Let us calculate the resulting physical linear displacement across three typical automation radii:

Short-arm indexer (R = 50 mm):
- At 15 arcmin (Standard gearbox): Δ x = 50 · (15 × 0.00029088) ≈ 0.218 mm (218 µm)
- At 3 arcmin (Precision gearbox): Δ x = 50 · (3 × 0.00029088) ≈ 0.044 mm (44 µm)
- At 1 arcmin (Ultra-precision): Δ x = 50 · (1 × 0.00029088) ≈ 0.015 mm (15 µm)
Long pick-and-place arm / SCARA axis (R = 500 mm):
- At 15 arcmin: Δ x = 500 · (15 × 0.00029088) ≈ 2.18 mm
- At 3 arcmin: Δ x = 500 · (3 × 0.00029088) ≈ 0.436 mm
- At 1 arcmin: Δ x = 500 · (1 × 0.00029088) ≈ 0.145 mm
Large automated gantry pinion / AGV steer radius (R = 1,000 mm):
- At 15 arcmin: Δ x ≈ 4.36 mm
- At 3 arcmin: Δ x ≈ 0.873 mm

These values are the full flank-to-flank displacement at the stated radius, not automatically a symmetric ± positioning error. At a 500 mm radius, 15 arcmin corresponds to about 2.18 mm. Define the position reference and approach direction before comparing that span with a ± tolerance. For a screw drive, use the screw lead per revolution rather than treating screw radius as the linear conversion radius.

2. Static Backlash vs. Lost Motion vs. Torsional Rigidity (`C_t21`)

The most dangerous specification trap in motion control is assuming that a gearbox with 3 arcmin of catalog backlash will only deviate by 3 arcmin under heavy working loads.

Two contributions must be distinguished:

Output angular backlash (b): Rotational play across opposite tooth contacts. A catalog value may include some low-load compliance, depending on the manufacturer's test.
Elastic torsional deflection: The twisting of teeth, shafts, carrier, and other components while transmitting torque.

The Hysteresis Loop: Demystifying Lost Motion

A torque-versus-angle test constrains the input and cycles output torque through an agreed range. The curve shows clearance, elastic stiffness, and loading-history effects. Ask for the manufacturer's definitions and test torque: a universal 1–2% test threshold cannot be assumed. For example, Nabtesco's RVW catalog glossary defines lost motion using ±3% rated torque for that reducer family. Its definition is useful context, not a planetary gearbox acceptance standard.

Original conceptual torque–angle sketch, not measured data. Loaded-region slope indicates torsional stiffness. Extract backlash and lost motion using the supplier's declared method; the drawing has no calibrated torque or angle scale.

Understanding Torsional Rigidity (`C_t21`)

Torsional rigidity (C_t21, expressed in N·m/arcmin) is the angular spring rate of the transmission. It determines how many Newton-meters of applied load torque are required to twist the output shaft by one arcminute.

For a first-pass allowance after crossing the full clearance gap and loading in one direction, use:

E_screen = b + |T| / C_t21

Here b is geometric angular play in arcmin, T is output torque in N·m, and C_t21 is stiffness in N·m/arcmin over the relevant torque interval. This is a quasi-static screening approximation, not a prediction of dynamic overshoot. A full reversal from +T to −T gives an approximate span of b + 2|T| / C_t21 for equal stiffness in both directions. If a catalog backlash value already includes loaded deflection, separate that component before adding stiffness again.

Illustrative Stiffness Comparison — Not a Measured Failure Case

Consider two hypothetical gearboxes at 120 N·m output torque:

Option A: b = 3 arcmin, C_t21 = 12 N·m/arcmin; screening allowance 3 + 120/12 = 13 arcmin.
Option B: b = 5 arcmin, C_t21 = 40 N·m/arcmin; screening allowance 5 + 120/40 = 8 arcmin.

Option B's calculated allowance is about 38% lower despite its greater clearance. This demonstrates the trade-off; it does not establish settling time, resonance, or suitability for a particular robot. Those require load inertia, mounting stiffness, controller tuning, and measured response.

3. Planetary Gearhead Backlash Tolerance Classes

The bands below are illustrative RFQ screening ranges, not standardized grades. Supplier labels such as “precision” differ. Specify the numeric maximum and measurement conditions instead of ordering by a class name. There is no fixed conversion from ISO/DIN tooth accuracy to assembled gearbox backlash, manufacturing method, cost, or encoder resolution.

Descriptive purchasing band	Output backlash b	Linear play at R = 100 mm	Engineering question	Supplier evidence to request
Ultra-low play	≤1 arcmin	≤0.029 mm	Is stiffness or thermal behavior the limiting error?	Torque–angle curve and hot/cold limits
Low play	>1–3 arcmin	>0.029–0.087 mm	Does loaded error still fit the axis budget?	Maximum backlash and minimum stiffness for the exact ratio
Moderate precision	>3–5 arcmin	>0.087–0.145 mm	Are couplings and mounting equally controlled?	Fixture method and interface tolerances
Reduced play	>5–8 arcmin	>0.145–0.233 mm	Can the axis tolerate the reversal span?	Reversing-load validation plan
General positioning	>8–12 arcmin	>0.233–0.349 mm	Is load-side feedback or one-sided loading available?	Feedback strategy and acceptance result
General drive	>12–20 arcmin	>0.349–0.582 mm	Does load torque retain one sign during operation?	Duty, peak torque, and backlash limits
Coarse positioning	>20 arcmin	>0.582 mm	Is free motion acceptable at the load?	End-position and impact-load acceptance criteria

All linear values use Δx = R × b × π/10800 and are rounded. These ranges do not qualify a product for medical, lifting, or other safety-related use.

4. Multi-Stage Gearbox Backlash: How Stages Accumulate Play

A frequent question from OEM purchasing teams is: "Why does a 100:1 three-stage planetary gear motor have higher catalog backlash than an equivalent 5:1 single-stage unit?"

To understand stage accumulation, we must look at how gear stages are coupled from input sun gear to output carrier:

The Mathematical Rule of Backlash Distribution

In a multi-stage planetary gearbox, backlash generated at an earlier stage is divided by the gear ratio of all subsequent downstream stages when viewed from the final output shaft!

Let us formalize this:

Let b_1 be the angular backlash at the output of the first stage (closest to the motor).
Let b_2 be the angular backlash at the output of the second stage.
Let b_3 be the angular backlash at the output of the final output stage.
Let i_2 and i_3 be the gear ratios of stages 2 and 3.

The total backlash measured at the gearbox output shaft is:

b_total = b_3 + b_2 / i_3 + b_1 / (i_2 × i_3)

The Final Stage Often Dominates — Calculate the Actual Split

For three hypothetical stages each having 5 arcmin at its own output, with downstream ratios i_2 = i_3 = 5, final-output play is 5 + 5/5 + 5/25 = 6.2 arcmin. The final stage contributes about 81% in this example; the proportion changes with stage clearances and ratio split.

This kinematic sum assumes rigid stages, a locked input, and all clearances taken up in the same sense. Include coupling or spline play separately. Additional stages do not impose a universal backlash increment or efficiency value; ask for the assembled unit's declared maximum and working-point ratings.

5. How Backlash Is Measured and Audited: The Inspection Protocol

Supplier and buyer measurements can disagree if they use different torque, fixtures, shaft positions, or temperature. Agree the test method before ordering.

Agreed Locked-Input / Lever-Arm Protocol

This is a suggested acceptance setup, not a test prescribed here by an ISO or DIN standard:

Constrain the input: With motor power isolated, lock the input using the supplier-approved fixture and bolt the housing to a rigid stand. If an integrated motor prevents access, agree an alternative fixture and document its compliance.
Measure the output: Attach a lever of known effective radius L using a rigid clamping connection. Align a displacement indicator tangentially at that radius, or use a calibrated output rotary encoder.
Declare conditions: Record model, ratio, serial number, lubricant, temperature, preconditioning, input restraint, fixture compliance, and output angular positions sampled.
Agree the torque and definition: Apply a defined positive and negative output test torque within permitted limits. Do not assume 2% rated torque isolates pure clearance: friction and elasticity remain relevant. For a hysteresis test, also agree sweep range and speed.
Convert the reading: For small angles, b_measured (arcmin) = (ΔS/L) × 10800/π ≈ (ΔS/L) × 3437.75, with displacement and lever radius in the same units. A 0.100 mm span at 200 mm gives about 1.72 arcmin.
Set acceptance: State the maximum permitted reading, sampled positions, measurement uncertainty, sampling plan, and any hot/cold or endurance checks. A lot average is not a guaranteed maximum for each unit.

Common Incoming-QC Pitfalls

Unrestrained rotor: Motor cogging or rotor movement can be mistaken for gearbox play. An unpowered motor is not a locked input.
Fixture or keyway play: Clearance at the lever, coupling, or input restraint adds to the reading. Check fixture repeatability before rejecting a gearhead.
Radial or axial loading: Bearing motion and misalignment can contaminate the indicator reading. Control the applied loads and sensor alignment.
Resolution versus accuracy: Encoder bit count and indicator resolution alone do not establish calibration accuracy or uncertainty.

6. The Engineering & Cost Trade-Off: How to Avoid Over-Specifying

Tighter backlash can increase manufacturing, assembly, or preload requirements, but there is no universal price multiplier. Compare quotations for the same frame, ratio, torque, bearings, interfaces, lubrication, quantity, warranty, and test documentation. Include feedback hardware and validation cost in the comparison.

Strategy 1: Unidirectional Motion Profiling

If the transmitted torque retains one sign, the gear mesh can remain loaded against the same flanks, making clearance less important to steady operation. Rotation in one direction alone does not guarantee this: braking, an overhauling load, or disturbances can reverse torque.

Compare a standard and a lower-backlash gearbox against the actual error and duty requirements. Do not assume equal accuracy or a fixed cost saving.

Strategy 2: Dual-Loop Feedback (Direct Load-Side Encoders)

In precision rotary indexers or ballscrew drives, position can be measured at two locations:

Motor-side feedback: An encoder mounted on the rear of the motor shaft. The controller assumes the output position based on motor angle divided by the reduction ratio. Here, any backlash or torsional windup in the gearbox produces undetectable load error.
Dual-loop feedback: A linear optical scale or magnetic ring encoder mounted directly on the driven load, while the motor encoder handles high-frequency velocity and commutation loops.

Load-side feedback can correct settled position error, but it does not remove mechanical clearance or guarantee stability during reversal. Include sensor accuracy, sampling, tuning, and settling requirements. FAULHABER's gearhead support guidance discusses output play, the limits of motor-side feedback, and external feedback compensation; its miniature-gearhead backlash ranges are not precision planetary specifications.

Strategy 3: Software Backlash Compensation

Some controllers offer direction-dependent backlash compensation. Confirm support for the exact drive and firmware, then tune and validate it under the real load. Compensation can correct a repeatable offset but does not add stiffness, suppress all impact, or track arbitrary wear and temperature changes.

Strategy 4: Mechanical Preloading (Split-Gear or Dual-Drive Systems)

Split-gear or opposing-drive preload can keep selected flanks in contact. The preload path must cover every relevant clearance, including the gearboxes and external rack or pinion interfaces. Assess the added torque, heat, bearing load, control complexity, and fail behavior rather than assuming a second motor eliminates all play.

Kinematic Sizing & Backlash Sourcing Decision Matrix by Application

Use a declared full span of linear motion for the clearance budget. If a requirement is written as ±e, establish whether a centered 2e span or a stricter one-sided e allowance applies before calculation. The examples below allocate half of the stated span to geometric backlash; the other half remains for elastic and other errors. This 50% allocation is an explicit planning assumption, not an industry rule.

b_budget (arcmin) = allocated linear clearance span / R × 10800/π

Illustrative axis	Full linear error span	Effective radius R	Clearance allocation	Calculated b ceiling	Qualification focus
Pick-and-place rotary arm	0.100 mm	450 mm	0.050 mm	0.382 arcmin	Sub-arcmin budget may require a different transmission or feedback architecture
Inspection rotary arm	0.160 mm	600 mm	0.080 mm	0.458 arcmin	Include sensor and structural errors
Packaging indexing turret	0.300 mm	300 mm	0.150 mm	1.719 arcmin	Check stiffness at acceleration torque and settling time
AGV traction wheel	4.000 mm	100 mm	2.000 mm	68.755 arcmin	Wheel slip, tire deformation, torque, and impact may govern; ceiling is not a recommendation
Tracking pivot	26.180 mm	1500 mm	13.090 mm	30.000 arcmin	Example span corresponds to 1° total; check wind deflection separately
Machining rotary fixture	0.040 mm	150 mm	0.020 mm	0.458 arcmin	Cutting-load stiffness and output feedback may dominate
Small rotary dosing cam	0.020 mm	25 mm	0.010 mm	1.375 arcmin	Tangential displacement is not a flow-accuracy guarantee
Sorter diverter arm	3.000 mm	200 mm	1.500 mm	25.783 arcmin	Assess reversal noise, end-position needs, and impact limits

For the packaging example, if 0.100 mm of the remaining 0.150 mm span is reserved for symmetric elastic reversal at ±20 N·m, its angular allowance is 1.146 arcmin. The linear-stiffness model requires C_t21 ≥ 2 × 20 / 1.146 ≈ 34.9 N·m/arcmin; 0.050 mm remains for other errors. A real stiffness curve may be nonlinear, so verify this allocation using the supplier's test data.

These are independent numerical examples, not complete machine designs. Gearbox ratings, safety functions, fatigue life, and production acceptance remain separate requirements. Discuss your error budget and RFQ conditions with engineering before selecting a catalog class.

7. OEM Procurement & Supplier Backlash Qualification Checklist

Include these eight items in the drawing, RFQ, and quality agreement:

Maximum output backlash: State arcmin, exact model and ratio, test torque, temperature, sampled positions, and whether the limit applies initially or after agreed endurance testing. Avoid undefined “zero backlash” claims.
Stiffness and error budget: Request minimum stiffness or the torque–angle curve over the working range. Distinguish one-direction loading from full reversal, and avoid counting compliance twice.
Input and output interfaces: Specify shaft fits, coupling or clamping design, tightening torque, and permitted interface play. Include mounting and bearing compliance in the axis budget.
Materials and wear evidence: Ask for the supplier's material, heat-treatment, finish, and endurance evidence appropriate to the design. A backlash number alone cannot dictate hardness or a manufacturing process.
Multi-stage configuration: Record stage count, exact ratio, and assembled-output guarantee. Request the ratio split or stage data where available, without substituting component values for an assembled test.
Temperature and lubrication: Agree ambient and operating limits, duty cycle, approved lubricant, mounting position, and any cold-start or hot-running acceptance test.
Traceable QC evidence: Define whether every unit or a lot sample is tested. Require model, ratio, serial or lot reference, method, conditions, result, and calibration traceability.
Incoming inspection alignment: Agree fixtures, torque, lever radius or encoder accuracy, uncertainty, sampling, and rejection criteria before production release.

FAQ

Does lower backlash necessarily reduce gearbox efficiency?

No universal relationship can be inferred from the backlash number. Preload, seals, lubrication, speed, torque, and temperature all matter. Request efficiency and thermal data at the working point instead of assigning a fixed penalty to a precision class.

Why can a 1-arcminute gearbox still produce much greater axis error?

Backlash is only one part of the error budget. Torsional windup under load, coupling play, shaft or mounting flex, sensor accuracy, and control response can dominate. Measure the load position and torque–angle behavior before specifying a tighter gearbox.

Does backlash increase over the service life of the machine?

Wear can increase play, but there is no universal arcmin-per-hour rate or guaranteed stable period. Preloaded designs can behave differently. Agree an initial limit and, if needed, an end-of-test limit under specified torque, speed, lubrication, and temperature. Do not convert a generic service-life figure into a backlash guarantee.

How does ambient temperature affect planetary gearbox backlash?

The direction and magnitude depend on material expansion, internal geometry, fits, bearing preload, and temperature distribution. Lubricant behavior also changes. A housing material alone cannot predict the result: request hot and cold measurements for the exact configuration and duty cycle.

Can software compensation replace a low-backlash gearbox?

It can correct a predictable position offset in some axes, but it does not eliminate physical clearance, increase torsional stiffness, or guarantee acceptable reversal dynamics. Validate settling time, contouring error, impact, and stability with the real load before relaxing the mechanical specification.

Sources and Measurement Boundaries

Source review: October 27, 2026. These manufacturer references support the stated concepts; they do not establish universal backlash classes, market prices, or performance of this site's products.

FAULHABER — Gearheads Support: output play and motor-side versus external position feedback. Its product-specific ranges should not be applied to all planetary gearboxes.
maxon — Gears, Technology: Short and to the Point (2023, printed pp. 72–73): selection limits, lubrication, and a reduced-backlash planetary design using output-stage preload. Product statements apply to maxon's described families.
Nabtesco — RVW Catalog, glossary, printed p. 32: an explicit distinction between zero-torque backlash, low-torque lost motion, and hysteresis. This is an RV reducer-family definition, not a planetary gearbox inspection standard.

The numerical tables and stiffness comparisons above are this article's derived examples using their displayed inputs and equations. ISO 1328-1 tooth-flank tolerance grades are a different specification from assembled output backlash; quote any applicable standard's exact edition and scope separately in the supplier agreement.

Engineering Support: Optimize Your Drivetrain Accuracy and Cost

Selecting the optimal planetary gear motor requires striking a precise balance between kinematic positioning accuracy, dynamic torsional stiffness, thermal limits, and total production cost. Over-specifying can increase cost, while an insufficient error budget can cause positioning failures and rework.

If your project involves complex mobile robotics, automated material handling, or demanding access systems, compare your kinematic requirements with our specialized engineering resources:

Review our detailed gear ratio calculation methodology in the planetary gear ratio selection guide for AGVs and mobile robots.
Address acoustic noise and vibration challenges with our low-noise (NVH) planetary gear motor sourcing guide.
Troubleshoot open-loop step loss and mechanical stiffness in our NEMA stepper gear motor troubleshooting guide.
Standardize your incoming receiving procedures with the pre-shipment QC checklist for gear motor orders.

Ready to validate your motion axis specification?
Share your target model or envelope, ratio, output torque, working radius or screw lead, error budget, temperature range, and reversing duty so the engineering team can review the available options.

Request an engineering review and OEM quotation.
Direct technical inquiries and drawing reviews can be sent to [email protected] with your target axis inertia, reversing duty cycle, allowable angular play (arcmin), and motor interface dimensions.