What Backlash Is — and the Common Misconceptions That Lead to Wrong Specifications
Backlash is the angular displacement of the gearbox output shaft when the input reverses direction and the output does not yet move — the “dead zone” at every direction reversal, caused by the clearance between meshing gear teeth. It is measured in arcminutes (1 arcminute = 1/60°) at the output shaft, with the input shaft locked.
Backlash is not a manufacturing defect. Every planetary gearbox requires tooth clearance for two unavoidable physical reasons: the lubricant film between gear teeth needs physical space to form — zero clearance means the oil film is squeezed out and metal-to-metal contact occurs. And thermal expansion of the gear teeth during operation requires clearance to accommodate the dimensional change without seizure. A gearbox with genuinely zero backlash would overheat and seize within minutes of reaching operating temperature.
What the P0/P1/P2 grade system controls is how much backlash — the precision of the tooth clearance. Tighter clearance (lower backlash) requires more precise gear grinding, closer dimensional tolerance in the housing bores, and selective assembly pairing of gear sets — all of which increase manufacturing cost. The engineering discipline is to specify the minimum grade that your application actually requires, not the minimum that the catalogue offers.
Systems marketed as “zero backlash” achieve this through mechanical preload — spring-loaded or mechanically-biased components that keep gear flanks in constant contact. Preload reduces effective backlash but introduces constant friction and wear, reducing efficiency and service life. Standard Korea Ever-Power EP planetary gearboxes use engineered tooth clearance (P0/P1/P2) rather than preload — the specified backlash is the genuine clearance, not a preloaded approximation.
The practical implication: when a servo axis with a P0 ≤1 arcmin gearbox shows positioning drift, the cause is almost never the gearbox backlash — it is more likely servo tuning, thermal expansion of the machine structure, or spindle runout. Backlash shows as a specific signature: consistent angular error in one direction only immediately following a direction reversal, equal to the backlash magnitude. If the error is random, grows with temperature, or appears in both directions simultaneously, the cause is not backlash.
Backlash — The Physical Mechanism
Input shaft moves
Output shaft: STATIONARY
(clearing tooth clearance)
←←← θ_backlash ←←←
Input continues reversing →
Output shaft begins moving again
Measured: θ_backlash = angular gap
at output shaft during reversal
The Arcminute Scale — Translating Grade Numbers into Real Positioning Error
Understanding planetary gearbox backlash explained in arcminute terms — rather than degrees — is essential for accurate specification. The arcminute is the unit of angular measurement used for backlash because it is appropriately sized — 1 arcminute equals 1/60 of one degree. For a typical servo application, backlash values span 1 to 30 arcminutes. Expressing these values in degrees (0.017° to 0.5°) is unwieldy; expressing them in radians (0.000291 to 0.00873 rad) loses intuitive scale. Arcminutes sit at exactly the right resolution for practical specification and measurement.
The linear positioning error produced by a given backlash depends on the radius at which the error is measured — the lever arm between the gearbox output shaft and the workpiece or tool contact point. The relationship is:
BACKLASH → LINEAR ERROR FORMULA
θ (rad) = arcmin × π / (180 × 60) = arcmin × 0.000291
At r=100 mm, 1 arcmin → Δx = 100 × 0.000291 = 0.029 mm
Engineering implication: the same gearbox backlash produces different linear errors at different radii. P1 ≤3 arcmin on a rotary table driving a 50 mm radius workpiece produces 0.044 mm error — within IT9 tolerance for aluminium parts. The same P1 on a 200 mm radius table produces 0.175 mm — outside IT9 tolerance for any material. This is why backlash grade must be calculated from the actual working radius, not read from a general application table.
Backlash Grade → Linear Positioning Error at Different Radii
| Grade / Backlash | r=25 mm | r=50 mm | r=100 mm | r=200 mm |
|---|---|---|---|---|
| P0 · ≤1 arcmin | ≤0.007 | ≤0.015 | ≤0.029 | ≤0.058 |
| P1 · ≤3 arcmin | ≤0.022 | ≤0.044 | ≤0.087 | ≤0.175 |
| P2 · ≤5 arcmin | ≤0.036 | ≤0.073 | ≤0.145 | ≤0.291 |
| AE/AER · ≤8 arcmin | ≤0.058 | ≤0.116 | ≤0.233 | ≤0.465 |
| Economic · 6–8 arcmin | 0.044–0.058 | 0.087–0.116 | 0.175–0.233 | 0.349–0.465 |
| Worm typical · 15–30 arcmin | 0.11–0.22 | 0.22–0.44 | 0.44–0.87 | 0.87–1.75 |
All error values in mm. Errors are reversal errors — the lost motion when direction is changed. Use r = distance from output shaft centreline to workpiece or tool contact point.
P0, P1, P2 — How the Grades Are Defined and What Produces the Difference
The three precision grades reflect three different manufacturing tolerance bands applied to the same gear geometry. P0 is not a different type of gearbox — it is the same planetary design produced to tighter dimensional tolerances throughout the manufacturing process. Understanding what manufacturing steps produce each grade helps in understanding why the cost increases with grade.
Single ≤1′ · Two-stage ≤3′
Gear teeth ground to minimum functional clearance. 100% selective assembly: each gear set is measured after grinding and paired with a ring gear whose bore deviation complements the tooth deviation to produce a total tooth clearance within ≤1 arcmin. Every unit measured at the output shaft before delivery — certificate issued with each gearbox confirming the measured value.
Single ≤3′ · Two-stage ≤5′
Close-tolerance gear grinding within a wider band than P0 — achievable without 100% selective assembly. The gear grinding process produces consistent tooth profiles within a tolerance that ensures the assembled total stays within ≤3 arcmin. Batch sampling rather than 100% individual measurement. Cost-efficient for applications requiring better-than-standard precision without the extreme tightness of P0.
Single ≤5′ · Two-stage ≤7′
Normal production tolerance gear grinding. Standard assembly process without special pairing or 100% individual measurement. Produces the baseline tooth clearance that the housing, bearing, and gear geometry naturally achieve without additional tolerance control steps. Correct specification for all non-precision servo axes and general actuators.
All Korea Ever-Power EP precision series are measured at the output shaft before delivery. For P0, a backlash certificate is issued with every unit confirming the measured value — not just a grade conformance statement. The EP-AB P0 series certificate shows the actual measured backlash (e.g. “0.82 arcmin measured at output shaft, 14 Dec 2025”) — verifiable at commissioning by measuring the new unit before installation.
Series Without P Grades — Fixed Backlash and When That Is Correct
Not all Korea Ever-Power series use the P0/P1/P2 grade system. Some series have a fixed backlash specification that does not change with any ordering option — either because the series design produces a fixed tolerance regardless of manufacturing adjustment, or because the application class the series targets does not require grade selection. Understanding which series lack P grades prevents specification errors.
EP-AFH — Fixed ≤1 arcmin Standard
Every EP-AFH unit, at every frame size (060–240 mm) and every ratio (3–100:1), is manufactured to ≤1 arcmin backlash as its only specification — there is no grade code because there is no lower-tier option. The entire series is built to P0-equivalent quality. Specifying EP-AFH eliminates the P grade selection step: ≤1 arcmin is guaranteed at delivery. Backlash up to 3,805 N·m at 240 mm frame.
EP-AE/AER — Fixed ≤8 arcmin (single), ≤12 (two-stage)
The large-flange series targets applications where IP67 protection or large overturning moment resistance is the primary requirement — food processing, outdoor conveyors, wet wash zones. Backlash precision is secondary and ≤8 arcmin fixed is adequate for all these drive types. No P grade selection available — attempting to specify P0 or P1 on AE/AER is not a valid option.
EP-AH/AHK New Line — Fixed 1–2 arcmin
The New Line heavy-duty series achieves 1–2 arcmin backlash as its inherent manufacturing standard — the structural housing and precision gear train designed for 9,585 N·m service produce this backlash as a result of their manufacturing process, without a selectable P grade option. Adequate for all heavy industrial rotary tables and crane slewing drives where P0-class precision is not a functional requirement.
Economic Line — 6–8 arcmin (inline), 10–12 (right-angle)
Economic Line targets induction motor and stepper motor applications where the drive system itself does not use closed-loop position control. Backlash at 6–8 arcmin is irrelevant for conveyor speed control, agitator drives, and open-loop positioning. No P grade option — specifying P0 on Economic Line is a category error; use EP-AB or EP-AF precision series if backlash below 5 arcmin is required.
Matching Backlash Grade to Application — The Complete Decision Table
With planetary gearbox backlash explained from first principles, the correct grade selection flows from three inputs: (1) the required positioning accuracy at the workpiece or tool tip, (2) the working radius (lever arm), and (3) what percentage of the total error budget the gearbox backlash is allowed to consume. Work through the formula Δx = r × θ (radians) first, then cross-reference to the table below.
| Sollicitatie | Required Accuracy | Working Radius | Grade | Korea Ever-Power Series |
|---|---|---|---|---|
| 5-axis titanium rotary table | ±0.02° (1.2′) | 100–200 mm | P0 | EP-AFH (no grade code needed) |
| Collaborative robot (all joints) | ±0.02° (1.2′) | 250–1,000 mm | P0 | EP-AB P0 042–090 |
| CNC register correction axis | ±0.05° (3′) | 50–100 mm | P0 or P1 | EP-AB P1 / EP-AF P1 |
| Packaging VFFS forming axis | ±0.1° (6′) | 30–80 mm | P1 | EP-AB P1 / EP-AF P1 |
| General servo positioner / turntable | ±0.15° (9′) | 50–150 mm | P2 | EP-BAB P2 |
| Food conveyor head drive | Speed only | N/A | No grade | EP-AER IP67 (6–8′ fixed) or Economic Line |
| Heavy rotary table (large workpiece) | ±0.1° (6′) | 200–400 mm | 1–2′ fixed | EP-AH New Line (1–2′ inherent) |
A 12-axis CNC machine tool with 4 precision axes (rotary table, B-axis ×2, and one linear register axis) and 8 auxiliary axes (chip conveyor, coolant, pallet shuttle, doors, ATC carousel) costs approximately 30–40% more in gearbox BOM if all 12 axes are specified at P0. Applying P0 to the 4 precision axes and P2 or Economic Line to the 8 auxiliary axes delivers the same machine performance at significantly lower component cost — the auxiliary axes cannot benefit from P0 precision because their functions (conveyor speed, door position, coolant flow) are not affected by backlash below ±0.15°.
How Backlash Grows in Service — Wear Mechanics and the Replacement Threshold
Backlash does not remain at its delivery value throughout service — it grows as gear tooth flanks wear at the reversal contact point. Understanding the wear mechanics helps engineers set realistic service life expectations and design appropriate monitoring intervals.
The wear mechanism: During normal driving motion (before reversal), the driven flank of each planet gear tooth is in continuous rolling contact with the driven flank of the ring gear — lubricated contact, manageable wear rate. At the instant of direction reversal, the previously unloaded coast flank of the planet gear tooth makes impact contact with the coast flank of the ring gear tooth as the clearance is closed. This reversal impact is a micro-shock: a brief high-contact-stress event that produces localised surface fatigue (pitting) on both tooth flanks. Over millions of reversals, the accumulated pitting increases the effective tooth clearance — which manifests as growing backlash.
The wear rate depends strongly on: the applied torque at reversal (higher torque = higher impact contact stress), the reversal frequency (more reversals per hour = more impact events per hour), the lubrication condition (degraded grease = higher metal-to-metal contact during impact), and the inertia ratio (higher load inertia = higher kinetic energy in the impact event).
BACKLASH GROWTH PROFILE — P0 GEARBOX
Hours 5,000–12,000: Gradual wear, ~1.2–1.4′
Hours 12,000–18,000: Accelerating, ~1.5–1.8′
Hours 18,000–20,000: Near threshold, ~1.9–2.0′
Hours 20,000+: Replacement recommendedThreshold: 2× delivery value (≤2.0′ for P0)
At 2×: functional accuracy is measurably
degraded and replacement warranted.
Factors that accelerate wear beyond this profile: operating continuously at 90–110% of rated torque, inertia ratio above 10:1 (higher impact energy at reversal), poor grease condition (seal degradation in contaminated environments), and high reversal frequency (robot joints at 3-shift operation reach 10× the reversal count of CNC table drives).
The EP-BAB precision planetary series uses permanently sealed grease — no periodic re-lubrication is required. The sealed construction also prevents the contamination-induced wear acceleration that affects non-sealed gearboxes in Korean factory environments with metallic dust, coolant mist, or food product residue.
Two-Stage Backlash Accumulation — Why P0 Single-Stage ≤1′ Becomes P0 Two-Stage ≤3′
The most frequently misunderstood backlash specification is the two-stage value. Engineers encountering “P0: single-stage ≤1 arcmin, two-stage ≤3 arcmin” for the first time often ask whether this means two-stage has worse precision — why does adding a stage triple the specification? The answer lies in how backlash from multiple stages combines at the output shaft.
The correct accumulation formula (not a simple sum of the two individual stages):
TWO-STAGE BACKLASH ACCUMULATION
θ_stage1 = input stage backlash (arcmin)
θ_stage2 = output stage backlash (arcmin)
i_stage2 = output stage ratio
Example (P0, each stage 0.9′):
i_stage2 = 5
θ_total = 0.9 + (0.9 / 5)
= 0.9 + 0.18
= 1.08 arcmin
Conservative upper bound at max values:
θ_total = 1.0 + (1.0 / 3) [i_stage2=3 minimum]
= 1.0 + 0.33 = 1.33 arcmin
Korea Ever-Power specifies ≤3 arcmin for P0
two-stage — providing margin for worst-case
stage ratio combinations and manufacturing
variation over production batches.
Why the output stage dominates: The input stage backlash is divided by the output stage ratio before adding to the output stage backlash. At a typical output stage ratio of i=5, a 1 arcmin input stage backlash contributes only 0.2 arcmin at the final output shaft — one-fifth of the input stage value. The output stage backlash contributes its full value (1 arcmin) directly. This is why precision at the output stage matters more than precision at the input stage for two-stage gearbox specifications.
The same logic applies to right-angle gearboxes with a bevel stage: the bevel stage contribution is divided by the bevel stage ratio before adding to the planetary stage output. However, bevel stage ratios are typically 1:1 (the bevel only changes direction without ratio change) — so in right-angle integrated gearboxes, the bevel contribution adds directly without division. This is accounted for in Korea Ever-Power’s EP-ABR/ADR/AFR specifications, which measure total backlash at the right-angle output shaft.
STAGE CONTRIBUTION DIAGRAM
↓
[STAGE 1] θ₁ = 1.0 arcmin
↓ (reduced by i₂ = 5)
contribution = 1.0/5 = 0.2′ at output
↓
[STAGE 2] θ₂ = 1.0 arcmin
↓ (adds directly at output)
contribution = 1.0′ at output
↓
TOTAL at output = 0.2 + 1.0 = 1.2 arcminKey insight:
Stage 1 backlash is attenuated by i₂
Stage 2 backlash is NOT attenuated
→ Output stage quality matters most
The specification provides margin above the typical calculated value (~1.0–1.2 arcmin) to account for manufacturing variation across production batches, the full range of ratio combinations (where lower i₂ increases stage-1 contribution), and realistic worst-case tolerance stack-ups. Korea Ever-Power guarantees ≤3 arcmin at the output shaft; typical delivered values for P0 two-stage units are 1.1–1.8 arcmin.
Right-Angle Gearbox Backlash — Why the Bevel Stage Is Already Included
The most persistent backlash misconception in Korean machine tool design is the belief that right-angle planetary gearboxes add bevel gear backlash on top of the planetary stage specification — that a P0 right-angle unit is effectively P1 at the output because of the extra bevel contribution. This is incorrect for Korea Ever-Power integrated right-angle series, and the measurement method explains why.
For EP-ABR, EP-ADR, and EP-AFR, the P0/P1/P2 backlash is measured at the right-angle output shaft — the actual shaft that connects to the machine. The measurement procedure locks the motor input shaft, applies a small torque to the right-angle output shaft in each direction, and records the angular displacement. This single measurement captures the combined contribution of all stages — planetary stages and bevel stage together — in one result.
The stated P0 ≤1 arcmin for EP-ABR is the total system backlash at the right-angle output shaft, including bevel. It is not the planetary-stage backlash with an additional bevel contribution on top. When a Korea Ever-Power certificate says “EP-ABR090 P0: 0.76 arcmin measured at output shaft,” that 0.76 arcmin includes the bevel contribution — and the total is within the P0 specification.
If a supplier states a backlash value for a right-angle gearbox, always verify: “Is this measured at the final right-angle output shaft with the bevel stage active?” If the answer is “this is the planetary stage specification — the bevel adds extra,” the total output backlash is higher than stated. Korea Ever-Power provides the measurement point confirmation in writing on the delivery certificate for all right-angle series.
| Configuration | P0 Spec | Measured at |
|---|---|---|
| EP-AB (inline) | ≤1 arcmin | Inline output shaft ✓ |
| EP-ABR P0 (R/A) | ≤1 arcmin | R/A output shaft ✓ (bevel included) |
| EP-AFR P0 (R/A) | ≤1 arcmin | R/A output shaft ✓ (bevel included) |
| Competitor: “P0 planetary + bevel” | ≤1′ stage only | Before bevel ⚠ (+3–5′ extra) |
How to Specify Backlash Grade When Ordering — and Frequently Asked Questions
KOREA EVER-POWER ORDER CODE FORMAT — BACKLASH GRADE EXAMPLE
Specify the Right Backlash Grade — Korea Ever-Power Application Support
Korea Ever-Power’s Korean application team calculates the required backlash grade from your positioning specification and working radius, confirms the appropriate series, and issues a grade certificate with every P0 and P1 delivery — in Korean, same working day.
Redacteur: Cxm
