Planetary Gearbox vs Harmonic Drive — Engineering Comparison Guide for Precision Servo Applications

30,000 hr
Planetary P0 L10 Life
5–15k hr
Harmonic Drive Flexspline Life
97–99%
Planetary Efficiency
70–85%
Harmonic Drive Efficiency
<0.1
HD Ultra-Prec. Backlash (arc-min)
≤1
Planetary P0 Backlash (arc-min)

Technology Comparison

Two Fundamentally Different Approaches to Precision Gear Reduction — And Why One Dominates 90% of Applications

EP-FAD planetary gearbox cross-section showing ring gear, planet gears, and sun gear — rolling contact efficiency 97-99% vs harmonic drive flexspline deformation 70-85%

Planetary gearbox cross-section: three planet gears in rolling contact with ring gear and sun gear. Energy transmission through rolling contact surfaces produces 97–99% efficiency. The rolling contact principle also means there are no continuously deflected components — bearing fatigue sets the life limit at 30,000 hr, not a consumable wear element.

Precision servo automation uses two technologies for high-ratio gear reduction: planetary gearboxes and harmonic drives. Both produce gear ratios in the 10:1–160:1 range with backlash performance in the arc-minute class. Both are used in robot joints, precision axes, and semiconductor handling equipment. The question of which to use is one of the most common specification decisions in machine design — and it is consistently answered incorrectly when engineers default to harmonic drives because of the “zero backlash” reputation without evaluating the full comparison.

This comparison is written with a specific premise: harmonic drives genuinely excel at one thing — absolute minimum backlash approaching zero arc-minutes — and at one application requirement — ultra-flat axial profile in the disc form factor. In the ten other parameters that precision servo engineers specify, planetary gearboxes are equal or significantly superior. An honest comparison acknowledges both technologies’ genuine advantages and gives the machine design engineer the framework to make the correct choice for each specific joint.

The harmonic drive principle (wave generator deflecting a thin-walled flexspline into engagement with a circular spline) produces near-zero backlash because the teeth are always in contact at two simultaneous engagement zones — there is no angular play because there is no gap between engaging teeth. This is a genuine engineering advantage for the minority of applications that require positional accuracy below 1 arc-minute without servo backlash compensation. For the majority of precision servo applications, planetary P0 at 0.8 arc-min with servo compensation achieves the required positioning accuracy at significantly lower cost, higher efficiency, longer service life, and greater shock tolerance.

🎯
The Right Specification Question
Do not ask: “Should I use a harmonic drive or a planetary gearbox?” Ask instead: “Does my application require backlash below 1 arc-minute that cannot be achieved by P0 planetary with servo compensation — and is the 5–15,000 hr flexspline life, 70–85% efficiency, and 150–300% price premium acceptable?” If the answer to the first part is yes and the second part is also yes, harmonic drive is the correct choice. For the overwhelming majority of robot joints, CNC axes, and precision automation axes, the answer to the first question is no — P0 planetary with compensation achieves the required accuracy — which makes the second question irrelevant. Use this framework, not the reputation of either technology, to make the selection.

10-Parameter Technology Selection Matrix — Planetary P0 vs Harmonic Drive

The following matrix compares three configurations across ten engineering parameters that precision servo engineers typically specify. Each cell shows the specific value or characteristic, with the column header colour indicating the relative advantage: navy for planetary advantage, amber for harmonic drive advantage, grey for parity. Read each row independently — the correct technology choice emerges from which parameters are decision-critical for your specific application.

Parameter EP-FAD P0
Planetary, precision grade		 Harmonic Drive
Standard series		 HD Ultra-Precision
Premium / superprecision		 Selection notes
Backlash
(arc-min)
≤1 arc-min
(0.78 typical, measured+stamped)
≤1 arc-min
(similar to P0; some models ±0.5)
<0.1 arc-min
(near-zero; true zero with preload)
 P0 planetary ≈ HD standard. HD Ultra wins if <0.1 arc-min is required without servo compensation.
Service life
(design life)
30,000 hr
(L10 bearing fatigue, S1 rated)
5,000–15,000 hr
(flexspline fatigue — consumable)
5,000–10,000 hr
(thinner flexspline = shorter life)
 Planetary wins decisively. Flexspline fatigue is 2–6× shorter than planetary bearing life. HD flexspline replacement is a scheduled maintenance cost.
Efficiency
(at rated torque)
97–99%
(DIN Class 5 rolling contact)
75–85%
(flexspline deformation energy loss)
70–80%
(higher preload = more loss)
 Planetary wins significantly. At 100W motor power, planetary wastes 1–3W; HD wastes 15–30W. Motor sizing, battery life (mobile robots), and thermal management all affected.
Shock load
tolerance
High
(3 planets share load; DIN Cl.5 steel; peak 2–3× rated)
Low
(flexspline thin wall fatigue; peak 2–4× rated but cumulative)
Very low
(thinner flexspline = faster fatigue under shock)
 Planetary wins decisively. Agricultural equipment, construction robots, any application with mechanical shock: harmonic drive flexspline fatigue failure is a known failure mode. Planetary standard choice.
Operating
temperature
−40°C to +125°C
(NYOGEL 792D; sealed lifetime)
0°C to +70°C
(standard grease; flexspline steel below 0°C)
0°C to +60°C
(tighter tolerance on flexspline)
 Planetary wins for cold and high-temp. HD flexspline steel becomes brittle below 0°C; NYOGEL-sealed planetary operates to −40°C. Cold-climate robotics, outdoor equipment: planetary.
Axial length
(disc-form HD)
Standard gearbox length
(50–120mm depending on frame/ratio)
20–40mm disc
(cup-form available; flat disc form very short)
15–30mm disc
(ultra-flat — thinnest available)
 HD wins for ultra-flat requirements. Surgical robot final wrist joint, thin collaborative robot link: disc-form HD is significantly shorter. Note: EP-FADS saves 22mm vs FAD, narrowing (but not closing) the gap.
Back-driveability
(output→input)
Medium — can back-drive
(helical gears; back-drive efficiency ~60–80%)
Low — difficult to back-drive
(high friction in flex engagement)
Very low
(preloaded flexspline increases friction)
 Depends on application: collaborative robots that must be back-driveable for safe human interaction → planetary advantage. Gravity-holding applications → HD low back-drive can be useful (reduces servo braking power).
Efficiency at
low speed / low load
95–98%
(consistent across load range)
50–70%
(flexspline deformation loss is load-independent)
40–65%
(preload adds constant losses)
 Planetary advantage increases at partial load. HD flexspline deformation losses are nearly constant (not proportional to torque), so at light loads HD efficiency drops significantly. Battery-powered mobile robots: planetary preferred.
Ratio range
(single stage)
5:1 to 100:1
(standard; special: up to 91:1)
50:1 to 160:1
(below 50:1 impractical; most common 80:1–160:1)
50:1 to 160:1
(same range)
 Planetary covers low and medium ratios (5:1–50:1) that HD cannot. For high ratios (80:1–160:1), both work. HD is impractical below 50:1 due to geometry. Planetary covers the entire practical range.
Unit price
(relative)
100%
(EP-FAD P0 — benchmark)
150–250%
(plus flexspline replacement cost at life limit)
250–400%
(superprecision grade premium + replacement)
 Planetary wins significantly. Also: planetary has no scheduled consumable (flexspline). Total cost of ownership gap is wider than unit price comparison shows.

Summary:
Planetary P0 wins: 6 parameters
HD Ultra wins: 2 parameters (backlash, axial length)
Context-dependent: 2 parameters (back-driveability, ratio range)

Specifications shown are representative typical values for comparison purposes. Actual values depend on specific model, frame size, and ratio within each technology. Verify from manufacturer datasheets for your specific application. Harmonic drive life figures are for standard operating conditions at rated load — life varies significantly with duty cycle and torque level.

Engineering Fundamentals

The Flexspline Fatigue Mechanism — Why Harmonic Drives Wear Out 2–6× Faster Than Planetary

Planetary gearbox types comparison — EP-FAD round flange vs EP-FAB square flange vs harmonic drive flexspline wear mechanism

Design Life Comparison
EP-FAD P0 (planetary)
30,000 hr
HD standard (flexspline)
~15,000 hr
HD ultra-precision
~8,000 hr
L10 bearing fatigue (planetary) vs flexspline material fatigue (HD). Flexspline replacement is a scheduled maintenance item for all HD installations.

The Fundamental Physics of Flexspline Fatigue

A harmonic drive’s operating principle requires its flexspline to deform continuously. Every revolution of the wave generator deflects the flexspline from its natural circular form into an ellipse, engaging the circular spline at two diametrically opposite points, then back to circle, then to ellipse again in the other orientation. In one wave generator revolution, the flexspline wall completes two full bending cycles from natural to deflected to natural. At a motor speed of 3,000 rpm, the flexspline undergoes 6,000 bending cycles per minute — 360,000 cycles per hour — 10.8 billion cycles in 30,000 hours of operation.

Metal fatigue is a cyclic damage accumulation mechanism. A thin-walled steel flexspline operating at the deflection amplitude required for tooth engagement will accumulate fatigue damage according to its S-N curve (stress vs cycles to failure). The flexspline is designed with sufficient wall thickness and material properties to achieve typically 5,000–15,000 hours of L10 flexspline fatigue life — after which the probability of a fatigue crack propagating through the flexspline wall rises to 10% or higher. This is not a design failure; it is the fundamental physics of a component that must deform continuously to transmit torque.

A planetary gearbox has no continuously deflected component. The planet gear teeth engage the ring gear and sun gear in rolling contact — each tooth pair engages and disengages once per planet revolution, but the tooth material itself is never bent. Bearing rolling elements experience cyclic Hertzian contact stress, which is the basis for the L10 bearing life calculation — but this contact stress is distributed across multiple rolling elements simultaneously, and the fatigue life of precision bearing steel at the contact stress level of a DIN Class 5 gearbox produces the 30,000-hour L10 life that Korea Ever-Power rates for EP-FAD. The planetary’s life-limiting mechanism (rolling contact bearing fatigue) has a significantly longer L10 life than the harmonic drive’s life-limiting mechanism (flexspline bending fatigue) at the forces involved in equivalent-ratio precision gearboxes.

The practical consequence: Harmonic drive installations require a planned flexspline replacement at the manufacturer’s recommended life interval — typically every 5,000–12,000 hours depending on the specific HD model and operating conditions. This replacement involves disassembling the joint, removing the failed or pre-emptively replaced flexspline, installing a new one, and re-commissioning the joint. For a 6-axis industrial robot with HD at all joints, this means periodic rebuilds at each joint — a maintenance cost that does not exist for a robot built with planetary gearboxes at equivalent joints. Over a 10-year machine service life at typical industrial duty cycles, the cumulative flexspline replacement cost can equal or exceed the initial HD purchase premium.

⚡
Why 70–85% Efficiency Matters More Than It Seems
At 97% planetary efficiency, a 100W motor output delivers 97W to the load. At 80% harmonic drive efficiency, the same motor delivers only 80W to the load — with 20W converted to heat in the flexspline deformation. This 20W of heat must be dissipated from the joint housing. In a compact robot wrist, this creates a thermal management problem: the joint temperature rises, the lubricant degrades faster, and the flexspline fatigue life accelerates. For battery-operated mobile robots (AMRs, surgical robots, inspection drones), the 15–20% efficiency penalty translates directly to reduced run time per charge — a constraint that is increasingly significant as robot payloads and working hours increase. The efficiency comparison is not academic: at robot joint torques of 10–100 N·m and speeds of 50–500 rpm output, the difference between 97% and 80% efficiency is 1.5–15W of continuous heat generation per joint, across six joints, for the life of the machine.

The efficiency penalty has a second-order effect on flexspline life that compounds the first-order fatigue argument. A harmonic drive operating at 80% efficiency generates heat within the gear mesh. This heat raises the internal temperature of the gear assembly, which affects the flexspline steel’s fatigue properties: as temperature increases above the steel’s design temperature, the endurance limit decreases, and the fatigue life at a given stress level shortens. In a thermally constrained robot joint where the housing temperature is already elevated by motor losses, the additional 15–20W of flexspline deformation heat pushes the joint temperature higher, further accelerating flexspline fatigue. This thermal-fatigue coupling is one reason why harmonic drive manufacturers specify operating temperature ranges that are tighter than equivalent planetary gearboxes, and why flexspline life in hot environments can be significantly shorter than the rated figure.

The torsional compliance of the flexspline — which contributes to the high torsional stiffness characteristic of harmonic drives — also plays a role in the shock tolerance comparison. The flexspline acts as a torsional spring between the wave generator input and the circular spline output. Under normal operating loads, this spring effect is absorbed by the flexspline material’s elastic range and does not cause immediate damage. Under shock loads — abrupt stops, collisions, or dropped payloads — the spring absorbs energy and deflects beyond the normal operating range, into the plastic deformation zone of the flexspline steel. Each shock event that causes even minor plastic deformation in the flexspline wall accumulates fatigue damage faster than continuous cyclic loading at the rated stress level. This is why harmonic drives are specifically not recommended for applications where shock loads are part of the duty cycle — agricultural equipment, construction robots, human-collaborative robots that may contact obstacles. In contrast, a planetary gearbox under equivalent shock loading distributes the impact across three planets and the ring gear simultaneously, with each contact zone operating well within the elastic range of the DIN Class 5 alloy steel, producing no fatigue damage accumulation from the shock event itself.

Honest Comparison

When Harmonic Drive Is Genuinely the Better Choice — Four Application Scenarios

EP-FADS direct-insert — the closest planetary gearbox to harmonic drive axial compactness. For robot wrist joints where the full disc-form HD is not needed, EP-FADS P0 with its 22 mm shorter profile vs EP-FAD often closes the axial gap sufficiently to avoid the HD specification. Contact Korea Ever-Power with your available axial envelope to confirm whether FADS fits before defaulting to HD.

This comparison would not be complete or credible without clearly identifying the applications where harmonic drive is genuinely the better engineering choice. These cases exist, and they are important enough that many high-performance robots and precision instruments correctly use harmonic drives for specific joints while using planetary gearboxes elsewhere on the same machine.

Scenario 1 — Sub-Arc-Minute Positioning Without Servo Compensation

For applications where positioning accuracy below 0.3 arc-minutes is required and the servo controller does not implement backlash compensation — or where the application is so position-sensitive that even residual post-compensation error of 0.3–0.5 arc-min is unacceptable — harmonic drive ultra-precision is the correct choice. Examples: interferometric measurement stage rotation axis, optical telescope final pointing axis, E-beam lithography beam steering, wafer handler with ±0.01 mm TCP requirement. In these applications, the near-zero backlash of HD ultra-precision is genuinely essential — it is not achievable by planetary P0 regardless of servo tuning.

Note: Verify this requirement carefully. Many applications specified for HD because of “zero backlash” requirements can be met by EP-FAD P0 at 0.78 arc-min with proper servo compensation — the actual positioning error is below ±0.1 mm at standard robot arm radii. Test with a P0 planetary on a commissioning prototype before committing to HD on the production BOM.
Scenario 2 — Ultra-Flat Joint in a Constrained Axial Envelope

When the robot or instrument joint must fit within an axial envelope of 25–40mm, disc-form harmonic drives are the only gear reduction technology that can achieve the required ratio (80:1–160:1) within this constraint. Thin collaborative robot joint modules, minimally invasive surgical instrument wrist joints with trocar diameter constraints beyond what EP-FADS can address, and flat gantry wrist axes where mounting depth is architecturally fixed: disc-form HD is the correct specification. Before committing, check whether EP-FADS P0 fits — its direct-insert geometry saves 22 mm vs EP-FAD and may close the gap without the HD efficiency and life penalties.

Note: Many robot wrist applications that have historically used HD because of axial constraints can now use EP-FADS, which eliminates the adapter ring and produces a shorter total assembly. Confirm dimensions with Korea Ever-Power before making the HD/planetary decision for a constrained wrist joint.
Scenario 3 — Gravity-Holding Without Motor Braking

Harmonic drive’s low back-driveability (output cannot easily drive the input) is an advantage in applications where the joint must hold a gravity load without motor braking power applied — for example, a robot arm that must hold position at rest without drawing current, or a valve actuation mechanism that must remain in position during a power interrupt without a mechanical brake. The high friction in HD tooth engagement provides natural gravity-holding at no electrical cost. Planetary gearboxes can be back-driven with moderate force at ratios below 50:1, requiring motor braking for gravity hold. This can be addressed with a separate brake, but the HD simplifies the design for gravity-sensitive joints.

Scenario 4 — Very High Ratio in a Single Stage (100:1–160:1)

Harmonic drives achieve ratios of 100:1 to 160:1 in a single stage at the same axial depth and diameter as lower-ratio versions. Planetary gearboxes require compound staging (two gearbox stages in series) to achieve ratios above 100:1 in standard catalogue configurations, which doubles the axial length and increases cost. For applications requiring 120:1 or 160:1 in a compact package, HD is the more natural choice. Note that Korea Ever-Power’s EP-FALR provides 180:1 in a right-angle single unit, but inline single-stage 160:1 planetary is not practical — HD wins at ultra-high ratios in compact form factors.

Selection Framework

The Five-Question Selection Framework — Planetary P0 or Harmonic Drive for Each Joint

Apply these five questions sequentially to each joint. The framework produces the correct technology choice for the overwhelming majority of precision servo applications. For cases where the answers are genuinely ambiguous, a commissioning prototype with both technologies provides the empirical data to resolve the selection.

Five-Question Technology Selection
Q1
Does the application require backlash below 0.5 arc-minutes without servo compensation?
Yes → Harmonic Drive ultra-precision. Planetary P0 at 0.78 arc-min with compensation achieves approximately 0.3–0.4 arc-min effective — if that is sufficient for your accuracy requirement, continue to Q2. If truly below 0.3 arc-min is needed (interferometric stages, E-beam lithography): HD ultra-precision. No → Continue to Q2. Calculate the required backlash using the grade selection calculator (θ = δ/r × 3438) to confirm P0 planetary with compensation meets your accuracy budget.
Q2
Is the available axial envelope less than 50mm for the gearbox at the required ratio?
Yes → First check EP-FADS dimensions for the required frame size (EP-FADS saves 22mm vs EP-FAD). If EP-FADS still does not fit: Harmonic Drive disc-form. No → Planetary gearbox fits — continue to Q3.
Q3
Is the required ratio above 100:1 in a single stage without compound gearboxes?
Yes → Harmonic Drive (100:1–160:1 single stage) or EP-FALR 180:1 right-angle (if right-angle motor layout is acceptable). No → Planetary gearbox catalogue covers the required ratio — continue to Q4.
Q4
Is the application subject to shock loads, cold temperatures (below 0°C), or continuous high duty?
Shock loads: Planetary. HD flexspline fatigue failure under shock is a known risk; planetary distributes shock across three planets. Below 0°C: Planetary (EP-FAD NYOGEL 792D rated to −40°C; HD flexspline brittle below 0°C). Continuous high duty: Planetary (30,000 hr L10 vs HD 5,000–15,000 hr flexspline life). If any of these apply: Planetary. None apply → Continue to Q5.
Q5
Is efficiency important (battery power, heat dissipation, motor sizing)?
Yes: Planetary. 97–99% efficiency vs 70–85% HD — significant for battery-powered mobile robots, heat-constrained wrist joints, and motor sizing for continuous duty. No specific constraint: Either technology satisfies Q5. At this point, the primary remaining differentiator is backlash (Q1 — confirmed adequate for planetary) and price: planetary P0 at 100% vs HD at 150–300%. Default: Planetary P0 for cost, life, and efficiency unless Q1–Q4 conclusively indicated HD. At this point in the decision process, the engineer has confirmed that the application does not require sub-0.5 arc-min backlash, that the axial envelope accommodates planetary (potentially using EP-FADS), that the ratio is achievable in single-stage planetary, and that the duty/shock/temperature conditions favour planetary over HD. The remaining decision is purely economic: planetary P0 at 100% vs HD at 150–300%, with the ongoing flexspline replacement cost added to HD. The five-question framework applied honestly produces the correct technology for each joint — and for the majority of joints in industrial automation, collaborative robotics, and precision instruments, that technology is planetary P0.

Worked Selection Example: 6-Axis Collaborative Robot Arm

Joint Q1: Backlash Q2: Axial Q4: Duty/Shock Selected Series
J1 Base rotation P1 adequate at large radius Large housing, no constraint High duty; occasional shock Planetary EP-FAB P1 (high torque)
J2 Shoulder P0 with comp adequate Link housing, ample space S5 duty; moderate shock Planetary EP-FAD P0
J3 Elbow P0 adequate Compact link, check FADS S5 duty; low shock Planetary EP-FADS P0 (compact)
J4 Wrist pitch P0 with comp adequate Tight wrist — check FADS vs HD S5 duty; cobotic safety Planetary EP-FADS P0 if fits; HD if not
J5 Wrist yaw P0 with comp adequate Very tight — HD disc likely needed S5 duty; light shock HD HD disc-form (axial constraint)
J6 Tool rotation P0 with comp adequate for collab Ultra-tight — HD disc only S5 duty; tool change shock HD HD disc-form (axial + ratio)

This worked example shows a realistic 6-axis collaborative robot where J1–J4 use planetary gearboxes and J5–J6 use harmonic drives — a mixed architecture that optimises cost, life, and efficiency for each joint’s specific constraints. The assumption that “a collaborative robot uses harmonic drives everywhere” leads to unnecessary cost and maintenance burden at J1–J3 where the axial and backlash constraints that justify HD do not apply.

💰
The Grade Selection Calculator Applied to HD vs Planetary
Before specifying harmonic drive for any joint, apply Korea Ever-Power’s grade selection calculator (θ = δ/r × 3438) to confirm whether the application’s accuracy requirement actually needs HD. For a typical collaborative robot wrist joint with ±0.05 mm TCP specification and 400 mm reach: required backlash without compensation = (0.05/400) × 3438 = 0.43 arc-min; with 60% servo compensation, the gearbox can have up to 0.43/0.4 = 1.07 arc-min raw backlash. EP-FAD P0 at 0.78 arc-min satisfies this requirement with 0.29 arc-min margin — no HD required. Only when the calculation produces a required backlash of less than 0.3–0.4 arc-min (which occurs at small arm radii below 80mm combined with tight accuracy requirements) does HD’s near-zero backlash provide a genuine benefit over P0 planetary. Run this calculation for every joint before committing to HD — it takes three minutes and it frequently reveals that P0 planetary is sufficient where HD was assumed necessary.

Frequently Asked Questions — Planetary vs Harmonic Drive

Can EP-FAD P0 with servo compensation truly match harmonic drive backlash performance?＋
For the accuracy requirements of most industrial robot applications (±0.1 mm TCP repeatability), EP-FAD P0 at 0.78 arc-min measured with servo backlash compensation produces effective positional error of approximately 0.78 × 0.4 = 0.31 arc-min — which at a 400mm arm radius corresponds to approximately 0.036 mm. This is well within the ±0.1 mm typical robot TCP specification. For reference, ISO 9283 defines ±0.05 mm as “high accuracy” for industrial robots — which EP-FAD P0 with compensation achieves at arm radii above 120mm. The comparison is not about whether planetary with compensation “matches” HD in absolute backlash value — it does not, because HD achieves true near-zero backlash. The question is whether the effective positional error after compensation meets the application’s accuracy requirement, which for the majority of precision servo applications it does. For the minority of applications that require error below 0.05 arc-min equivalent regardless of compensation effectiveness, HD ultra-precision is genuinely required. A practical calibration exercise for engineers who are uncertain: commission a prototype joint with EP-FAD P0, measure the effective TCP positioning error over 100 consecutive moves in each direction, and compare the worst-case position error to the application specification. If the measured error is within specification with margin, the HD specification can be avoided. If the measured error exceeds specification even after servo compensation optimisation, HD is confirmed as necessary. This empirical commissioning test is more reliable than a theoretical calculation alone, because it incorporates the actual servo tuning, encoder resolution, bearing clearance, and thermal effects of the real installation — all of which affect the practical positioning error and none of which the theoretical calculation captures with full accuracy.
How significant is the efficiency difference in practice for a 6-axis robot?＋
For a typical 10kg payload 6-axis industrial robot operating at moderate speed: each joint motor delivers approximately 50–300W of mechanical power at peak. At 80% HD efficiency vs 98% planetary, the difference per joint is 10–60W of heat generation. Across 6 joints, this is 60–360W of additional heat that the robot arm must dissipate — in a compact arm structure where thermal mass is limited. This heat contributes to joint temperature rise, which accelerates lubricant degradation and flexspline fatigue. For battery-operated autonomous mobile robots with integrated arms, the efficiency penalty is even more significant: a robot arm consuming 120W more than necessary at 80% HD efficiency reduces run time by approximately 10–15% on a typical battery pack. For collaborative robots where motor thermal limits constrain maximum continuous force, the efficiency penalty directly reduces the achievable payload at thermal equilibrium. For fixed-installation industrial robots on AC power, the efficiency difference may be less operationally critical — though it still represents a meaningful increase in electricity consumption over the machine’s service life.
What is the total cost of ownership difference between HD and planetary over 10 years?＋
A representative calculation for a 6-axis industrial robot operating at 6,000 hours per year (typical 3-shift production): Planetary EP-FAD P0 initial cost: 6 × €300 = €1,800. No scheduled replacements at rated load in 10 years (60,000 hours; above 30,000 hr L10, but at typical 60% duty and partial load, calculated life exceeds 10 years). Harmonic Drive HD equivalent: 6 × €700 = €4,200 initial. Flexspline replacement at 10,000 hr intervals: 6 joints × 3 replacements × €250 per flexspline kit = €4,500 over 10 years. Total HD 10-year cost: approximately €8,700. Total planetary 10-year cost: approximately €1,800 plus minor maintenance. This example is illustrative and uses representative pricing — actual costs depend on specific robot model, joint size, HD model, and regional labour rates. The principle holds: the scheduled flexspline replacement cost accumulates to exceed the initial HD premium within the first replacement cycle, making the 10-year total cost of ownership for HD significantly higher than planetary despite the appearance of a moderate initial price difference. For robot OEMs building and selling machines under warranty, the TCO comparison extends to warranty claims: a robot with planetary gearboxes operating within their L10 rated life will generate near-zero gearbox warranty claims over a 2-year warranty period. A robot with harmonic drives at joints approaching their flexspline life limit may generate warranty claims for flexspline-related failures — soft stops, position drift, or catastrophic flexspline crack — that the OEM bears during the warranty period. This warranty exposure is a real procurement cost that does not appear in unit price comparisons but is visible in warranty reserves and field service budgets. OEMs who have transitioned from HD to planetary on high-volume robot production lines consistently report a reduction in gearbox-related warranty claims alongside the BOM cost reduction — making the total financial case for planetary even stronger than the unit price comparison alone suggests.
Does EP-FADS actually eliminate the axial advantage of harmonic drives for most wrist joints?＋
For many robot wrist applications — not all. EP-FADS saves the adapter ring and external clamp ring stack (typically 18–22mm) compared to EP-FAD, bringing the total motor-to-output length closer to what disc-form HD achieves. For wrist J4 axes where the available axial envelope is 60–80mm, EP-FADS often fits where EP-FAD does not, and the remaining gap to HD disc-form may be small enough that the engineer accepts the slightly larger planetary to gain the life, efficiency, and cost advantages. For wrist J5 and J6 axes where the envelope is 20–35mm, disc-form HD is typically the only viable option regardless of whether planetary is FAD or FADS — the physics of fitting a ratio-capable gear train into 20mm of axial depth requires the flexspline deformation principle that HD uses. The honest answer is: EP-FADS eliminates the HD axial advantage for a significant proportion of wrist J4 applications, but not for the most compact J5/J6 configurations. Send Korea Ever-Power your specific joint envelope dimensions to confirm whether FADS fits before defaulting to HD.
Are there any cycloidal or other technologies to consider alongside planetary and HD?＋
Cycloidal drives (also called RV reducers, commonly used in industrial robot J1–J3 base joints) and strain wave gears (another name for harmonic drives) are the main alternatives to planetary gearboxes for precision robot joints. Cycloidal drives offer very high torsional stiffness and shock tolerance at high ratios (typically 30:1–120:1), slightly lower backlash than helical planetary (1–3 arc-min class), and good life — but at significantly higher cost and weight than equivalent planetary gearboxes, and with a more complex internal mechanism. For robot base joints requiring extremely high torsional stiffness under heavy payload shock loads (automotive press tending robots, large palletising robots), cycloidal drives are commonly used at J1/J2 where their stiffness advantage over planetary justifies the cost premium. For the majority of industrial robot joints and precision automation axes in the 50–300 N·m output torque range, planetary P0 provides the best balance of stiffness, life, efficiency, cost, and availability — which is why the global installed base of precision servo automation uses planetary gearboxes at higher volumes than any other technology. A future article in this series will cover the planetary vs cycloidal vs harmonic comparison for robot base joints specifically. For engineers currently specifying base joints: cycloidal drives (RV reducers) achieve very high torsional stiffness and excellent shock tolerance at ratios of 30:1–120:1, making them the traditional choice for heavy industrial robot J1/J2 joints with payloads above 20kg. EP-FAB P0 (square-flange, high-stiffness) is competitive for J1/J2 joints in the 5–20kg payload class where the torque requirement is within the FAB rated range and the torsional stiffness of helical planetary is adequate for the application. For lighter collaborative robots (6kg and below), EP-FAD or EP-FAB P0 at J1/J2 is typically the correct specification — the cycloidal and HD cost and complexity premiums are not justified at these payload levels.

Confirm EP-FADS Fits Your Wrist Joint Before Specifying Harmonic Drive
Send your available axial envelope, required ratio, and output torque — Korea Ever-Power will confirm whether EP-FADS P0 fits within your joint geometry and provide dimensional drawings for direct comparison with disc-form harmonic drive. Most wrist J4 applications can use EP-FADS; J5 and J6 are evaluated case by case.

Compare EP-FADS vs Harmonic Drive for Your Joint →

Editor: Cxm

VR Tour of Our Factory

TAGs:

Recent Comments

No comments to show.