Motor Sizing for Robotic Arm Joints: A Rigorous Engineering Guide

What Is This Calculation and Why It Matters

Motor sizing for a robotic arm joint is the systematic determination of the minimum torque and continuous power output required to reliably actuate a joint under worst-case dynamic and static loading conditions. Unlike generic motor selection, this process is mission-critical in robotics—where undersizing leads to stalling, thermal runaway, position error, or catastrophic joint failure; while oversizing introduces unnecessary mass, inertia, cost, and energy inefficiency that degrade system-level performance (e.g., reduced payload-to-weight ratio, slower acceleration, increased control complexity).

In collaborative robots (cobots), surgical manipulators, and high-precision industrial arms, joint-level motor sizing directly governs repeatability, bandwidth, safety compliance (e.g., ISO/TS 15066), and functional lifetime. A miscalculated motor may meet nominal specs on paper but fail during transient events—such as sudden payload shifts, inertial coupling from adjacent links, or frictional spikes during start-up—because it lacks sufficient torque headroom or thermal margin. Moreover, regulatory standards mandate traceable, verifiable sizing rationale—not just rule-of-thumb approximations.

This calculation bridges mechanical load analysis and electromechanical component selection. It transforms physical requirements (force, geometry, motion profile) into electrical specifications (torque, power, voltage, current) while respecting real-world constraints: transmission losses, thermal derating, duty cycle, and dynamic amplification factors. As such, it is not merely arithmetic—it is a foundational systems engineering activity anchoring the entire robot’s kinematic and dynamic integrity.

Theory and Formula Walkthrough

The core calculation comprises two interdependent equations derived from classical mechanics and energy conservation:

1. Required Torque: τ = F × r

• τ (tau): Required output torque at the motor shaft (N·m). This is the mechanical torque the motor must deliver after gear reduction (if present) to overcome the maximum static or quasi-static load at the joint.
• F: Maximum force applied perpendicular to the lever arm at the joint center (N). Critically, this is not the end-effector payload alone—it includes inertial forces (F_inertial = m·a), gravitational components (F_grav = m·g·sin(θ) projected radially), and worst-case friction or external disturbance forces. For rotary joints, F represents the tangential equivalent of the net radial load acting at radius r.
• r: Effective moment arm—the perpendicular distance from the joint axis to the line of action of F (m). In a simplified model, this is often the link length or gear pitch radius. However, per ISO 9283 §4.2, “the effective radius shall be determined from the actual force application geometry, including transmission elements (e.g., harmonic drive output flange, belt pulley radius), not nominal link dimensions.” Misidentifying r is among the top causes of 20–30% torque errors.

⚠️ Note: This formula assumes static equilibrium and pure rotational motion. For dynamic cases with angular acceleration α, the full equation is τ = I·α + F·r, where I is the reflected inertia. The calculator provided focuses on the dominant load torque term (F·r)—but engineers must verify that acceleration torque (I·α) does not exceed 15–20% of F·r or add it explicitly if significant.

2. Required Power: P = τ · ω / η

• P: Electrical input power required (W). This is the input power the motor must draw from the drive, not mechanical output power.
• τ: Same torque as above (N·m).
• ω: Maximum continuous angular velocity (rad/s) at which the joint must operate while delivering τ. Do not confuse with peak or short-duration speed—ISO 9283 §4.2 requires power rating verification at “rated operating speed under sustained load conditions.”
• η: Overall system efficiency (dimensionless, expressed as decimal). Here, η = efficiency / 100. This factor accounts for losses across all stages: motor copper/iron losses, gearbox friction (typically 70–95% per stage), coupling hysteresis, and encoder feedback overhead. Using η = 0.85 (85%) is reasonable for a well-designed planetary gearmotor—but harmonic drives may drop to η = 0.65–0.75 under high torque, demanding recalibration.

Why divide by η? Because the motor must draw more electrical power than the mechanical power it delivers: P_electrical = P_mechanical / η. Neglecting η yields optimistic (and unsafe) power estimates—commonly causing drive overcurrent trips or thermal shutdown.

Standard Requirements

Compliance is non-negotiable in certified robotic systems. Two key standards govern this calculation:

IEC 60034-1:2017 Rotating electrical machines — Part 1: Rating and performance

• Clause 5.1 defines “rated output” as “the mechanical output power which the machine can supply continuously at its terminals when operating under specified conditions.” Crucially, it mandates that rating must account for all loss mechanisms and be validated at the intended ambient temperature and cooling method. For robotic joints, this means:
  • Motor must sustain P (calculated) at 40°C ambient with its specified cooling (e.g., convection only, no forced air).
  • Derating applies if ambient exceeds 40°C or if duty cycle exceeds continuous (S1) rating.
• The standard prohibits using “peak” or “intermittent” ratings for sizing unless explicitly validated for the robot’s thermal time constant and duty cycle (e.g., < 30 s ON, > 300 s OFF).

ISO 9283:2016 Manipulating industrial robots — Performance criteria and related test methods

• Clause 4.2 specifies test methodology for “positioning accuracy and repeatability,” directly tying motor capability to performance claims. It states: “The robot shall be tested at rated payload, maximum speed, and maximum acceleration simultaneously.” Therefore, motor sizing must ensure torque and power margins exist at the intersection of these three parameters—not just at isolated maxima.
• Furthermore, ISO 9283 requires reporting of “maximum permissible torque at rated speed” and “thermal time constant”—both of which must be cross-checked against the calculated τ and P.

Non-compliance risks type certification failure, voided warranties, and liability exposure—especially in medical or aerospace applications governed by FDA 21 CFR Part 820 or DO-178C.

Common Mistakes and How to Avoid Them

1. Confusing Force with Payload Mass

Mistake: Using F = m·g (9.81 N/kg) for all orientations—even when the joint is horizontal and gravity produces zero torque about its axis. Fix: Resolve forces vectorially. For a shoulder joint lifting vertically, F_grav = m·g; for an elbow joint rotating horizontally, F_grav ≈ 0 but F_inertial = m·α·r_link dominates. Always compute the component perpendicular to r.

2. Ignoring Reflected Inertia During Acceleration

Mistake: Assuming τ = F·r suffices, neglecting τ_acc = J_total·α, where J_total includes motor rotor inertia, gear ratio squared, and link inertia referred to the motor shaft. Fix: Calculate J_total using parallel-axis theorem and gear ratio i: J_reflected = J_link / i² + J_motor. Then verify τ_acc ≤ 0.2·τ_load. If exceeded, use full dynamic equation or select higher-torque motor.

3. Applying Efficiency Incorrectly

Mistake: Using η as a single fixed value without considering its torque-speed dependence. Efficiency peaks near 75% of rated torque and collapses near stall. Fix: Consult manufacturer’s efficiency map (not datasheet “typical” values). For conservative sizing, use η at 75% of calculated τ—or apply 10–15% derating to P.

4. Omitting Safety Factor and Thermal Margin

Mistake: Sizing exactly to calculated τ and P with no margin. Fix: Apply minimum safety factors per application:

• Industrial arms: 1.5× torque, 1.3× power
• Cobots (human interaction): 2.0× torque (for ISO/TS 15066 contact force limits)
• Surgical robots: 2.5× torque (FDA guidance) Then validate motor thermal rise using manufacturer’s t_thermal and your duty cycle.

5. Overlooking Transmission Backlash and Compliance

Mistake: Assuming ideal rigid-body transmission; ignoring how gear backlash or shaft torsion distorts torque delivery and induces control instability. Fix: Add 10–20% torque margin for systems with >0.1° backlash or low-stiffness couplings. Use finite element analysis (FEA) to estimate torsional compliance k_t and verify closed-loop bandwidth remains >5× disturbance frequency.

Worked Example with Realistic Numbers

Scenario: A 6-DOF collaborative robot’s elbow joint (Link 3) must lift a 5 kg payload at full reach (0.4 m from shoulder), while rotating at up to 2.5 rad/s (≈143°/s). Worst-case orientation places payload directly below elbow—maximizing gravitational torque. Gearmotor uses a 100:1 planetary gearbox (η_gear = 0.92) and a brushless DC motor (η_motor = 0.88). Ambient: 35°C.

Step 1: Determine Maximum Force at Joint

• Gravitational force on payload: F_grav = 5 kg × 9.81 m/s² = 49.05 N
• Geometry: At elbow joint, radius r = distance from joint axis to payload CG projected perpendicular to axis. With full extension and vertical payload, r = 0.4 m (link length).
• But: Link 3 itself has mass (2.1 kg) and CG at 0.18 m. Its gravitational torque adds 2.1 × 9.81 × 0.18 = 3.71 N·m.
• Total static torque: τ_static = (49.05 × 0.4) + 3.71 = 19.62 + 3.71 = 23.33 N·m
• Add 25% safety factor (cobots): τ_required = 23.33 × 1.25 = 29.16 N·m

Step 2: Verify Dynamic Contribution

• Max angular acceleration: α = 10 rad/s² (from motion profile)
• Reflected inertia: J_link3 = 0.012 kg·m², J_motor = 0.0008 kg·m², i = 100 → J_reflected = 0.012 / 100² + 0.0008 = 0.0008012 kg·m²
• τ_acc = J_reflected × α = 0.0008012 × 10 = 0.008 N·m (negligible vs. 29.16 N·m → acceptable)

Step 3: Compute Power

• Overall efficiency: η = η_motor × η_gear = 0.88 × 0.92 = 0.8096
• P = τ × ω / η = 29.16 × 2.5 / 0.8096 = 89.9 W
• Add 15% margin for controller losses and thermal derating: P_final = 89.9 × 1.15 = 103.4 W

Step 4: Validate Against Standards

• Per IEC 60034-1 §5.1: Select motor rated ≥103.4 W continuous at 40°C ambient. A 120 W motor meets this.
• Per ISO 9283 §4.2: Confirm motor delivers ≥29.16 N·m at 2.5 rad/s on its published speed-torque curve—not just at stall or no-load.
• Cross-check thermal time constant: If motor t_thermal = 120 s, and duty cycle is 60 s ON / 180 s OFF, temperature rise stays within limits.

Final Selection: A 120 W, 30 N·m continuous, 100:1 planetary BLDC motor with IP65 enclosure and integrated absolute encoder—validated against manufacturer’s thermal and torque maps at 35°C ambient.

This example illustrates why “plug-and-play” calculators are starting points—not endpoints. Every variable demands physical justification, every margin requires risk-based justification, and every number must trace to a standard clause or test protocol.