What’s the Difference Between Single and Dual Encoder Robot Joint Modules?

Should you choose a single or dual encoder robot joint module? The answer depends on your positioning requirements, transmission characteristics, force control strategy and budget.

A dual encoder system measures both the motor position and the actual joint output position. This additional feedback can improve control performance, but achieving those benefits requires suitable hardware, control algorithms and calibration.

This guide explains how dual encoder robot joint modules work, when a single encoder may be sufficient, and what dual feedback can contribute to precision, force estimation and functional safety.

What Is a Dual Encoder Robot Joint Module?

A dual encoder robot joint module typically includes two angular position sensors located on opposite sides of the transmission.

Motor-Side Encoder

The motor-side encoder measures the rotor angle at the motor shaft, on the high-speed side of the reducer.

Its feedback is commonly used for motor commutation and velocity control.

Output-Side Encoder

The output-side encoder measures the actual output angle at the reducer output shaft or joint output flange, on the low-speed side.

It provides direct feedback on the joint position rather than relying only on a calculation based on motor rotation.

Together, these encoders provide motor-side and load-side position information. With a compatible controller, they can support a dual-loop control architecture. Installing two encoders alone does not automatically establish dual-loop control.

What Are the Main Benefits of Dual Encoder Robot Joint Modules?

The main benefit is greater visibility into transmission behavior. By measuring the actual output position, the controller can reduce the effects of errors that a motor-side encoder cannot directly observe.

Common sources of transmission-related positioning error include:

Backlash and lost motion: During direction changes, motor movement may not immediately produce the expected output movement.

Elastic deformation: Reducers, belts, couplings and other transmission components can twist under load.

Transmission error: Manufacturing and assembly variations can cause deviations from ideal angular transmission, even when the nominal gear ratio is fixed.

Temperature and wear: Changes in operating conditions and component condition can affect transmission behavior.

With a single motor-side encoder, the output angle is estimated from motor position and the reduction ratio. For a reduction ratio defined as motor revolutions per output revolution:

Estimated output angle = Motor angle ÷ Reduction ratio

This calculation must also account for rotation direction and reference offsets. It does not directly measure load-side deflection or lost motion.

More Accurate Output Position Control

In a typical dual-loop arrangement, the output encoder supplies feedback to the outer position loop, while the motor encoder supports commutation and the inner velocity loop. Motor current feedback comes from current sensing, rather than from the encoder itself.

This arrangement allows the controller to respond to the measured output position and compensate for some transmission-related errors.

However, dual feedback does not eliminate every source of positioning error. Performance still depends on encoder accuracy, mounting quality, calibration, mechanical behavior and controller tuning.

It also cannot directly measure structural deflection occurring beyond the output encoder, such as bending in a robot link or end effector.

Compensation for Backlash and Elastic Deformation

After converting both encoder readings to the same reference frame, the controller can compare motor-derived output position with the measured output position.

The difference can help identify lost motion, elastic deflection and other transmission effects. Depending on the control strategy, this information can support feedback correction or model-based compensation.

The practical benefit is reduced positioning error. Mechanical backlash remains present, and reversal performance must still be verified under representative loads and speeds.

Torque Estimation in Suitable Joint Designs

In a joint with a known, calibrated elastic element, the angular deflection across that element can be used to estimate transmitted torque.

Dual encoder feedback can therefore support torque estimation in some compliant joints and series elastic actuators.

The estimate depends on the stiffness model, encoder accuracy, friction, hysteresis, temperature and dynamic behavior. An ordinary gearbox does not automatically become an accurate torque sensor simply because encoders are installed at both ends.

Where precise torque measurement is required, a dedicated joint torque sensor may still be appropriate.

Additional Diagnostic Information

Unexpected differences between motor-side and output-side measurements can indicate abnormal behavior, such as transmission slip, excessive deflection or certain sensor and mechanical faults.

These comparisons can support fault monitoring. They do not guarantee detection of every collision or failure, and safety-related use requires a validated diagnostic and response strategy.

Is a Dual Encoder Joint Module Always Better Than a Single Encoder Module?

No. Dual encoders provide additional feedback, but they also introduce costs and integration requirements.

The appropriate choice depends on whether direct output measurement provides a meaningful benefit for the application.

A single encoder robot joint module may be sufficient when the complete system meets its positioning, motion and safety requirements.

When Can a Single Encoder Be Sufficient?

1. Transmission errors remain within the application’s tolerance

A suitably selected precision reducer may provide adequate performance with motor-side feedback.

However, low backlash and high stiffness are different characteristics. A reducer can have very low backlash while still exhibiting significant torsional deflection under load.

Evaluate the actual reducer specifications and load conditions rather than assuming that every harmonic or RV reducer behaves the same way.

2. The required positioning performance is achievable without output feedback

Some handling, welding, spraying and repetitive automation tasks can be performed effectively with motor-side feedback.

The result must be verified on the complete robot. Encoder count alone cannot establish a particular end-effector accuracy, such as ±0.1 mm. Accuracy and repeatability must also be assessed separately.

3. Force control uses another measurement source

Some systems use an end-effector six-axis force/torque sensor to measure interaction forces while the joints operate primarily under position control.

In these applications, direct joint torque estimation from dual encoders may not be necessary.

4. Cost and integration simplicity are priorities

A second encoder can increase component cost, mechanical integration work, signal processing requirements and commissioning effort.

The actual cost difference depends on the encoder technology, module design and production volume. It should be evaluated from supplier quotations rather than a universal percentage.

5. Loads and operating conditions are predictable

If the load profile is stable and transmission errors remain acceptable throughout the operating cycle, the performance benefit of a second encoder may be limited.

Operating speed alone does not determine the answer. Dynamic behavior and control stability must also be considered.

Example: Force-Controlled Cleaning with a Mobile Manipulator

Consider a mobile manipulator that uses an end-effector six-axis force/torque sensor to regulate cleaning contact force.

The robot joints may operate in position control while an admittance controller adjusts the commanded motion using the measured contact force.

For this architecture, a single encoder joint module combined with a suitable reducer may be sufficient, provided that testing confirms acceptable tracking, contact-force regulation and stability.

If future requirements include hand-guided teaching or joint-space impedance control, additional feedback may be beneficial. Possible options include output encoders, dedicated joint torque sensors or other validated sensing and estimation approaches.

The choice should follow the intended control architecture and performance requirements.

Single vs Dual Encoder Robot Joint Modules: Comparison

How Can Dual Encoders Contribute to Robot Functional Safety?

Dual feedback can contribute to a safety architecture, but two encoders do not automatically provide Category 3, PL d, PL e or SIL 3 capability.

Position feedback for motion control and redundancy for functional safety are related but distinct design questions.

1. Channel Architecture and Common Cause Failures

Two sensors at different locations are not necessarily independent safety channels. They may share power supplies, electronics, communication paths or environmental vulnerabilities.

For the relevant ISO 13849-1 assessment, measures against common cause failure are evaluated using a points system. The commonly referenced threshold is at least 65 points out of 100, not “CCF ≤65%.”

Different mounting locations or communication protocols alone do not demonstrate compliance.

2. Integration with the Required Safety Functions

Safety-related feedback must be integrated into the complete sensing, logic and drive architecture.

Depending on the application, the required functions may involve safe stopping, speed limitation, position limitation or brake control.

Fault thresholds and response times must be derived from the system requirements. A fixed encoder disagreement limit, such as ±0.5°, should not be treated as a universal setting.

For collaborative applications using power and force limiting, acceptable contact conditions depend on factors including the affected body region, contact type, force and pressure. A single 150 N limit is not a universal criterion for safe human–robot contact.

3. Reliability and Diagnostic Coverage

A safety assessment considers architecture, component reliability, diagnostic coverage and common cause failures.

The probability of a dangerous failure per hour, or PFHD, cannot be calculated simply by placing two encoder failure-rate values in parallel.

For reference, the PFHD bands associated with PL d and PL e are:

PL d: 10⁻⁷ ≤ PFHD < 10⁻⁶ per hour.

PL e: 10⁻⁸ ≤ PFHD < 10⁻⁷ per hour.

Meeting a numerical band alone does not establish the performance level; the other applicable requirements must also be satisfied.

4. Fault Detection and a Suitable Safety Response

A properly designed system can use feedback disagreement to detect certain faults and initiate an appropriate response.

However, encoder count alone does not determine whether a system is fail-safe. A single physical encoder can contain independent measurement channels and diagnostics intended for safety-related applications, provided that it is integrated with a suitable safety controller.

Therefore, it is incorrect to assume that every single encoder system becomes uncontrolled after any encoder fault, or that every dual encoder system meets a particular safety integrity level.

Why Can Dual Encoder Feedback Be Useful in Humanoid Robot Joints?

Humanoid robots combine many joints with changing loads, balance requirements and frequent interaction with the environment.

These characteristics make accurate joint-state feedback valuable. However, the sensing architecture varies by robot design, and dual encoders should not be assumed to be universal.

Better Observation of Transmission Behavior

In a geared joint, motor movement and output movement may differ because of backlash, transmission error or elastic deformation.

A motor-side encoder measures only the motor position. An output encoder adds direct information about the actual joint angle.

This can improve the joint-position information available to the robot controller.

Measurement of Load-Dependent Deflection

When a robot transfers its weight, lands a foot or handles an external load, its joints and transmission components may deflect.

A deliberately compliant joint, such as a series elastic actuator, may exhibit additional relative motion across its elastic element.

Measuring both sides of the transmission can help observe this behavior. When combined with an appropriate model, the feedback may support control and estimation.

However, dual encoders alone do not ensure dynamic balance. Overall performance also depends on the robot’s mechanical design, inertial sensing, contact information, state estimation and control algorithms.

Absolute Output Position at Startup

An absolute output encoder can report joint position within its measurement range at power-up, reducing the need for a homing movement.

This capability comes from the encoder’s absolute measurement technology, not from having two encoders.

A single encoder can also be absolute, while a dual encoder arrangement may use incremental sensors.

Calibration, turn tracking where required, and startup checks remain necessary. Knowing the joint angle at power-up does not by itself guarantee that a humanoid robot can safely stand or begin moving.

How to Choose Between Single and Dual Encoder Joint Modules?

Begin with the requirements of the complete robot, then determine what feedback is needed at each joint.

Consider the following questions:

What output positioning accuracy and repeatability are required?

How much backlash, lost motion and torsional deflection occur under the expected load?

Will force control use motor-current estimation, an elastic-element model, a joint torque sensor or an end-effector force/torque sensor?

Does the controller support dual-loop operation and suitable tuning?

Is absolute output position required at startup?

What diagnostics and safety functions must the system provide?

Does the performance benefit justify the additional integration cost?

A single encoder can be an effective choice when the complete system meets its requirements with motor-side feedback.

A dual encoder robot joint module is worth considering when direct output measurement offers a clear advantage in positioning, transmission monitoring or compliant-joint control.

If you are selecting robot joint modules for a new project, discuss your positioning targets, load profile, force control strategy and feedback requirements with HONPINE.