Proportional, Servo and Closed-Loop Motion Control

Introduction and Context

Conventional hydraulic control often uses directional valves that are either fully open or fully closed. Such systems are effective for basic extension, retraction, clamping and lifting, but they provide limited control of position, velocity, acceleration and force. Precision machinery—such as machine tools, injection-moulding machines, test rigs, robotics and forming equipment—requires the actuator to follow a commanded motion accurately despite changing loads and disturbances.

Proportional and servo valves provide continuously variable control of hydraulic flow. When these valves are combined with sensors, electronic controllers and suitable feedback algorithms, they form closed-loop electrohydraulic motion-control systems. The essential engineering problem is to regulate a physical variable—usually position, velocity or force—so that its measured value tracks a reference command with acceptable accuracy, speed and stability.

The distinction between proportional and servo systems is not absolute. Both may use electromagnetic actuation and feedback. In general, proportional valves offer robust, economical variable control with moderate bandwidth, whereas servo valves provide finer control, faster response and greater sensitivity to contamination and installation quality.

Learning Objectives

By the end of this chapter, the reader should be able to:

1. Explain the operating principles of proportional and servo valves.
2. Distinguish open-loop, closed-loop and cascade motion-control architectures.
3. Select and evaluate position, velocity, pressure and force sensors.
4. Relate valve flow, actuator motion and load dynamics.
5. Interpret rise time, settling time, overshoot, bandwidth and steady-state error.
6. Explain the sources of instability and methods of controller tuning.
7. Analyse a practical electrohydraulic precision-motion application.
8. Identify safety, maintenance, environmental and employability implications.

1. The Motion-Control Chain

An electrohydraulic axis normally contains the following functional elements:

• Command generator: produces the desired position, velocity or force signal.
• Controller: compares the command with measured output and calculates a valve command.
• Valve amplifier or drive: converts the low-power controller signal into current for the valve solenoid or torque motor.
• Proportional or servo valve: meters hydraulic flow and establishes actuator direction and speed.
• Actuator: converts fluid power into linear or rotary mechanical motion.
• Mechanical load: includes mass, friction, stiffness, inertia and external forces.
• Sensor: measures the controlled variable.
• Feedback path: returns the measurement to the controller.

A useful conceptual relationship is that valve flow determines actuator velocity, while actuator displacement is the time integral of velocity. For an ideal linear hydraulic cylinder with piston area A:

Here Q is volumetric flow rate in m³/s, A is effective piston area in m² and v is piston velocity in m/s. This relation assumes incompressible flow and neglects leakage. Therefore:

For a double-acting cylinder, the effective area differs between extension and retraction because the rod occupies part of the return-side area. Consequently, equal valve flow commands do not generally produce equal extension and retraction speeds.

The force generated by a cylinder is approximately:

where F is force in newtons, p is pressure difference across the piston in pascals and A is effective area in m². Friction, seal forces, acceleration and back pressure must be included in a more realistic force balance.

2. Proportional Valves

A proportional valve changes spool position approximately in proportion to an electrical input, commonly a current signal. A typical command may be 0–10 V, ±10 V or 4–20 mA, although the valve amplifier generally regulates coil current because coil resistance changes with temperature.

A proportional directional valve can control both flow magnitude and direction. A proportional pressure valve regulates pressure according to command, within the limits of supply pressure, flow capacity and compensation design. A proportional flow valve may include pressure compensation so that flow is less affected by load-pressure changes.

2.1 Operating characteristics

Important characteristics include:

• Rated flow: flow at a specified pressure drop.
• Null position: spool position at which the commanded output is nominally zero.
• Overlap: the amount by which lands cover ports near neutral.
• Hysteresis: difference in output for increasing and decreasing command signals.
• Deadband: input range producing little or no output.
• Repeatability: ability to return to the same output under identical conditions.
• Frequency response: ability to follow changing commands.
• Dither: a small high-frequency signal used to reduce static friction and improve repeatability.

A valve with positive overlap tends to reduce leakage but creates a deadband. A valve with underlap responds rapidly around null but may leak internally. The correct choice depends on efficiency, controllability and safety requirements.

2.2 Advantages and limitations

Proportional valves are relatively economical, rugged and suitable for many industrial applications. They can control acceleration and deceleration more smoothly than on/off valves. However, spool friction, hysteresis, nonlinear flow characteristics, oil temperature and supply-pressure variation limit precision. Without feedback, a proportional valve system cannot automatically correct position errors caused by load changes.

3. Servo Valves

A hydraulic servo valve is a high-performance flow-control device designed to respond accurately to small command changes. A common two-stage design uses a torque motor and pilot stage to move a main spool. The pilot stage amplifies the small force produced by the torque motor, allowing the main spool to meter substantial flow.

Servo valves commonly have a low or zero-lapped spool, giving high sensitivity near null. They offer low hysteresis and high dynamic response, but they require very clean fluid. Small clearances in the pilot stage can be blocked by contamination. Their installation may require high-quality filtration, flushing, correct alignment and protection from electrical noise.

3.1 Proportional valve versus servo valve

The boundary between the two categories is technological rather than strictly definitional. Modern digital proportional valves may approach servo-valve performance, while some servo systems use robust electronics and diagnostics similar to proportional systems.

4. Feedback Sensors and Measurement Quality

Closed-loop performance is limited by the quality of measurement. A sensor must be selected according to range, resolution, accuracy, linearity, dynamic response, environmental resistance and failure behaviour.

4.1 Position sensors

Common position sensors include:

• Linear potentiometers: simple and inexpensive, but subject to wear.
• LVDTs: contactless differential transformers with good repeatability and long life.
• Magnetostrictive sensors: often integrated into hydraulic cylinders and capable of measuring piston position without external linkage.
• Linear encoders: high resolution, but require careful alignment and protection.
• Rotary encoders: measure motor or screw rotation; mechanical transmission backlash must be considered.

4.2 Velocity and acceleration measurement

Velocity may be obtained directly using a tachometer or calculated by differentiating position. Differentiation amplifies high-frequency noise, so filtering is normally required. Accelerometers are useful in vibration and dynamic testing but are less commonly the primary sensor for ordinary hydraulic positioning.

4.3 Pressure, force and load measurement

Pressure transducers measure pressure at one or more actuator ports. Force may be inferred from pressure difference, but this estimate is affected by seal friction, acceleration and unequal areas. A load cell or force transducer provides more direct measurement where force accuracy is important.

4.4 Sensor errors

If the true position is x and the measured position is xₘ, measurement error may be represented as:

Sensor resolution, bias, drift, hysteresis, noise and calibration error contribute to eₘ. A controller cannot correct an error that is hidden by sensor limitations. Excessive filtering reduces noise but introduces phase lag, which can reduce stability margins.

5. Open-Loop and Closed-Loop Architectures

In an open-loop system, the controller sends a valve command without measuring the resulting motion. It is simple but sensitive to load variation, leakage, friction and supply pressure.

In a closed-loop system, the measured output is compared with the reference command. The difference is the tracking error:

where r(t) is the reference, y(t) is the measured output and e(t) is the error. A controller uses e(t) to calculate the actuator command.

Figure: Basic closed-loop electrohydraulic motion-control architecture.

5.1 Cascade control

High-performance axes often use nested loops. The outer position loop generates a velocity command; the velocity loop generates a flow or valve command; an inner current loop regulates solenoid or torque-motor current. Each inner loop should normally be substantially faster than the loop surrounding it. Cascade control improves disturbance rejection and reduces the effect of valve nonlinearity.

A force-control architecture may use force as the outer loop and position or velocity as an inner loop. The choice depends on whether the process requires accurate location, controlled contact force or both.

6. Dynamic Response and Mathematical Models

Hydraulic motion control is dynamic because fluid compressibility, valve spool motion, fluid inertia, mechanical mass and damping create energy storage and delay.

For a simplified cylinder-load model, neglecting leakage and friction, the continuity relationship is:

Here Q is flow into the control volume, A is effective actuator area, v is actuator velocity, Cₜ is total leakage coefficient in m³/(s·Pa), p is pressure difference, Vₜ is effective fluid volume in m³ and βₑ is effective bulk modulus in Pa. The equation assumes small perturbations about an operating point and approximately constant parameters. The final term represents fluid compressibility.

The mechanical force balance can be written as:

where M is equivalent moving mass in kg, B is viscous damping in N·s/m and Fᴸ is external load force, including friction when represented as a disturbance. The model is simplified; real systems may include nonlinear friction, hose elasticity, valve dynamics and load stiffness.

6.1 Important response measures

• Rise time: time for the response to move between specified percentages, commonly 10% and 90%, of final value.
• Peak time: time to reach the first maximum response.
• Overshoot: amount by which the response exceeds its final value, expressed as a percentage.
• Settling time: time required for the response to remain within a specified tolerance band.
• Steady-state error: long-term difference between reference and output.
• Bandwidth: frequency range over which the system follows input without unacceptable attenuation or phase lag.

Increasing gain can reduce tracking error and increase response speed, but excessive gain may produce oscillation. Delays caused by sensors, digital sampling, filters and valve dynamics reduce the available stability margin.

7. Controllers and Stability

A proportional controller is:

where u(t) is controller output and Kₚ is proportional gain. Proportional control is simple and responsive, but a nonzero steady-state error may remain under load.

A proportional–integral–derivative controller is:

Kᵢ is integral gain and Kᴅ is derivative gain. The integral term removes constant steady-state error in an ideal model; however, it can accumulate during saturation, causing integrator windup. Anti-windup logic is therefore essential. The derivative term can improve damping but amplifies measurement noise, so practical systems often use a filtered derivative or omit it in favour of velocity feedback.

For a linear time-invariant feedback system with forward transfer function G(s) and feedback transfer function H(s), the closed-loop transfer function is:

Here s is the Laplace variable. This expression applies to a linear model with consistent signal definitions and negative feedback. Stability depends on the roots of the characteristic equation:

All characteristic roots must lie in the left half of the complex s-plane for continuous-time asymptotic stability. Nonlinearities such as saturation, deadband, Coulomb friction and backlash mean that linear stability analysis is an approximation, usually valid near a selected operating point.

7.1 Practical stability assessment

Engineers assess stability using step tests, frequency-response measurements, root-locus reasoning, Bode plots or margins calculated from a validated model. A stable response should settle without sustained oscillation. Excessive oscillation may result from too much gain, excessive integral action, sensor delay, poorly tuned filters, trapped air, mechanical looseness or a resonance in the hydraulic circuit.

Tuning should proceed conservatively:

1. Verify mechanical integrity, fluid cleanliness, pressure settings and sensor polarity.
2. Confirm valve command limits and emergency-stop behaviour.
3. Establish inner current and velocity loops before the outer position loop.
4. Begin with low proportional gain.
5. Increase gain until the response is sufficiently fast but remains well damped.
6. Add integral action only if steady-state error requires correction.
7. Use feedforward for known velocity or acceleration demands where appropriate.
8. Test across the full load, speed, temperature and pressure range.

8. Accuracy, Nonlinearity and Compensation

Control accuracy is not determined by controller resolution alone. It depends on sensor accuracy, valve hysteresis, fluid temperature, leakage, friction, structural compliance, quantisation, electrical noise and calibration.

Feedforward can improve tracking by supplying an estimated command based on desired velocity or acceleration, while feedback corrects residual error. However, feedforward cannot replace feedback because the plant model is never exact.

Friction compensation is useful at low speed, where static and Coulomb friction may cause stick–slip motion. Dither can reduce stiction, but excessive dither increases noise, wear and heat. Notch filters may reduce the effect of a known structural resonance, although filtering should not conceal an unsafe mechanical defect.

Digital controllers introduce sampling and computation delay. The sampling frequency should be high relative to the highest significant closed-loop frequency, and aliasing must be prevented with appropriate analogue or digital filtering. Encoder counts, analogue-to-digital resolution and numerical scaling must be checked against the required position tolerance.

9. Worked Example: Cylinder Speed and Position Error

A hydraulic cylinder has an effective extension area of 25 cm² and receives 15 L/min during extension. Estimate the ideal extension velocity. If the measured position is 0.8 mm below the reference, determine the position error using the convention e = r − y.

Solution. Convert units:

• A = 25 cm² = 0.0025 m²
• Q = 15 L/min = 0.00025 m³/s

Using v = Q/A:

Thus the ideal extension speed is 0.10 m/s, or 100 mm/s. This ignores leakage, compressibility, valve pressure drop and acceleration.

If the reference is 100.0 mm and the measured position is 99.2 mm:

The positive error indicates that the actuator must move farther in the positive direction. The calculation illustrates why feedback is necessary: the commanded flow predicts motion, but only the sensor confirms actual position.

10. Applied Case: Precision Hydraulic Slide

Background

The following is an illustrative industrial case. A hydraulic slide in a materials-testing machine must move a specimen to a commanded position and apply a controlled compressive force. The slide has a 500 mm stroke and must achieve ±0.05 mm positioning accuracy at moderate speed.

Problem

An open-loop proportional valve system produced different final positions under different specimen stiffnesses. At low speed, the slide also displayed stick–slip motion. Pressure fluctuations and changing oil temperature altered the result.

Analysis

The position error originated from load-dependent leakage, seal friction and fluid compressibility. A flow command alone could not compensate for these effects. Force could not reliably be inferred from supply pressure because return pressure and friction varied.

Approach and solution

The machine was redesigned with:

• a magnetostrictive cylinder position sensor;
• a calibrated load cell in the force path;
• a pressure transducer for diagnostic monitoring;
• a pressure-compensated proportional valve;
• an inner valve-current loop;
• a velocity loop nested inside the position loop;
• a force loop used during specimen contact;
• anti-windup limits and ramped command profiles;
• high-quality filtration and contamination monitoring.

During free travel, position control generated the velocity command. During contact, a controlled transition transferred authority to the force loop, preventing a sudden force spike. A modest dither signal reduced low-speed stiction, while software limits prevented motion beyond the safe stroke.

Outcome

The closed-loop architecture substantially reduced load-dependent position error and produced repeatable force application. The improvement did not result from the valve alone; it depended on sensor calibration, mechanical stiffness, suitable loop hierarchy, conservative tuning and disciplined maintenance.

Learning points

1. The controlled variable must match the process requirement.
2. A pressure sensor is not automatically a force sensor.
3. Nested loops should be tuned from the inside outward.
4. Mechanical and fluid defects cannot be solved safely by increasing controller gain.
5. Precision is a system property involving valve, actuator, sensor, controller and structure.

11. Safety, Maintenance and Professional Practice

A closed-loop controller can command motion rapidly and can mask developing faults if monitoring is inadequate. Safety design should therefore include independent pressure relief, guarded motion, emergency-stop functions, safe valve states, stroke limits and fault detection. A software limit is not a substitute for a physical safeguard where hazardous motion is possible.

Important diagnostic checks include sensor plausibility, valve command versus actual motion, abnormal following error, excessive pressure, filter differential pressure and unexpected drift at zero command. A redundant sensor or independent limit switch may be required for high-risk machinery.

Hydraulic cleanliness is particularly important for servo systems. Fluid sampling, correct filter selection, flushing procedures, leak prevention and contamination-control records support reliability. Maintenance personnel should never adjust gains, pressure settings or safety limits without documented authorization and a controlled test procedure.

Environmental considerations include energy losses from throttling, heat generation, oil leakage and disposal of contaminated fluid. Variable-displacement pumps, load-sensing supply systems and energy recovery can reduce consumption. Engineers should consider lifecycle impact rather than only initial valve cost.

12. Contemporary Developments

Modern electrohydraulic systems increasingly use digital valve electronics, fieldbus communication, onboard diagnostics and model-based commissioning. High-resolution sensors and programmable controllers support automatic calibration and condition monitoring. Data-driven methods can identify valve hysteresis, leakage or abnormal friction from trends in command, pressure and position.

Artificial intelligence can assist anomaly detection and predictive maintenance, but it should not replace deterministic safety functions. Training data may not represent rare hazardous failures, and an opaque model may be difficult to validate. Safety-critical control should retain verified limits, fail-safe states and human-understandable diagnostic evidence.

Challenges and Limitations

The principal limitations of electrohydraulic motion control are nonlinear friction, compressibility, temperature dependence, valve deadband, contamination sensitivity, sensor noise, structural resonance and load variation. A mathematically accurate model at one operating point may be inaccurate at another. High bandwidth also increases sensitivity to noise and mechanical flexibility. Consequently, design requires a balance among speed, accuracy, energy efficiency, robustness, cost and maintainability.

Chapter Summary

Proportional and servo valves convert electrical commands into continuously variable hydraulic flow or pressure. Proportional valves provide economical and robust control, while servo valves generally offer superior sensitivity and bandwidth at the cost of greater contamination and maintenance requirements. Motion accuracy depends on the complete control chain: command, controller, valve, actuator, load and sensor.

Closed-loop control compares a reference with measured output and corrects disturbances. Position, velocity, force and pressure sensors each serve different purposes and introduce resolution, noise and delay considerations. Hydraulic compressibility, leakage, inertia and friction determine dynamic behaviour. Rise time, overshoot, settling time, bandwidth and steady-state error provide useful performance measures.

Controller gain must be selected with regard to stability. PID control, cascade loops, feedforward, friction compensation, filtering and anti-windup can improve performance when applied within a validated architecture. Safe and precise operation also requires clean fluid, calibrated sensors, independent safeguards, disciplined tuning and attention to environmental impact.

Key Terms

Proportional valve: A valve whose output is continuously varied in relation to an electrical command.

Servo valve: A high-performance valve designed for precise, rapid flow control.

Closed-loop control: Control based on measurement of the controlled output.

Open-loop control: Control without feedback measurement of the output.

Feedback: Return of measured output information to the controller.

Following error: Difference between commanded and measured motion.

Hysteresis: Difference in output for increasing and decreasing input at the same command value.

Deadband: Input range producing little or no effective output.

Dither: Small oscillatory signal used to reduce static friction effects.

Bandwidth: Frequency range over which a system follows input adequately.

Overshoot: Amount by which a transient response exceeds its final value.

Settling time: Time required to remain within a specified error band.

Steady-state error: Persistent error after transient effects have decayed.

Cascade control: Nested control loops arranged with faster inner loops.

Bulk modulus: Measure of a fluid’s resistance to compression.

Stiction: Static friction that impedes initial motion.

Anti-windup: Method preventing excessive integral accumulation during saturation.

Feedforward: Control action based on predicted plant requirements.

Following-error limit: A safety or diagnostic threshold for excessive tracking error.

Review Questions

5 Short Answer Questions

1. What is the primary functional difference between an open-loop and a closed-loop hydraulic motion system?
2. Why is valve current often regulated instead of merely applying a fixed voltage to a solenoid?
3. Define hysteresis and deadband in a proportional valve.
4. What role does a position sensor play in a closed-loop axis?
5. Why are servo valves more sensitive to contamination than many proportional valves?

5 Long Answer Questions

1. Compare proportional and servo valves in terms of construction, performance, cost, contamination tolerance and application.
2. Explain the operation of a nested position–velocity–current control architecture.
3. Discuss how fluid compressibility, leakage and mechanical inertia influence electrohydraulic dynamic response.
4. Explain the meaning and engineering significance of rise time, overshoot, settling time, bandwidth and steady-state error.
5. Describe a safe and systematic procedure for tuning a PID-controlled hydraulic axis.

3 Analytical / Application Questions

1. A cylinder has an effective area of 0.003 m² and receives 18 L/min. Calculate its ideal speed in m/s, state the assumptions, and explain why the actual speed may differ.
2. A positioning axis oscillates after proportional gain is increased. Identify at least four possible hydraulic, mechanical, sensing or control causes and propose a diagnostic sequence.
3. Design a control architecture for a hydraulic press that requires rapid approach, accurate final position and controlled pressing force. Justify the sensors, loop hierarchy, transition strategy and safety provisions.

Multiple Choice Questions with Answers

Question 1

Which variable is directly related to ideal cylinder velocity by v = Q/A?

A. Pressure divided by area

B. Flow rate divided by effective area

C. Area divided by flow rate

D. Force multiplied by flow rate

Correct Answer: B. Flow rate divided by effective area

Question 2

What is a principal advantage of closed-loop control?

A. It eliminates the need for sensors

B. It automatically corrects measured output errors

C. It guarantees stability at any gain

D. It prevents all hydraulic leakage

Correct Answer: B. It automatically corrects measured output errors

Question 3

Which sensor is commonly used for direct measurement of hydraulic-cylinder piston position?

A. Magnetostrictive sensor

B. Relief valve

C. Flow restrictor

D. Accumulator

Correct Answer: A. Magnetostrictive sensor

Question 4

What is a likely consequence of excessive integral action?

A. Integrator windup and oscillatory response

B. Elimination of all sensor noise

C. Reduced fluid compressibility

D. Permanent removal of valve deadband

Correct Answer: A. Integrator windup and oscillatory response

Question 5

For negative feedback, the characteristic equation is:

A. G(s) − H(s) = 0

B. G(s)H(s) = 1

C. 1 + G(s)H(s) = 0

D. G(s) + H(s) = 1

Correct Answer: C. 1 + G(s)H(s) = 0

Question 6

A cylinder receives 12 L/min and has an effective area of 0.002 m². What is its ideal velocity?

A. 0.01 m/s

B. 0.06 m/s

C. 0.10 m/s

D. 6.0 m/s

Correct Answer: B. 0.06 m/s

Question 7

Why is a force transducer often preferred to estimating force from pressure alone?

A. Pressure cannot be measured hydraulically

B. Pressure-based force estimates are affected by friction and other forces

C. Force transducers increase valve bandwidth automatically

D. Pressure is independent of actuator area

Correct Answer: B. Pressure-based force estimates are affected by friction and other forces

Question 8

Which characteristic generally distinguishes a servo valve from a basic proportional valve?

A. Lower sensitivity and slower response

B. Higher sensitivity and dynamic response

C. Complete immunity to contamination

D. Operation without hydraulic pressure

Correct Answer: B. Higher sensitivity and dynamic response

Question 9

What should normally be tuned first in a cascade control system?

A. The slowest outer position loop

B. The safety relief valve

C. The fastest inner loop

D. The mechanical load

Correct Answer: C. The fastest inner loop

Question 10

If the reference position is 50.0 mm and the measured position is 49.6 mm, using e = r − y, what is the error?

A. −0.4 mm

B. 0.4 mm

C. 49.6 mm

D. 50.0 mm

Correct Answer: B. 0.4 mm