Speed - Torque - Position control with servo?

Servo speed, torque and position control

Which variable should your motion system regulate?

Keeping a conveyor at a specified rotational speed, producing the required winding pull and stopping a cutting axis at a target coordinate are different control tasks. Servo systems use speed, torque and position modes to address these needs. The selected mode determines which physical variable of the servo motor takes priority; the other variables must remain within suitable limits.

The servo drive compares the command with feedback and regulates the current supplied to the motor. An encoder or resolver provides position feedback, from which speed can be derived. Selecting a control mode does not disable every other feedback or protection function. Check the drive model for supported modes, command interfaces and mode-switching conditions.

Speed mode: maintaining speed as the load changes

In speed control, the reference might be 1,000 rpm. The speed loop evaluates the difference between actual and commanded speed and sends a torque demand to the inner current loop. If products increase the load on a conveyor, the drive produces more torque, provided the current limits and motor capability allow it.

The objective is not necessarily high speed: stable low-speed operation is also a speed-control task. A speed command alone does not guarantee that an axis stops at the same position on every cycle. When cut length or a stopping coordinate matters, consider the difference between speed and position control. Speed can drop when the torque limit is reached; this does not automatically indicate incorrect gain tuning.

Torque mode: regulating the turning effort

In servo torque control, the command may be expressed in N·m or as a manufacturer-defined percentage of rated torque. In common permanent-magnet servo systems, electromagnetic torque is approximately related to the torque-producing current component within the applicable operating region: T ≈ K_t × I_q. Interpret the current convention, torque constant and percentage scaling using the manufacturer's documentation.

The torque displayed by a drive is usually calculated from measured current and a motor model; it is not an independent measurement of actual load-side torque. Gearbox losses, friction and acceleration create differences. Direct monitoring of load torque requires a suitable torque sensor. A torque command does not hold motor speed constant. The axis can accelerate if its load decreases, so evaluate the manufacturer's torque-mode speed-limiting function and appropriate machine safeguards.

Constant winding torque does not maintain constant web tension as roll diameter changes. For an ideal, lossless drum at steady state, F = T / r: 3 N·m at a radius of 0.05 m produces 60 N of pull, whereas a radius of 0.10 m gives 30 N. Real systems use diameter calculation, friction compensation and, where required, load-cell feedback. Assess the relationship between torque and rotational speed, the motor's torque-speed curve and its continuous operating limits together.

Position mode: reaching a target coordinate

In position control, the target may be a 90° rotation, 50 mm of travel or an absolute coordinate. The controller generates a motion profile, and the following error between commanded and feedback position is monitored. Travel speed, acceleration and motion smoothness must also be defined. Increasing position-loop gain alone does not make a machine more accurate; excessive settings can cause vibration and overshoot.

For example, a directly coupled ball screw with a 5 mm/revolution lead requires 10 motor revolutions for 50 mm of travel. If the drive input is configured for 20,000 command pulses per motor revolution, the move requires 200,000 pulses. This is command scaling, not encoder resolution. Include the ratio if a gearbox is fitted. On CNC axes, backlash, compliance, screw lead and mechanical connections affect actual load position; a motor encoder does not directly measure all these effects.

How do the three loops work together?

A common cascaded architecture places the current loop innermost, with the speed loop around it and the position loop outermost. Position error produces a speed demand; speed error produces a torque or current demand. The current loop has the fastest response. This is a common architecture, rather than a claim that every manufacturer uses an identical control algorithm.

The position loop can run in the drive or in a suitable PLC and motion controller. Consequently, a drive operating in speed mode can still form part of a machine whose external controller closes the position loop. The principles of PID control help explain the relationship, but using identical PID settings for every servo loop is inappropriate.

Distinguish the command interface from the control mode

On compatible drives, pulse-and-direction signals carry position commands: total pulse count affects travel, while pulse frequency affects travel speed. Electronic gearing defines the scaling between the input command and motor movement. Depending on the model and assigned function, an analog input can instead carry a speed or torque reference.

The presence of an EtherCAT interface does not automatically make every operating mode available. On suitable systems, CSP, CSV and CST identify cyclic synchronous position, velocity and torque modes respectively. Controller support, cycle time, process-data mapping and drive configuration must be checked together. Communication protocol, control mode and motion profile are separate concepts.

What should be checked during commissioning?

Motor and drive selection requires more than a kW rating. Consider continuous torque, short-duration peak torque, maximum speed, duty cycle and the load and transmission calculation. A torque limit in speed or position mode does not convert that mode into torque control; it caps the available torque. If the limit is too low, speed deviation or following error increases.

Before tuning servo gains, check motor data, feedback direction, mechanical connections and command scaling. Follow the manufacturer's automatic tuning procedure, then inspect speed, following-error and torque traces with the real load. On frequently decelerating axes, assess regenerative energy and, where required, braking-resistor capacity. When regulating a process variable such as force or tension, a PID application using a servo needs a clearly defined outer-loop measurement point and drive reference.

Editor's note: “The servo runs at 1,000 rpm, so its torque must also be constant” is incorrect. In speed mode, torque can change with load; in torque mode, speed can change. In position mode, the motor can draw current while stationary to hold its target position. First decide which variable must be regulated, then set appropriate limits for the others.