What is a speed loop?

A speed loop, also called a velocity loop, keeps a motor running close to its commanded speed by comparing speed feedback with the speed reference and adjusting the required torque. In a servo motor system, the regulator usually runs inside the servo drive. It is a form of closed-loop control: the drive checks what the motor actually does and corrects the difference.

A conveyor example: same speed, heavier load

Suppose a conveyor is commanded to run at 1500 rpm. A heavy item arrives, increasing the opposing load torque, and actual speed briefly falls to 1420 rpm. The regulator now sees an 80 rpm speed error and requests more motor torque. The inner current loop delivers the corresponding current. If the motor and drive have enough available capacity, speed returns towards the reference.

The graphic pairs actual speed in blue with motor torque in orange. Press “Start” to follow both signals along the same timeline. Notice that torque remains higher after speed has recovered: the heavier load still needs to be driven. The traces and eight-second playback are illustrative, not a performance specification. With reduced motion enabled, the complete graph remains visible without moving markers.

What closes the loop?

The controller repeatedly reads the reference, obtains actual speed, calculates the error and updates the torque request. The basic relationship is speed error = commanded speed − feedback speed. For a positive direction command, positive error calls for more accelerating torque. Negative error may require less driving torque or braking torque, depending on the operating point.

An encoder or resolver commonly supplies the underlying feedback. The drive can calculate velocity from changes in shaft position over time. Feedback design matters: incremental encoders and absolute encoders report position differently, while velocity quality also depends on resolution, sampling, signal integrity and filtering.

Where it sits in a servo system

The current loop is normally the fastest inner loop. It makes motor current follow its reference, providing the electrical control needed to develop torque. Current-loop gain belongs to this part of the system.

The speed loop sits outside that current loop and generates a torque-producing current request from the speed error. In field-oriented AC motor control, this is commonly associated with the q-axis current reference. It should not be treated as identical to total phase current under every operating condition.

For position control, an additional position loop produces a speed reference from position error. The familiar cascade is position → speed → current → motor. Position-loop gain and speed-loop gain therefore serve different purposes. A speed-controlled machine does not necessarily need an active outer position loop.

Speed reference and speed-loop gain are different settings

The reference answers “How fast should the motor run?” Loop gain determines “How should the controller respond when speed is wrong?” Doubling the gain does not double the commanded rpm.

Many drives use a PI speed regulator. The proportional term responds to the present error; the integral term accumulates error over time and helps remove a persistent speed offset when the system is stable and its output is not limited. Insufficient gain can give sluggish recovery. Excessive gain can produce oscillation, vibration or audible noise.

Kp, Ki and integral-time settings are not universally interchangeable across drive manufacturers. With some integral-time conventions, a shorter time means stronger integral action. Use the drive manual's definitions rather than copying numbers from another model. Loop bandwidth describes dynamic response; it is different from shaft rpm and PWM switching frequency.

Does every speed loop need an encoder?

No. Some sensorless vector-control systems estimate speed from electrical measurements and a motor model. Their speed regulator uses that estimate instead of a shaft-mounted speed sensor. Performance depends on the motor, algorithm and operating conditions; very low-speed or near-zero-speed capability must be checked against the particular drive's specifications.

Basic V/f control sets the relationship between voltage and frequency. That alone does not establish accurate measured shaft-speed regulation. Induction-motor slip changes with load, so shaft speed can change even at a fixed output frequency. Slip compensation can reduce the difference, but it is not the same measurement arrangement as encoder feedback.

Read the traces before changing the gain

Speed drops while torque or current stays at its limit: investigate available motor torque, excessive load, binding mechanics and overload conditions. More gain cannot create additional motor capacity. If the required torque exceeds what is available, the motor may lose speed or stall.

Speed repeatedly swings above and below the reference: investigate gain settings alongside coupling flexibility, backlash, resonance and noisy feedback. For drive commissioning, verify motor and feedback data first, then use the supported autotuning procedure or the manufacturer's tuning method. Record commanded speed, feedback speed and current or torque together. Servo accuracy depends on the complete system, not a single parameter.