What Is Voltage Sag?

What is voltage sag?

A voltage sag — also called a voltage dip — is a short-duration reduction in the RMS voltage of a supply. It is defined by two quantities: the retained voltage (or depth of the reduction, usually expressed as a percentage of nominal or of a declared reference voltage) and the duration for which the reduction persists. In industrial power quality practice, a voltage sag is generally understood as a drop to between 90% and 10% of nominal, lasting from half a cycle up to a few seconds. Events shorter than half a cycle are usually treated as transients; events longer than a few seconds shade into sustained undervoltage, which is a different phenomenon with different causes and different corrective measures.

For plant electrical engineers, the practical significance of voltage sag is not the voltage itself but what connected equipment does while the voltage is depressed. A sag that is invisible on a slow-responding analogue meter can still trip a contactor, reset a PLC, or force a variable speed drive offline. This page explains how voltage sags arise, how they affect common industrial loads, how severity is characterised, and what the general mitigation options are. It is an educational overview; equipment selection and protection design are covered separately.

How a voltage sag differs from other supply disturbances

Voltage sag is often confused with other events on the same network. Distinguishing them matters because the remedies are not interchangeable.

Voltage sag versus undervoltage

Undervoltage is a steady-state condition: the RMS voltage remains below its tolerance band for minutes or hours, typically because of overloaded transformers, undersized conductors, or a weak supply. A voltage sag is transient and returns to normal on its own once the initiating event clears. Tap changers, conductor upgrades and load balancing address undervoltage; they do nothing useful for a 100 ms sag caused by a remote fault.

Voltage sag versus interruption

An interruption is a reduction to below 10% of nominal (in some conventions, below 1%). Many practical events begin as a sag and deepen into an interruption as protection clears the fault. For equipment ride-through, the distinction is often less important than the total energy the load must supply from its own inertia or storage during the event.

Voltage sag versus transient

Transients are fast, oscillatory or impulsive deviations measured in microseconds to milliseconds. They are characterised by peak amplitude and rise time rather than RMS magnitude and duration. A capacitor switching transient, for example, may produce a brief voltage magnification at a downstream bus without registering as a sag in RMS terms.

What causes voltage sag

Voltage sag is an Ohm's law phenomenon: a current flowing through the source and distribution impedance produces a voltage drop. Any event that sharply increases current drawn from the supply, or that changes the network configuration so that a point of common coupling is fed through a higher impedance, will depress voltage at nearby loads. The most common industrial causes fall into four groups.

Network faults

Short circuits on the utility network or within the plant are the dominant cause of deep voltage sag. A fault on a parallel feeder, or on a higher-voltage network several kilometres away, produces a voltage depression across a wide area until protective devices isolate the faulted section. The depth depends on the electrical distance between the fault and the affected bus and on the source impedance; the duration is set by protection clearing time, typically from a few cycles to several hundred milliseconds.

Motor starting

Across-the-line starting of large induction motors draws a starting current several times rated current at a low power factor. The resulting voltage drop at the motor terminals and at adjacent buses is a voltage sag whose depth depends on motor size relative to source strength (short-circuit capacity) and on the impedance between the motor and other loads. Starting sags are predictable and repeatable, which distinguishes them from fault-induced events.

Transformer energising

Switching a transformer onto the network can produce an inrush current with a substantial DC offset and significant harmonic content. The resulting voltage sag is asymmetric and may persist for several cycles to a few seconds as the core settles. Adjacent transformers may also be affected through sympathetic inrush.

Load switching and other causes

Switching of large capacitor banks, sudden connection of heavy resistive or rectifier loads, and loose or high-resistance connections can all produce measurable voltage sag. High-resistance joints tend to produce repetitive, load-dependent sags at a specific bus and are worth investigating when sags occur without a corresponding network event.

How voltage sag affects industrial equipment

Susceptibility varies widely between device types. A sag of 70% retained voltage for 100 ms may be invisible to one load and fatal to another. The following sections describe typical behaviour; actual ride-through capability must always be confirmed against the specific device datasheet and its DC-link or control supply design.

Contactors and relays

AC contactors hold in by electromagnetic force from the coil, balanced against a spring. When coil voltage falls, holding force falls faster than the spring force at the point of release, and the contactor drops out. Typical AC contactors release somewhere in the range of 60–80% of rated coil voltage, and drop-out occurs within tens of milliseconds. This is why a shallow but not particularly deep sag can stop a motor that the motor itself could have ridden through. The restart is then manual or requires a dedicated control scheme, so a brief network event becomes a prolonged production stop. DC coils and electronic coil interfaces behave differently and generally tolerate wider voltage bands.

Variable speed drives

A modern PWM drive rectifies the incoming supply onto a DC link. During a voltage sag, the DC-link voltage follows the supply down until the input diodes or the pre-charge circuit can no longer maintain it. The drive trips on undervoltage when the DC link falls below its protection threshold. The ride-through time therefore depends on DC-link capacitance, load level, and the undervoltage threshold setting. Drives on light load ride through longer than drives on full load. Some drives offer kinetic buffering, which uses the motor and load inertia to hold the DC link up during the sag; others offer restart-on-the-fly to reconnect to a spinning motor after a trip. Both behaviours are configuration-dependent and should be verified per model.

PLCs and control systems

PLC susceptibility is dominated by the power supply, not the CPU. A switched-mode supply with adequate holdup time will carry the controller through short sags; the same controller on a marginal supply will reset. Discrete I/O and safety circuits add another layer: a contactor that drops out removes feedback, and a safety relay that sees an inconsistent state may latch a stop that requires deliberate reset. Control transformers, if used, introduce their own drop-out behaviour and should be sized with the same attention as the contactor coils they feed.

Motor groups and process lines

In a linked process, the weakest device sets the outcome. A line with fifteen drives and one contactor-fed pump will stop when the contactor drops out, regardless of the drives' ride-through. Conversely, a group of motors that all trip and then all attempt restart simultaneously can produce a second, self-inflicted voltage sag from starting current — a common mechanism for repeated trips after a single network event.

Characterising severity: magnitude, duration and ride-through curves

Because two numbers — retained voltage and duration — describe a sag, severity is naturally plotted on a magnitude-duration plane. This is the basis of the ride-through curves used in industry. The general concept behind ITIC and CBEMA curves is a boundary defining the region in which information technology equipment is expected to operate; events inside the boundary should be tolerated, events outside may cause malfunction. Original equipment manufacturers publish equivalent curves for their own products, and those curves, not a generic industry plot, are what matter for a specific installation.

Two practical points follow. First, a single point measurement is rarely enough: a site may experience many shallow sags per year and a few deep ones, and the mitigation decision depends on the distribution of both depth and duration. Second, sag depth varies with location on the network. A survey at the incoming supply does not describe conditions at a remote motor control centre fed through a long cable and a transformer. Monitoring at the buses that actually feed sensitive loads gives a more useful picture.

Measurement considerations

Voltage sag measurement is standardised in general terms by IEC 61000-4-30, which defines measurement methods and classes of instrument performance, including how RMS values are aggregated over cycles and how events are detected with hysteresis to avoid chattering event records. EN 50160 describes the characteristics of the voltage supplied in public distribution networks, including the expected frequency of dips, and is useful context for what a plant should reasonably expect from its utility supply. When comparing monitors or reviewing survey reports, confirm the measurement class, the aggregation interval, and the sag threshold settings — otherwise event counts from different instruments are not comparable.

Mitigation options: an overview

Mitigation works at three levels: reduce the number of sags, make the load tolerate the sags that remain, or ride through the event. In practice a combination is normal.

• Reduce sag incidence at source. Review utility supply arrangements, protection settings and fault levels; reduce motor starting current with soft starters, star-delta, or drives; sequence large load switching to avoid simultaneous inrush.
• Improve ride-through at the device. Select contactors with wide-tolerance or DC coils; set drive undervoltage thresholds and enable kinetic buffering or restart-on-the-fly where the process allows; specify power supplies with adequate holdup.
• Protect at the bus level. Uninterruptible power supplies, constant voltage transformers, and dynamic voltage restorers can hold a sensitive bus up during a sag. Each has constraints on depth, duration, and load type that must be checked against the site's actual sag profile.
• Restore quickly after a trip. Where ride-through is not economic, a controlled restart sequence — including motor restart after a trip, with restarting of motor groups sequenced to avoid a second sag — converts a long stoppage into a short one.
• Measure and verify. Continuous power quality monitoring at the critical buses establishes the baseline, validates the mitigation, and provides the event records needed for root-cause analysis.

Selection checklist

Before choosing any mitigation, confirm the following:

• The actual sag profile at the affected bus: typical depth, typical duration, and frequency per year.
• The ride-through curve of each sensitive device, from its own datasheet.
• The weakest device in each process line, not the most complex one.
• Whether the process can tolerate a controlled stop and restart, or requires continuous operation.
• Whether restarting the load after an event will itself cause an unacceptable sag.
• How the result will be verified: which monitor, at which bus, over what period.

Further reading on this site

Understanding voltage sag is the first step; specifying protection and restart behaviour is the next. Two pages on this site cover the engineering detail:

• Voltage sag protection and motor restart — how protection schemes and controlled restart sequences are applied to motor groups after a voltage dip.
• Power quality monitoring — instrument selection and monitoring arrangements for establishing a sag baseline and verifying mitigation.

FAQ

How long does a voltage sag last?

Most fault-induced sags last from a few cycles to a few hundred milliseconds, set by protection clearing time. Motor-starting sags typically last from a few hundred milliseconds to a few seconds, depending on motor and load inertia. Sags lasting longer than a few seconds are usually classified as sustained undervoltage and have different causes.

What is the difference between voltage sag and voltage dip?

The two terms describe the same event. "Voltage dip" is the term used in IEC publications and in European practice; "voltage sag" is common in North American and general industrial usage. Both refer to a short-duration RMS reduction characterised by retained voltage and duration.

Can a voltage sag damage equipment?

In most cases the effect is a trip or reset rather than physical damage. Damage risk arises indirectly: repeated restarts stress motors and contactors, uncontrolled reconnection to a spinning motor can stress drives, and process upsets can damage product or tooling. The engineering effort is therefore usually directed at ride-through and controlled restart rather than at surge-style protection.