-
Leon Zhang Sales Manager
-
Email: sale@relayprotect.com
-
Phone/WhatsApp: +86 13655813266

Power Quality Meter: A Practical Guide
A power quality meter logs voltage dips, harmonics and flicker to IEC 61000-4-30:2025. See Class A versus Class S, input wiring and when a survey is enough.
What is a power quality meter and what do Class A and Class S mean?
A power quality meter records voltage, frequency, harmonics, unbalance, flicker, dips and interruptions against a defined measurement method. IEC 61000-4-30:2025 sets that method and splits instruments into Class A and Class S; IEEE 1159-2019 names the disturbances a survey looks for. Edition 4.0 (October 2025) cancels the 2015 third edition. The letters are not quality grades: Class A requires uncertainty of 0.1% from 10% to 150% of Udin, contiguous 10 or 12 cycle blocks, and timing to ±20 ms at 50 Hz or ±16.7 ms at 60 Hz; Class S relaxes uncertainty and timing. Harmonics come from a discrete Fourier transform over a 10 or 12 cycle window (200 ms at 50 Hz, 167 ms at 60 Hz); IEC 61000-4-7:2002 defines the harmonic subgroup. IEC 62586-1:2017 and 62586-2:2017 set the minimum parameter set and the functional tests. Measurement is not correction: the record tells you where to spend the money. Setting review and quotation within 12 hours, MOQ 1, 7-15 working days.
Power quality meter is a three phase instrument that records supply voltage magnitude, frequency, harmonics, unbalance, flicker, dips and interruptions against a defined measurement method. IEC 61000-4-30:2025 sets that method and splits instruments into Class A and Class S. IEEE 1159-2019 names the disturbances a survey looks for. Accuracy, not the display, decides which class you need.
At a glance
- A power quality meter measures conducted phenomena; it does not correct them.
- IEC 61000-4-30:2025 Edition 4 defines Class A and Class S measurement methods.
- Class A holds supply voltage uncertainty to 0.1% of declared input from 10% to 150% of Udin.
- Harmonics use the IEC 61000-4-7:2002 grouped transform; flicker uses IEC 61000-4-15:2010.
- Dips, swells and interruptions are logged as events carrying a UTC timestamp.
- IEEE 519-2022 sets harmonic current limits at the point of common coupling.
- Class S suits surveys; Class A suits contractual or dispute measurements.
- Voltage inputs come from a VT or direct connection, current inputs from a CT.
What a power quality meter measures
IEC 61000-4-30:2025 lists the conducted phenomena an instrument may report on a 50 Hz or 60 Hz AC system. The catalogue covers power frequency, supply voltage magnitude, flicker, dips and swells, interruptions, transient voltages, unbalance, voltage harmonics and interharmonics, rapid voltage changes, mains signalling voltage, current magnitude, current harmonics and current unbalance. Emissions between 2 kHz and 150 kHz sit in the annexes. None of those figures are raw samples. Each one is the output of a named algorithm with a defined time window, which is the reason two instruments on the same bus can agree or disagree.
Pick the subset before the instrument. Name the class first.
Measurement is not correction. A power quality meter will show that a 30 ms sag dropped a contactor and let a trip coil release on a 110 V DC circuit. It will not add a ride through supply, change a breaker setting or stop the inrush current that started the event. The record tells you where to spend the money.
Revenue meters and protection relays answer narrower questions. A billing meter integrates energy to a tariff class and ignores a half cycle disturbance. A protection relay watches differential or overcurrent quantities and operates in milliseconds, following device numbers drawn to ANSI/IEEE C37.2-2022. Neither one keeps a week of 10 minute harmonic trends.
Class A and Class S under IEC 61000-4-30:2025
Selecting a relay for a live project?
Send the system voltage, CT ratio and tripping scheme. Our engineers confirm the model and reply within 12 hours.
Edition 4.0 of IEC 61000-4-30 was published in October 2025 and cancels the 2015 third edition. Class A stands for advanced and Class S for surveys. The letters are not quality grades. They describe how strictly the instrument must compute, aggregate and time stamp each value.
Class A requires a measurement uncertainty of 0.1% of declared input voltage over a range from 10% to 150% of Udin, contiguous non overlapping 10 or 12 cycle blocks, an external time source accurate to ±20 ms at 50 Hz or ±16.7 ms at 60 Hz, and flagging of data recorded during a dip or swell. Class S relaxes the uncertainty and the timing but keeps the same parameter definitions. A plant looking for the feeder that starts a weekly flicker complaint can use Class S.
Buy the class the dispute needs. Proof beats a datasheet.
| Parameter | Class A | Class S | Why the distinction matters |
|---|---|---|---|
| Supply voltage uncertainty | 0.1% of Udin over 10% to 150% | Wider tolerance permitted | Decides whether a reading survives a contractual challenge |
| Measurement blocks | Contiguous, non overlapping, 10 or 12 cycles | Gaps permitted | A gap can hide the first half cycle of a sag |
| Time synchronisation | GPS or NTP, ±20 ms at 50 Hz | Relaxed clock requirement | Events from two sites cannot be correlated without a shared clock |
| Aggregation | 3 s, 10 min and 2 h intervals | Same intervals, fewer obligations | Long term reports still need a common structure |
| Flagging | Data during an event is marked | Not mandatory | Prevents a dip being averaged into a clean 10 min value |
| Compliance testing | IEC 62586-2:2017 type tests | Not required | A claim on a datasheet needs a test report behind it |
Table 1. The two measurement classes, read clause by clause rather than as marketing labels.
IEC 62586-1:2017 sets the minimum parameter set for a power quality instrument in portable and fixed installations, and IEC 62586-2:2017 defines the functional tests and uncertainty checks that prove the claim. Ask for the test report, not a badge on the nameplate. Long term drift in the analogue front end is a real failure mode on unattended panels, and a recalibration interval belongs in the maintenance schedule.

Harmonics and interharmonics: the IEC 61000-4-7:2002 method
Harmonic magnitudes come from a discrete Fourier transform over a rectangular 10 or 12 cycle window, 200 ms at 50 Hz and 167 ms at 60 Hz. IEC 61000-4-7:2002 with its 2008 amendment defines the harmonic subgroup, which adds the adjacent spectral bins into each order so that a small frequency drift does not move energy between neighbouring orders. Class A sampling must be gapless.
Interharmonics sit between the integer orders and often point to a variable speed drive or an arc furnace. Above 2 kHz the 2025 edition of IEC 61000-4-30 separates the 2 kHz to 9 kHz method from a newer 9 kHz to 150 kHz method. A meter quoted only to the 40th order will not see them.

Total harmonic distortion is a summary, not a diagnosis. Two feeders can both read 5% THD while one carries a dominant 5th and the other a dominant 11th. Filter sizing, capacitor rating and neutral conductor heating all depend on which order dominates. Ask for the individual spectrum with the summary, and check the nameplate of any capacitor bank before attributing a rising THD to the supply.
Flicker, dips and rapid voltage changes
Flicker is measured by the flickermeter chain in IEC 61000-4-15:2010, which models the lamp and the human eye rather than reporting a plain voltage swing. Short term severity Pst comes from a 10 min interval and long term Plt from twelve of those intervals over 2 h. IEC 61000-4-11:2020 covers the immunity test that tells you whether a piece of equipment survives a given dip, which is a different question from measuring the dip itself.
Dips, swells and interruptions are events. A Class A instrument updates the RMS value every half cycle and declares an event against a threshold, commonly 90% of Udin for a dip and 110% for a swell, with a duration that can run from 10 ms to several seconds. Rapid voltage changes sit below the dip threshold and were one of the methods corrected in the 2025 edition. An event record needs the phase, the residual voltage, the duration and a shared clock.
A dip is a timestamp, not a number. Clock drift breaks correlation.
IEEE 1159-2019 and the disturbance categories
IEEE 1159-2019 groups power quality disturbances by duration and typical magnitude so that engineers describing an event use the same words. Transients cover sub cycle impulses and oscillatory bursts. Short duration variations cover sags, swells and interruptions measured in cycles to a minute. Long duration variations cover overvoltage, undervoltage and sustained interruption. Waveform distortion covers harmonics, interharmonics, notching and DC offset. Voltage fluctuation covers flicker, and frequency variation covers the small departures a 50 Hz or 60 Hz system tolerates. IEEE 519-2022 then sets harmonic current and voltage limits at the point of common coupling rather than inside the plant.
| IEEE 1159 category | Typical duration | Typical magnitude | What a survey should log |
|---|---|---|---|
| Impulsive transient | Microseconds to milliseconds | Up to several kV | Peak, rise time, phase and trigger waveform |
| Oscillatory transient | 0.3 ms to 5 ms | 0 to 4 pu | Frequency of the ring and the switching instant |
| Voltage sag | 0.5 cycle to 1 min | 0.1 to 0.9 pu | Residual voltage, duration, phase, UTC start |
| Momentary interruption | 0.5 cycle to 3 s | Below 0.1 pu | Recloser operation and downstream contactor drop |
| Harmonic distortion | Steady state | 0 to 20% THD | Individual orders to the 40th and the subgroup method used |
| Voltage flicker | 10 min to 2 h | 0.1% to 7% swing | Pst, Plt and the load step behind it |
Table 2. The categories a monitoring campaign is normally asked to separate.
EN 50160:2010 describes the voltage characteristics a European public supply is expected to deliver, including 230 V ±10% and a nominal 50 Hz within a stated band. It is a supply characteristic, not a guarantee that every load will tolerate it. IEC 61000-2-4:2002 sets compatibility levels inside industrial plants, where the environment is harsher than the public network. Confusing the two documents leads to a specification that no instrument can police.
Wiring a power quality meter: VT, CT and direct inputs
A three phase four wire installation needs four voltage inputs, three current inputs and one earth reference. On a low voltage board the voltage inputs may connect directly at 400 V or 690 V through fused taps. On an 11 kV or 33 kV feeder they come from the secondary of a voltage transformer, usually 100 V or 110 V line to line, and the current inputs come from the 5 A or 1 A secondary of a protection or metering CT. Ratio and polarity errors are the two faults that survive every other check. A reversed CT on one phase quietly doubles the reported negative sequence unbalance and leaves the voltage readings looking perfect.
Harmonic work puts an extra demand on the transducer. A 0.2S or 0.5 metering CT is specified at power frequency, and its ratio error grows well above the 50th harmonic. Where current harmonics are the point of the survey, ask for the frequency response of the CT or fit a Rogowski coil with a stated bandwidth. The current transformer construction and polarity guide explains why a protection core behaves differently from a metering core. Cable routing matters too. Run the twisted measurement pairs away from the trip circuit and the breaker closing leads, keep the shield earthed at one end, and respect the interlock and utilization category rules on the protection and control panel where the instrument is mounted.

Auxiliary supply is the last item on most schedules and the first cause of a lost record. A fixed meter fed from a 24 V DC or 110 V DC station battery needs a varistor or suppressor on the input and a hold up time long enough to write the event that follows a supply dip. Solid state inputs tolerate a narrower transient than the relay contacts they replaced, and a heatsink rated for a 55°C cubicle is not optional inside a sun exposed kiosk.
Portable versus fixed installations
A portable survey runs for seven to thirty days on a suspicious feeder, then the instrument moves on. Fixed installations stay for years and feed a historian. The choice changes the specification more than the measurement class does. Portable units need safe clamp on CTs, a carrying case and a battery that survives a shift. Fixed units need a panel cut out, a permanent 110 V DC or 230 V AC supply, an enclosure rated to at least IP54, and a temperature range covering –10°C to 55°C in an unconditioned switchroom.
Fixed meters usually push data over Modbus TCP or IEC 61850 into a SCADA historian. The SCADA polling and tag design guide covers why a 10 min aggregated value belongs in the historian while the raw event waveform does not. Bandwidth, not accuracy, is usually the limit on a remote site. Keep the local event buffer on the instrument and pull the summary.
Reading an accuracy and sampling specification
Suppliers like to quote a single accuracy at full scale. Class A does not work that way. Supply voltage uncertainty is 0.1% of Udin over an input range from 10% to 150%, which is a much harder claim on the low end of the scale. Frequency uncertainty is typically within 10 mHz. Harmonic voltage uncertainty is stated per order above a measurement threshold. Time accuracy is ±20 ms at 50 Hz and ±16.7 ms at 60 Hz.
| Quantity | Class A requirement | Measurement interval | Reference document |
|---|---|---|---|
| Supply voltage magnitude | 0.1% of Udin, 10% to 150% range | 10 or 12 cycles, contiguous | IEC 61000-4-30:2025 Clause 4 |
| Power frequency | Within 10 mHz | 10 s interval | IEC 61000-4-30:2025 |
| Harmonic voltage | Stated per order above a threshold | 10 or 12 cycle DFT, gapless | IEC 61000-4-7:2002 |
| Flicker Pst and Plt | Per flickermeter chain | 10 min and 2 h | IEC 61000-4-15:2010 |
| Dip residual and duration | Half cycle update | Event triggered | IEC 61000-4-30:2025 |
| Time stamp | ±20 ms at 50 Hz | Continuous UTC reference | IEC 61000-4-30:2025 |
Table 3. The numbers worth asking a supplier to put in writing.
Full scale accuracy alone means little. Ask for the range.
Aggregation windows and why timestamps matter
Raw samples are compressed in layers. The basic block is 10 cycles at 50 Hz or 12 cycles at 60 Hz, roughly 200 ms. Twelve of those build the 150 or 180 cycle, 3 s value. Each layer inherits the clock of the one below it. A GPS receiver or an NTP feed over Ethernet is what keeps the 10 min boundary aligned between instruments, and Class A asks for ±20 ms at 50 Hz. When the external reference is lost, a drift better than ±1 s in 24 h keeps the record usable, but that relaxed tolerance no longer proves Class A compliance.
| Window | Duration at 50 Hz | Built from | Used for |
|---|---|---|---|
| Basic measurement | 10 cycles, 200 ms | Sampled waveform, contiguous | RMS, harmonics, event detection |
| Short aggregate | 150 cycles, 3 s | 15 basic blocks | Dip and swell characterisation |
| Survey aggregate | 10 min, UTC aligned | 200 short aggregates | Compliance reports and trends |
| Flicker long term | 2 h | 12 survey aggregates | Plt and weekly flicker comparison |
Table 4. One clock governs every layer of the record.
Flagging is part of the same mechanism. A dip that occurs inside a 10 min interval must be marked so that a compliance report does not quietly average the disturbance away. Without flagging, a site can pass a weekly THD and voltage report while its contactors dropped out four times on Tuesday.
Compliance limits: EN 50160:2010 and IEEE 519-2022
Measurement and compliance are separate jobs. IEC 61000-4-30:2025 tells the instrument how to compute a value. EN 50160:2010 describes the voltage characteristics of public distribution, including 230 V ±10%, 50 Hz within ±1% for most of a week, a total harmonic distortion generally below 8%, and a stated number of dips per year. IEEE 519-2022 sets current distortion limits at the point of common coupling as a function of the short circuit ratio, so a weak rural feeder is allowed less harmonic current than a stiff urban one.
Standards and editions behind a measurement
Six documents appear in most specifications. The edition matters, because the 2025 revision of the measurement method changed the rapid voltage change and voltage event rules, and a report written against the 2015 third edition is not the same evidence. Device numbers in the surrounding switchgear drawings follow ANSI/IEEE C37.2-2022, and the relay performance standard IEC 60255-1:2022 sits behind the protection side of the same panel.
| Standard | Edition | What it fixes | Where it applies |
|---|---|---|---|
| IEC 61000-4-30 | 2025, Edition 4.0 | Measurement methods and Class A / Class S | The instrument itself |
| IEC 61000-4-7 | 2002 with AMD1:2008 | Harmonic and interharmonic subgroups | Harmonic and 2 kHz to 9 kHz bands |
| IEC 61000-4-15 | 2010 | Flickermeter chain, Pst and Plt | Flicker measurement |
| IEC 62586-1 | 2017 | Minimum PQI parameter set | Portable and fixed instruments |
| IEC 62586-2 | 2017 | Functional and uncertainty type tests | Compliance proof for Class A |
| IEEE 1159 | 2019 | Disturbance categories and terminology | Survey scoping and reports |
| IEEE 519 | 2022 | Harmonic limits at the PCC | Utility interface studies |
| EN 50160 | 2010 | Public supply voltage characteristics | Utility supply compliance |
Table 5. Editions quoted in this guide, checked against the current catalogue.
A specification that lists a standard without an edition is incomplete. Ask the supplier which edition the type test was run against and whether a 2025 firmware release is available for an instrument shipped in 2023.
Where a power quality meter belongs in a substation
One instrument at the incomer answers a utility question. Three instruments, at the incomer, the bus coupler and the offending feeder, answer a plant question. Synchronised records let you see a sag arrive on the incomer and deepen on the feeder, which separates a supply event from an inrush current or a short-circuit inside the site. On generator and transformer circuits the meter sits beside differential and overcurrent protection, but it stays on the measurement CT core and never in the tripping path. A tripping contact or trip coil circuit is not a convenient place to borrow a signal.
Placement also decides what a breaker operation looks like. A meter downstream of an automatic transfer switch will log the interruption caused by the changeover; one upstream will log the sag that started it. Both are correct and neither tells the whole story alone. Draw the single line, mark each voltage transformer and current transformer, and choose the points before ordering hardware.
One instrument rarely answers a network question. Place three, not one.
Selecting a model for this scheme? Send the voltage level, CT/VT ratios and your relay list. An engineer returns a model mapping and quotation within 12 hours →
Commissioning checks before you trust a record
A meter that has never been checked will happily produce a year of confident nonsense. Work through the inputs, ratios, clock and event replay before the survey window starts.
| Check | Method | Pass criterion |
|---|---|---|
| Input polarity and phase order | Compare the instrument phasors against a known rotation | L1, L2, L3 sequence matches the single line |
| CT ratio and burden | Inject a known 1 A or 5 A secondary current | Reported current within the CT class error |
| VT ratio and fuse state | Measure the secondary against the primary schedule | 100 V or 110 V present on all three phases |
| Clock synchronisation | Compare the displayed UTC against a GPS or NTP reference | Within ±20 ms at 50 Hz |
| Event replay | Force a controlled load step or a test dip | One tagged event with a waveform and timestamp |
| Export and storage | Pull one 10 min record to the historian | Values match the local display exactly |
Table 6. Six checks that take an hour and save a month of argument.

Calibration traceability closes the loop. Ask for the certificate, note the drift allowance of the analogue front end, and set a recall interval. Coil contacts that stick, a nuisance trip on a humid morning, a burnt terminal or a failed fan can all corrupt a record long before anyone reads it. Check the certificate.
Specifying a power quality meter: a short checklist
Write the schedule in the order an installer works, not in the order a catalogue is printed. Six lines settle most of it.
- Name the measurement class and the parameter subset to IEC 61000-4-30:2025.
- State the voltage inputs, CT ratios, CT frequency response and the earth reference.
- Require a UTC clock source and say how the 10 min boundary is aligned.
- Fix the aggregation intervals, flagging rule and the export format for the historian.
- Ask for IEC 62586-2:2017 type test evidence and a calibration certificate.
- Set the enclosure rating, auxiliary supply, hold up time and recalibration interval.
Leave the waveform storage on the instrument. Pull the summary over the network and fetch the raw event files only when a report needs them. Store waveforms locally.
When a power quality meter is the wrong tool
A monitoring instrument is a poor answer to several questions that sound similar. Fit the tool to the decision, not to the available panel space.
- Do not use it to prove a protection relay operated correctly. A digital fault recorder captures the trip coil current, breaker auxiliary contacts and the operate instant at a far higher rate.
- Do not buy a Class A survey to find a single nuisance trip that happened once. An event log from the existing relay and the SCADA sequence of events will usually name the cause first.
- Do not expect it to resolve a sub cycle switching transient. A 200 ms harmonic window is the wrong instrument for a 5 µs impulse; use a high speed transient recorder or an oscilloscope.
- Do not rely on it for revenue settlement. Billing accuracy follows a metering class and a tariff, not IEC 61000-4-30.
- Do not install one at a point with no voltage transformer and no safe CT access. A current clamp on an unshielded 11 kV cable gives neither the accuracy nor the safety the report needs.
- Do not treat a seven day survey as a seasonal conclusion. A site with a summer air conditioning load and a winter heating load needs at least one full operating cycle.
A digital fault recorder selection guide covers the first two cases. A power quality meter earns its keep when the question is what the supply actually did over weeks, in numbers both parties will accept.
Failure modes and what the waveform shows
The instrument is also a witness to the plant around it. A nuisance trip on a humid morning often appears as a 40 ms sag followed by a contactor drop, with the trip coil current collapsing at the same instant. A short-circuit on an adjacent feeder produces a deeper sag with a clean recovery when the breaker clears. Transformer inrush shows a decaying second harmonic and a current waveform that does not match the voltage distortion. Stuck auxiliary contacts leave an event with no corresponding breaker state change, and a drifting CT or a burnt input terminal shows up as a slow offset that no plant event explains.
Read the current channels with the voltage channels. A falling voltage with rising current points downstream; a falling voltage with falling current points upstream. Welded contacts and a failed auxiliary supply produce flat traces that look like a healthy feeder until somebody checks the channel against the single line. Correlate the event with the breaker and the interlock state before blaming the utility. Read the single line first.
Frequently asked questions
What is the difference between a power quality meter and a power quality analyzer?
The two names are often used for the same hardware. A meter is normally specified for the measurement method and accuracy class in IEC 61000-4-30:2025, while an analyzer is marketed on waveform capture and diagnostic depth. Check the Class A or Class S declaration rather than the product name.
Does a Class A instrument need GPS time synchronisation?
It needs an external time reference accurate to 20 ms at 50 Hz or 16.7 ms at 60 Hz. A GPS receiver or an NTP feed over Ethernet provides it. When the reference is lost, drift better than 1 s in 24 h keeps the record usable but no longer proves Class A compliance.
How many cycles does IEC 61000-4-30:2025 use for the basic measurement interval?
The basic interval is 10 cycles at 50 Hz or 12 cycles at 60 Hz, about 200 ms, measured contiguously without gaps. Fifteen of those blocks build the 3 s value and 200 of the 3 s values build the 10 min value.
Can a power quality meter measure DC side ripple on a solar plant?
Standard IEC 61000-4-30 measurements cover conducted phenomena on 50 Hz or 60 Hz AC systems. DC side ripple needs a separate DC input with a stated bandwidth, or a dedicated recorder. Check that the instrument offers it before assuming an AC survey covers the array.
What CT class do I need for harmonic current measurements?
A 0.2S or 0.5 metering class is specified at power frequency only. Ratio error grows well above the 50th harmonic, so ask for the frequency response of the CT or fit a Rogowski coil with a stated bandwidth when current harmonics are the point of the survey.
How long should a survey run before the data is representative?
Seven to thirty days is the usual portable survey window, and it should include the heaviest and lightest load days of the week. A site with seasonal air conditioning or heating load needs at least one full operating cycle before a compliance conclusion is drawn.
Does EN 50160:2010 define legal limits for every site?
It describes the voltage characteristics of a public distribution supply, including 230 V plus or minus 10 percent and nominal 50 Hz within a stated band. It is a supply characteristic, not a guarantee that every load will tolerate it. Compatibility levels inside a plant come from IEC 61000-2-4:2002.
Conclusion
A power quality meter earns its place when the question is what the supply actually did over weeks, in numbers a utility and a plant engineer will both accept. Start with the decision: a contractual dispute needs Class A with a traceable clock, a flicker complaint needs a survey at the right feeder, and a single nuisance trip usually needs the protection event log instead. Write the parameter subset, the input ratios, the aggregation intervals and the edition of IEC 61000-4-30 into the schedule before comparing datasheets. The rest is installation discipline: correct CT polarity, a shared UTC clock, a commissioned event replay and a calibration certificate kept with the report.
Related reading
- Online versus portable analyzer buying guide. Where a fixed installation pays for itself.
- Power distribution architectures. Where the point of common coupling sits in the network.
- Current transformer cores, turns and polarity. Why a metering core behaves differently from a protection core.
- How SCADA systems poll, tag and log. Which values belong in the historian and which stay on the instrument.
For a quotation with drawings and a test certificate, send the feeder schedule.
Working on a similar scheme? Send the single-line diagram, relay list and CT/VT data. Quote in 12 hours →




