Current IntelPower LibraryShop main store

How-to

Assessing VFD harmonics when one small drive runs a plant

Power quality and harmonics Published 7 min read

Short answer

No filter is justified on ratings alone: 6 hp × 0.746 kW/hp = 4.48 kW is 6.6% of 68 kW connected capacity. Measure voltage THD at the shared bus with the drive off and then running, and read the drive's input current. If the bus stays under IEEE 519's 5% THD limit with no single harmonic above 3%, leave it alone; if it climbs over, an input reactor on the drive is the fix.

On this page
  1. The screening math: 6 hp against 68 kW
  2. What IEEE 519 limits, and where
  3. The assessment, step by step
  4. Worked example: 6 hp on a 400 V bus
  5. When a small drive still causes trouble
  6. The decision
  7. Questions asked

No. One 6 hp drive on 68 kW of three-phase capacity is not, by itself, a harmonics problem: 6 hp is 4.48 kW of rated output, 6.6% of the connected load, and a drive that small will not push a healthy bus past the voltage-distortion limit. What it will do is read badly on an analyzer, with an input current THD that looks alarming while the amps behind it are few, and collect blame for symptoms it did not cause. Give it one afternoon of measurement before anyone buys a filter: voltage THD at the shared bus with the drive off and then running, plus the drive's input current. Under 5% bus THD with no single harmonic above 3% and nothing misbehaving, you are finished. Over that, the answer is an input reactor, not a filter bank.

The screening math: 6 hp against 68 kW #

Convert rated shaft output to kilowatts: 6 hp × 0.746 kW/hp = 4.48 kW. Against 68 kW of connected three-phase capacity that is 4.48 ÷ 68 = 0.066, or 6.6%. Input power runs somewhat higher after motor and drive losses, about 5.4 kW at typical efficiencies, which is still under 8% of the connected base.

Two things move the ratio. If 68 kW is connected capacity rather than metered demand, the honest denominator is smaller: on a true 45 kW demand the drive output share is 4.48 ÷ 45 = 10%. And confirm the 6 hp is one motor on one drive, read from the nameplate, not from a project summary. Either way the drive sits at single-digit percent of the plant: too small to justify mitigation on ratings alone, big enough to deserve a measurement.

What IEEE 519 limits, and where #

IEEE Std 519-2022 sets steady-state voltage and current distortion limits at the point of common coupling, the PCC, where the installation meets the utility, normally the service entrance or meter position. It does not impose one THD number on every drive, panel or electronic load inside the plant.

On the voltage side, the steady-state limit for general systems is 5% THD at the PCC with no individual harmonic above 3%. On the current side, the limits are expressed as TDD, total demand distortion: harmonic current referenced to the maximum demand load current at its fundamental component, not to whatever the drive happens to draw at the moment. A six-pulse rectifier concentrates its characteristic harmonics at h = 6n ± 1, so the 5th, 7th, 11th and 13th are the rows to watch in the analyzer spectrum. Editions, service-voltage rows and utility interconnection agreements differ, so confirm the binding limits and the PCC definition with your utility and the authority having jurisdiction.

The assessment, step by step #

  1. Fix the demand base. Pull maximum demand and power factor from the billing meter or a 30-minute demand record covering a normal production period. Connected capacity is the loose upper bound; the metered number is what TDD divides by.
  2. Read the drive's input side. Record the rated input current from the nameplate and manual, whether a DC-link choke is built in, and whether an input reactor is already wired in, verified against the as-built circuit rather than a purchase description.
  3. Map the shared bus. List every load on the same distribution board that distortion can disturb, including control power supplies, metering and any capacitor bank, and note anything that switches on a timer.
  4. Connect the analyzer safely. De-energize and test for absence of voltage before opening any panel the connection allows; where a connection must be made at energized terminals, that is energized work under NFPA 70E, 2024 edition, justified, planned and performed with the rated instruments, tools and personal protective equipment that standard requires. Use a power-quality analyzer conforming to IEC 61000-4-30 Class A, with current clamps rated for the voltage category at that board and ranged so a distorted waveform cannot clip.
  5. Record the baseline with the drive off: bus voltage, voltage THD and the harmonic spectrum at the distribution board that feeds the drive.
  6. Repeat with the drive running at its normal load, and capture the drive feeder at the same time: fundamental input current, current THD, spectrum and true power factor, with the drive's load percentage noted in the record.
  7. Compute the two screening numbers from the records: the drive's harmonic current as a share of demand current, and the change in bus voltage THD from drive-off to drive-on.
  8. Decide against the cases below, and file the records; they are the baseline the next added drive gets compared to.

Worked example: 6 hp on a 400 V bus #

Illustrative inputs, stated in full: 400 V three-phase supply; maximum demand taken as 68 kW at 0.85 power factor; one 6 hp drive-motor at rated load; analyzer readings of 9.0 A fundamental input current, 40% current THD, bus voltage THD 2.4% with the drive off and 2.9% with it running, 5th harmonic at 1.6%.

Demand current: I = P ÷ (√3 × V × PF) = 68 000 W ÷ (1.732 × 400 V × 0.85) = 68 000 ÷ 589 = 115 A.

Drive harmonic current: 0.40 × 9.0 A = 3.6 A. Its share of demand: 3.6 ÷ 115 = 3.1%, which is this drive's entire contribution to the TDD assessed at the PCC.

Bus check: 2.9% THD against the 5% limit, 1.6% fifth harmonic against the 3% individual limit, and a drive-on minus drive-off difference of half a point.

Decision for these inputs: pass, with margin. No filter and no reactor is required by the numbers. These are screening estimates; the drive-off and drive-on comparison and the metered demand base carry the decision, and the utility's stated limits and the drive documentation govern the final call.

When a small drive still causes trouble #

Symptom First place to look Why
Bus THD above 5% only when the drive runs Source impedance: small service transformer, long feeder Harmonic current times supply impedance sets the voltage distortion, so a stiff bus barely moves
Bus THD high with the drive off Upstream: other rectifier loads, the utility supply, a neighbouring consumer The drive is not the source, and filtering it changes nothing
Breakers trip at the same clock time daily Utility capacitor switching, scheduled plant loads Steady distortion does not keep a schedule; transients do
Supply transformer or feeder runs hot Loading, terminations and cooling first, then the spectrum Heat is usually load plus losses; confirm with the spectrum before assigning cause
Controls or meters misread near the drive Grounding and bonding of the drive and its cable Output common-mode current returns on grounding conductors, a bonding diagnosis rather than a harmonics one

Where the sufferer is the supply transformer rather than the bus voltage, the heating mechanism and its mitigation math are a subject of their own: troubleshooting transformer harmonics and overheating walks that case end to end. Current found returning on equipment grounding conductors likewise has its own diagnostic path, covered in diagnosing 200 A currents on equipment grounding conductors.

The decision #

  • Bus THD at or below 5%, no harmonic above 3%, drive running, nothing misbehaving: no mitigation. Record the readings and close the file.
  • Bus THD above 5% only with the drive running: fit an input line reactor, or use the drive's DC-link choke option if the manual lists one, sized to the rated input current; choosing a line reactor or motor reactor for a VFD covers the selection. Re-measure afterward; one small drive with a reactor is normally the end of it.
  • Bus THD above 5% with the drive off: the source is upstream. Take the drive-off record to the utility and audit the plant's other rectifier loads; money spent on this drive's harmonics is wasted.
  • Trips at fixed clock times: correlate trip timestamps with an event record and check the utility's capacitor-switching schedule. A reactor also damps transients reaching the drive's DC bus, which is protection for the drive rather than harmonics correction.
  • Symptoms with every number passing: stop treating it as harmonics. Chase loading, cooling, connections and grounding, in that order.

For a 6 hp drive on 68 kW, the assessment is a measurement job, and in most plants the measurement is also the conclusion: the readings pass, the record goes in the file, and the next look only comes when another drive joins the bus.

Questions asked

The analyzer shows over 80% current THD at the drive input. Is the plant in trouble?

No, not on that number alone. Current THD is referenced to the fundamental at the moment of reading, so a lightly loaded drive shows a large percentage over small amps. Convert to amps: at 9 A fundamental, 80% THD is 0.80 × 9 = 7.2 A of harmonic current, about 6% of a 115 A demand base. Judge the drive by those amps, by the bus voltage THD, and by whether anything malfunctions; the percentage by itself describes neither transformer heating nor disturbance to other loads.

Does IEEE 519 apply inside my plant, at the drive terminals?

No. IEEE Std 519-2022 applies its steady-state voltage and current distortion limits at the point of common coupling with the utility, which for most plants is the service entrance or meter position. Inside the plant, the numbers that matter are the voltage THD other loads actually see on the shared bus and how those loads behave. The utility enforces at the PCC; equipment compatibility inside the plant is your own acceptance criterion.

Should I add a line reactor even when the measurements pass?

Many plants do, for reasons beside harmonics: an input reactor limits the rectifier peak charging current, damps voltage transients and utility capacitor-switching spikes reaching the DC bus, and cuts the 5th and 7th harmonic current at the same time. Size it to the drive's rated input current per the drive documentation. As a harmonics decision, though, passing measurements need no reactor; it is drive protection, not a compliance requirement.

Breakers trip at the same time every day. Is that harmonics from the drive?

Probably not. Steady-state distortion is present whenever the loads run, so it does not keep a schedule. Fixed-time trips point at a scheduled event, and utility capacitor-bank switching is the classic one, along with plant load steps. Match the trip timestamps against a time-stamped event record from the analyzer; if the trips continue with the drive off, the source is upstream of it entirely.

What demand current should I divide by for the TDD-style share?

Use the maximum demand load current from the billing meter or a 30-minute demand record over a representative production period, taken at its fundamental frequency. Connected capacity overstates the base: 68 kW connected might be 45 kW metered, which moves the drive share from 6.6% to 10%, still small but honest. The billing demand is also the base the utility would use when assessing current distortion at the PCC.

Keep reading

  • Selection guide

    Choosing a line reactor or motor reactor for a VFD

    Pick the side first. A line reactor goes between supply and drive input, sized on drive input amps at 3% impedance for transient buffering or 5% for harmonic reduction; a motor reactor goes between drive output and motor, sized on motor FLA, for reflected-wave protection on long shielded leads. The two are not interchangeable.

    Variable frequency drives

  • Troubleshooting

    Motor running over nameplate amps after soft starter bypass

    The motor averages 192.7 A against 164.0 A FLA with only about 0.90% voltage-magnitude unbalance, so the supply is not the proven cause. Treat excess shaft load or excess motor loss first, and confirm negative-sequence voltage with a three-phase power-quality analyzer reading V2/V1, total kW and speed before replacing the starter or the motor.

    Motors and motor starters

  • Explainer

    Motor starter vs contactor: using a starter as one

    A motor starter is a contactor with an overload relay added: the relay's normally closed contact opens the coil circuit on sustained overcurrent. A starter works wherever a contactor is called for, provided three checks pass — horsepower rating at the supply voltage, coil voltage, and the overload setting against the motor's nameplate full-load current. A NEMA Size 1 covers a 3 HP motor.

    Motors and motor starters