Neutral Current in Unbalanced 3-Phase Systems (and Why Harmonics Make It Worse)

Quick answer: In a balanced three-phase, four-wire system the neutral current is zero because the three phase currents are 120° apart and cancel. When the loads are unbalanced, the neutral carries the vector difference. For linear loads use the formula I_N = √(I_a² + I_b² + I_c² − I_aI_b − I_bI_c − I_cI_a). With non-linear loads (computers, LED drivers, VFDs) third-harmonic currents add in the neutral and can push neutral current above the phase current — which is why the NEC now treats the neutral as a current-carrying conductor in many cases. Check your worst case with the neutral current calculator.

Why a balanced system has zero neutral current

A three-phase, four-wire wye system has three "hot" conductors and one neutral. The three phase voltages — and, for equal resistive loads, the three phase currents — are each separated by 120° in time. At any instant, when one phase is at its positive peak, the other two are at negative fractions, and the three add up to zero. The neutral is the return path, so if the three returns cancel, the neutral carries nothing.

That is the ideal. It is also why the old rule — "the neutral only carries the unbalance, so you can undersize it" — was allowed for decades. But modern buildings are full of loads that break both assumptions: they are unequally distributed and they draw current in sharp non-sinusoidal pulses. Both conditions put real current on the neutral, and the second one can put a lot.

The linear unbalanced case: vector math, not simple subtraction

When the loads are linear (motors, incandescent lighting, resistive heat) but unequal across the three phases, the neutral carries the vector sum of the three phase currents. Because the currents are 120° apart, you cannot just subtract them arithmetically — you have to add them as vectors. The result simplifies to a clean formula:

I_N = √(I_a² + I_b² + I_c² − I_a·I_b − I_b·I_c − I_c·I_a)

where I_a, I_b, I_c are the three phase (line) currents in amps. A few things fall straight out of this equation:

• If all three are equal (I_a = I_b = I_c), the expression under the root becomes zero — zero neutral current, confirming the balanced case.
• If only one phase carries current (I_b = I_c = 0), the formula reduces to I_N = I_a — the neutral carries the full phase current, exactly as a single-phase circuit would.
• For any partial unbalance, the neutral current lands somewhere between zero and the largest phase current.

Worked example: unbalanced lighting panel

A 208Y/120 V lighting panel feeds three single-phase branch groups:

• Phase A: 80 A
• Phase B: 60 A
• Phase C: 40 A

Plugging in:

I_N = √(80² + 60² + 40² − 80×60 − 60×40 − 40×80)
= √(6,400 + 3,600 + 1,600 − 4,800 − 2,400 − 3,200)
= √(11,600 − 10,400) = √1,200 ≈ 34.6 A

So even though phase A carries 80 A, the neutral only sees about 35 A for this linear load. Notice how much less that is than the naive "80 − 40 = 40 A" guess — the vector math matters. The neutral current calculator does this instantly and also reports how badly the panel is unbalanced so you can rebalance the branch circuits.

The harmonic case: where the neutral turns dangerous

Here is what the linear formula misses entirely. Non-linear loads — switch-mode power supplies in computers and servers, LED and fluorescent drivers, variable-frequency drives, UPS rectifiers — draw current in narrow pulses near the voltage peak, not as smooth sine waves. That pulsed current is rich in harmonics, and the most troublesome is the 3rd harmonic (180 Hz on a 60 Hz system).

The 3rd harmonic and its odd multiples (9th, 15th — the "triplen" harmonics) behave very differently from the fundamental. On the three phases, the fundamental currents are 120° apart and cancel in the neutral. But the triplen harmonics on all three phases are in phase with each other — so instead of cancelling, they add arithmetically in the neutral. Three phases each carrying a 3rd-harmonic component send that component into the neutral three times over.

The consequence is counter-intuitive and hazardous: in a heavily non-linear system with balanced phase currents, the neutral can carry up to 1.73× (and in extreme cases even more than) the phase current. A neutral sized the same as the phase conductor — or worse, a shared/undersized neutral from an older design — overheats. Because the neutral has no overcurrent protection of its own, this overheating is silent until insulation fails.

Why this changed the NEC

Because of triplen harmonics, the National Electrical Code (NEC 310.15(E)) now states that on a 4-wire wye circuit where the major portion of the load is non-linear, the neutral is a current-carrying conductor and must be counted when applying conductor ampacity adjustment (derating) factors. In practice, designers serving data centers, office IT loads and LED-heavy installations frequently full-size or oversize the neutral — a common approach is a 200% neutral (double the phase conductor's capacity) for known high-harmonic feeders.

Neutral current at a glance

How to control neutral current in design

• Balance the phases. Distribute single-phase branch loads as evenly as possible across A, B and C at every panel. Good balancing is the cheapest fix and reduces neutral current, voltage unbalance and losses all at once.
• Size the neutral for the real load. On feeders serving mostly non-linear loads, do not undersize the neutral. Treat it as current-carrying per NEC 310.15(E) and consider a 200% neutral for data-center and IT-heavy feeders.
• Watch voltage unbalance on motors. Even a few percent of voltage unbalance causes disproportionate motor heating — NEMA guidance is to keep voltage unbalance under 1% and derate motors above it.
• Consider harmonic mitigation. Harmonic-mitigating transformers, K-rated transformers, zig-zag grounding transformers, or active harmonic filters reduce the triplen content reaching the neutral and upstream equipment.
• Don't share neutrals on non-linear circuits. Multi-wire branch circuits that share one neutral can overload it under harmonics; give non-linear circuits dedicated neutrals.

Why voltage unbalance matters as much as current

Unbalanced loading does not just heat the neutral — it distorts the phase voltages. When one phase is heavily loaded, its voltage sags relative to the others, producing voltage unbalance. For three-phase motors this is especially punishing: a small percentage of voltage unbalance produces a much larger percentage of current unbalance and a sharp rise in winding temperature, shortening motor life. This is why balancing panels is not just an efficiency nicety — it protects every three-phase motor downstream. The percent unbalance is defined as the maximum deviation of any phase from the average, divided by the average, times 100.

Common mistakes with neutral sizing

• Subtracting phase currents arithmetically. The neutral current is a vector result — use the √ formula, not simple subtraction.
• Assuming a balanced system needs no neutral capacity. True only for linear loads. Under harmonics, a perfectly balanced system can still overload the neutral.
• Undersizing the neutral on IT/LED feeders. The old "half-size neutral" habit is dangerous for non-linear loads and non-compliant where the major load is non-linear.
• Ignoring the neutral in ampacity derating. When the neutral counts as a current-carrying conductor, it changes the number of conductors used in the adjustment factor.
• Balancing on connected kVA instead of measured current. Nameplate balance and real, diversified operating current differ — measure or estimate the actual load.

Delta systems and the three-wire case

Everything above assumes a four-wire wye system with a neutral, which is where neutral-current problems live. It is worth noting the contrast: a three-wire delta system — common for three-phase motor loads — has no neutral at all, so there is no neutral conductor to overload and triplen harmonics circulate within the delta winding rather than flowing on a shared return. This is one reason large motor loads are often served delta while mixed single-phase lighting and receptacle loads, which need a neutral for 120 V, are served from a wye. When a facility mixes both, the designer chooses the transformer configuration deliberately: a delta-wye transformer gives the wye secondary that feeds the 120/208 V panels, and the delta primary conveniently traps much of the triplen harmonic current so it does not propagate upstream. Understanding which loads sit on wye versus delta tells you immediately where to look for neutral-current and harmonic issues — they concentrate on the wye, four-wire branches serving single-phase electronic loads, not on the three-wire delta motor feeders.

Standards and references

A harmonic worked example: the neutral that overheats

Numbers make the harmonic problem concrete. Consider a 208Y/120 V feeder to an office floor packed with computers, each drawing current that is roughly 80% third-harmonic content — a realistic figure for older switch-mode supplies. Suppose the three phases are balanced, each carrying 100 A of fundamental current.

• Fundamental in the neutral: balanced, so it cancels — 0 A.
• Third-harmonic per phase: about 80% of 100 A = 80 A on each phase.
• Third-harmonic in the neutral: the triplen components are in phase, so they add: 80 + 80 + 80 = 240 A of third-harmonic current in the neutral.

The phase conductors each carry roughly √(100² + 80²) ≈ 128 A, while the neutral carries about 240 A — nearly 1.9× the phase current. A neutral sized equal to a 128 A phase conductor is now catastrophically overloaded, and because there is no breaker in the neutral, nothing trips. It simply runs hot until the insulation cooks. This is the exact scenario that drove NEC 310.15(E) and the widespread use of oversized neutrals in IT spaces. Modern power supplies with power-factor correction have far less third harmonic, which has eased the problem, but legacy equipment, cheap LED drivers and large drive populations keep it alive.

Measuring neutral current in the field

You do not have to model harmonics to catch this problem — you can measure it. A true-RMS clamp meter on the neutral of a suspect feeder tells you immediately whether the neutral is carrying more than you expect. Two field rules:

• Use a true-RMS meter. An averaging meter reads distorted, harmonic-rich current incorrectly — often low — hiding the very overload you are hunting.
• Compare neutral to phase. If the neutral current approaches or exceeds the phase current on a feeder that should be reasonably balanced, you have a harmonic problem, not just an unbalance problem, and the neutral and upstream transformer need a closer look.

A warm or hot neutral bus, discolored terminations, or a transformer that runs hotter than its loading suggests are all field symptoms of triplen harmonics loading the neutral. The neutral current calculator gives you the linear baseline to compare a measurement against — if the meter reads far above the calculated linear value, harmonics are the difference.

The neutral is not the ground — and why that matters here

A frequent source of confusion is treating the neutral and the equipment grounding conductor as interchangeable. They are not. The neutral (grounded conductor) is a normal current-carrying conductor — it is supposed to carry the return current discussed throughout this article. The equipment grounding conductor carries current only during a fault. They are bonded together at exactly one point (the service), and keeping them separate everywhere else is what keeps normal neutral current off the grounding system and the metal enclosures people touch.

This distinction matters for neutral-current design because a mis-wired system that lets neutral current flow on the grounding conductor or on building steel — a "shared" or improperly bonded neutral — creates objectionable current (NEC 250.6), causing stray voltages, noise on sensitive equipment, and overheating in paths never meant to carry current. When you oversize a neutral for harmonics, you must still keep that neutral isolated from ground except at the service, or the harmonic current finds its way onto the grounding system.

Shared neutrals and multi-wire branch circuits

A multi-wire branch circuit (MWBC) shares one neutral among two or three phase conductors to save copper. On linear loads this is efficient — the shared neutral carries only the unbalance. But two hazards apply:

• Harmonics defeat the savings. On non-linear loads the triplen harmonics add in the shared neutral just as they do in a feeder neutral, so a shared neutral on three phases of computer loads can be overloaded even though each phase looks modest. Give non-linear circuits dedicated, full-size neutrals.
• Open-neutral danger. If a shared neutral opens (a loose terminal), the single-phase loads on the two remaining phases end up in series across 208 V, and the voltage divides according to their impedance — lightly loaded equipment can see well over 120 V and be destroyed. NEC 210.4 requires simultaneous disconnection of all phases of an MWBC partly for this reason.

Practical design checklist

Harmonics also cook the transformer — the K-factor

The neutral is not the only casualty of triplen harmonics; the upstream transformer suffers too. Harmonic currents circulate in the delta primary of a delta-wye transformer and drive up eddy-current losses, which rise roughly with the square of the harmonic frequency — so a modest amount of high-order harmonic current produces a disproportionate amount of extra heating in the windings. A standard transformer serving heavy non-linear load can overheat even when its kVA loading looks conservative, because the losses it was rated for assumed a clean sine wave.

The industry answer is the K-factor transformer, built with oversized neutrals, transposed or stranded windings, and extra thermal margin to tolerate a defined level of harmonic loading. Common ratings are K-4, K-13 and K-20, chosen from the expected harmonic spectrum of the load — K-13 is typical for office and IT spaces, K-20 for very heavy non-linear concentrations. Specifying a K-rated transformer, oversizing the neutral, and keeping THD within IEEE 519 limits are the three coordinated moves that keep a harmonic-rich system from slowly baking its own distribution equipment. Treating the neutral in isolation while ignoring the transformer just moves the hot spot upstream.

Why balancing pays off three ways

Balancing the phases is the cheapest and highest-leverage action in this whole topic because it improves three things at once. First, it minimizes neutral current from unbalance, keeping the neutral cool. Second, it minimizes voltage unbalance, protecting every three-phase motor from the disproportionate heating that even a few percent of unbalance causes. Third, it minimizes I²R losses in the feeder, because losses are lowest when current is spread evenly rather than concentrated on one hot phase. A panel schedule that looks balanced on paper but was never verified against real, diversified operating current is a common gap — the connected kVA can be balanced while the actual running current is badly skewed, so balance on measured or realistically estimated load, not just nameplate.

The bottom line

Neutral current is zero only in the textbook case of balanced linear loads. In real buildings you face two departures: unbalance, handled by the vector formula, and harmonics, which make the neutral carry more than the phases. Size the neutral for the worse of the two — and on IT-, LED- and drive-heavy feeders, that almost always means a full-size or oversized neutral treated as current-carrying under the NEC. Run your worst-case numbers through the neutral current calculator, then confirm the feeder and derating with a licensed electrical engineer against NEC 310.15(E) and IEEE 519.

Frequently asked questions

How do you calculate neutral current in a 3-phase system?

For linear loads, the neutral carries the vector sum of the three phase currents, which simplifies to IN = square root of (Ia² + Ib² + Ic² − Ia·Ib − Ib·Ic − Ic·Ia), where Ia, Ib and Ic are the phase currents in amps. If the three currents are equal the result is zero, confirming that a balanced linear system has no neutral current; if only one phase is loaded the neutral carries the full phase current.

Why is neutral current zero in a balanced system?

Because the three phase currents are equal in magnitude and separated by 120 degrees in time, so at every instant they add up to zero. The neutral is the shared return path, and when the three returns cancel there is nothing left for the neutral to carry. This holds only for balanced linear loads — unbalance or harmonics break the cancellation and put real current on the neutral.

Can neutral current be higher than phase current?

Yes, with non-linear loads. Switch-mode power supplies, LED and fluorescent drivers, and VFDs draw pulsed current rich in third-harmonic (triplen) content. Triplen harmonics on all three phases are in phase with each other, so instead of cancelling in the neutral they add arithmetically. In heavily non-linear systems the neutral can carry up to 1.73 times the phase current, or even more, even when the phase currents themselves are balanced.

What are triplen harmonics?

Triplen harmonics are the third harmonic and its odd multiples — the 3rd, 9th and 15th. Unlike the fundamental and most other harmonics, the triplen components on the three phases of a wye system are in phase with each other rather than 120 degrees apart. That means they do not cancel in the neutral; they add, which is why non-linear loads can overload a neutral that carries almost no current from the fundamental.

Does the NEC require a full-size neutral?

NEC 310.15(E) states that on a 4-wire wye circuit where the major portion of the load is non-linear, the neutral is a current-carrying conductor and must be counted when applying conductor ampacity adjustment factors. In practice designers full-size or oversize the neutral on feeders serving heavy non-linear loads — a 200% neutral is common for data-center and IT-heavy feeders — because the triplen harmonics can make the neutral current exceed the phase current.

How do I reduce neutral current?

Balance the single-phase branch loads as evenly as possible across the three phases, which reduces neutral current, voltage unbalance and losses at once. For non-linear loads, do not undersize the neutral, avoid shared neutrals on those circuits, and consider harmonic mitigation such as K-rated or harmonic-mitigating transformers, zig-zag grounding transformers, or active harmonic filters to reduce the triplen content reaching the neutral.

Why does load unbalance matter for motors?

Because unbalanced loading distorts the phase voltages, and three-phase motors are very sensitive to voltage unbalance. A small percentage of voltage unbalance produces a much larger percentage of current unbalance and a sharp rise in winding temperature, shortening motor life. NEMA guidance is to keep voltage unbalance under about 1% and to derate motors above it, which is why balancing panels protects every downstream motor.

Is a neutral current calculator accurate for design?

The vector formula it uses is exact for linear loads and gives the correct neutral current from the three phase currents. For non-linear loads you must add the harmonic contribution, which depends on the load's harmonic spectrum, so treat the linear result as a floor and size the neutral for the harmonic case per NEC 310.15(E) and IEEE 519. Confirm the feeder, neutral and derating with a licensed electrical engineer.