On paper, a balanced three-phase load returns zero current through the neutral. In practice, the neutral almost always carries something — and it can come from two genuinely different mechanisms that behave nothing alike.
In a perfectly balanced three-phase four-wire system with purely linear loads, the three phase currents are equal in magnitude and evenly spaced 120° apart — and their vector sum is exactly zero. That's the textbook answer to "what does the neutral carry," and it's correct as far as it goes. Real systems, though, almost never sit at that ideal. Neutral current shows up for two distinct reasons that get conflated constantly: plain load imbalance, and harmonic distortion from nonlinear loads. One behaves exactly the way intuition suggests. The other doesn't — see the harmonic-current mechanism below for why. For the full three-phase background, see the three-phase power systems guide.
Apply Kirchhoff's current law at the neutral node of a wye-connected system: the neutral current is the vector sum of the three phase currents, In = Ia + Ib + Ic. If Ia, Ib, and Ic are equal-magnitude sinusoids spaced exactly 120° apart, that vector sum collapses to zero — draw them tip-to-tail and they close into a perfect triangle, returning right back to where they started. Unbalance the magnitudes even slightly and the triangle stops closing. Whatever gap is left between the end of the last vector and the starting point is the actual neutral current — a residual, not a cancellation failure.
Unbalance the load — a heavier single-phase circuit on phase A than on B or C, for example — and the phasor chain no longer closes. The gap left between the last vector's tip and the starting point is a real, physical current that has to flow somewhere, and the neutral conductor is what carries it.
There's nothing mysterious happening at the neutral point — it's a plain application of Kirchhoff's current law at a node with four conductors meeting. Whatever the three phase currents don't manage to cancel among themselves has to flow through the fourth conductor to balance the node, and that fourth conductor is the neutral. When the three phase currents are equal-magnitude sinusoids 120° apart, they cancel completely and the neutral is idle. The moment that symmetry breaks — even a little — the leftover has to go somewhere, and it's directly proportional to how far off-balance the phases are. That's why neutral current from imbalance is so predictable: it scales cleanly with the size of the mismatch between phases.
True as far as load imbalance goes — but incomplete, and the missing half is the surprising one. Harmonic distortion from nonlinear loads is a second, independent cause of neutral current, and it doesn't need any imbalance at all. A perfectly balanced load of electronic ballasts, switch-mode power supplies, or LED drivers can produce a neutral current that isn't just nonzero — it can actually exceed any individual phase current, purely from harmonic content. That mechanism has nothing to do with the vector-cancellation story above; it's covered in harmonics and triplen neutral overload, and it's exactly why neutral conductor sizing can't stop at checking for load balance.
Explains why a perfectly balanced three-phase load theoretically returns zero neutral current, and why real neutral current comes from two genuinely different causes — simple load imbalance, and harmonic distortion from nonlinear loads — that behave in very different ways and matter for very different reasons.
It's easy to treat neutral current as a single phenomenon that simply tracks how unbalanced a system is. That's true for one of its two causes but not the other. Load imbalance produces neutral current proportional to the mismatch between phases, and vanishes entirely on a balanced load. Harmonic distortion from nonlinear loads is a completely separate mechanism that can produce large neutral current even when the fundamental-frequency load is perfectly balanced, which is why neutral conductor sizing has to consider both causes independently rather than assuming a balanced system means an idle neutral.
At the neutral node, Kirchhoff's current law requires In = Ia + Ib + Ic. For equal-magnitude sinusoidal phase currents spaced 120° apart, that vector sum is identically zero — drawn tip to tail, the three phasors close into an equilateral triangle. Unbalancing the magnitudes (while keeping the 120° spacing) breaks the closure, and the residual gap between the end of the phasor chain and its starting point is the actual neutral current, scaling directly with the size of the imbalance.
Neutral conductor sizing has traditionally focused on load imbalance, since that's the intuitive and easily measured cause. But a facility with heavy imbalance-free but harmonic-rich loading — electronic ballasts, switch-mode supplies, VFDs, LED drivers — can still see substantial, even excessive, neutral current from the second mechanism. Distinguishing which cause is driving an elevated neutral reading determines the fix: rebalancing single-phase loads across phases addresses imbalance-driven neutral current, but does nothing for harmonic-driven neutral current, which instead calls for oversized neutral conductors or harmonic mitigation at the source.
Only as an idealization. Real single-phase loads (lighting circuits, receptacles, small motors) are rarely distributed with perfect precision across three phases, and most modern facilities carry at least some nonlinear load, so some neutral current is normal. The goal of good design is keeping it within the neutral conductor's rating, not eliminating it entirely.
Yes, for the imbalance mechanism specifically — neutral current from load imbalance scales directly with how far apart the phase magnitudes are, which is why panel balancing (redistributing single-phase circuits across phases) is a standard, effective fix for that cause.
Yes — that is exactly what makes it counterintuitive. If the load on each phase is identical in both magnitude and harmonic content, the fundamental-frequency components cancel at the neutral just as the balanced case here shows, but certain harmonic components do not cancel the same way. See the companion explainer on harmonics and triplen neutral overload for the mechanism.
A true RMS clamp meter reading on the neutral compared against the phase currents is a start, but distinguishing the two causes usually requires a harmonic-capable meter: if the neutral current tracks the arithmetic imbalance between phases, it is load-imbalance-driven; if it is elevated even when the phases look balanced, harmonic content — particularly triplen harmonics — is the more likely cause.
Try our Electrical Studio
More calculators, simulators, and guides for this discipline.