What Size Neutral Wire Do You Need - Service and Feeder Rules

The neutral is the one conductor that is sometimes allowed to be smaller than the others, because it carries only the unbalanced share of the load. This guide works the 220.61 sizing rule, the service floor in 250.24(C), and the cases where a reduced neutral is not permitted at all.

Built to NEC 2023 tables. No reviewer claim until a licensed reviewer is signed.

The neutral carries only the unbalanced load

On a 120/240 V single-phase service, the two ungrounded conductors are opposite legs of one winding. A 240 V load draws on both and returns nothing on the neutral. A 120 V load draws on one leg and returns its current on the neutral, but a matching 120 V load on the other leg returns current in the opposite phase, and the two cancel in the neutral conductor itself. What the neutral actually carries, at any moment, is the difference between the legs: the unbalanced load. NEC 220.61 sizes the neutral for exactly that, the maximum unbalanced load, computed by applying the Article 220 demand factors to the 120 V loads and taking the worst credible imbalance between the legs. That is the whole legal basis for a smaller neutral, and its boundary: the neutral may be smaller precisely to the extent that the loads genuinely balance.

The service floor: 250.24(C)

The calculated unbalanced load is not the only constraint at a service. Under 250.24(C), the grounded conductor brought to the service equipment may not be smaller than the grounding electrode conductor that Table 250.66 requires for the service's ungrounded conductors, whatever the calculation says. The floor exists because the service neutral also serves as the fault return path to the source in the events that matter most, and the code will not let arithmetic shrink it below the grounding system's own scale.

Worked on the commonest case: a 200 A dwelling service on 4/0 aluminum ungrounded conductors. Table 250.66 puts that conductor size in the band requiring a grounding electrode conductor of 2 AWG copper or 1/0 aluminum, so the service neutral may not be smaller than 1/0 if it is aluminum, with 120 A of ampacity at 75 C. Most 200 A dwellings calculate an unbalanced neutral load below that figure, which means the floor, not the calculation, is what sizes most residential service neutrals, and why they so often land one size below the ungrounded conductors rather than at a calculated minimum. The grounding conductor size calculator works both 250.66 and the equipment grounding table beside it, and keeps the neutral's floor distinct from the two grounding conductors it is so often confused with.

The permitted reduction on larger services

Above the floor, 220.61(B) permits one further reduction, aimed at services large enough that perfect balance is a statistical certainty. For the portion of the calculated unbalanced load above 200 A, a demand factor of 70 percent applies. A service whose unbalanced load calculates to 260 A therefore sizes its neutral for 200 A in full plus 70 percent of the remaining 60 A: 242 A, which lands on 250 kcmil copper at 255 A in the 75 C column. The same section gives household cooking equipment and dryers their own neutral treatment, at 70 percent of their demand for the unbalanced figure, recognising that their 120 V components, motors, lights and controls, are a small and diversified share of the appliance. Both reductions are permissions with arithmetic attached, not habits: each must be traceable to a calculated load, which is what the service size calculator and the load calculation behind it are for.

When the neutral may not be smaller

Three situations withdraw the permission. The first is nonlinear load on a four-wire three-phase wye system. Switch-mode power supplies, LED drivers and variable-speed drives draw current in pulses that produce triplen harmonics, and harmonics in a shared neutral add instead of cancelling: the neutral of a wye system feeding mostly nonlinear load can carry more current than any phase conductor. Section 220.61(C) therefore prohibits applying any reduction to the portion of the load consisting of nonlinear loads on such a system, and data-heavy buildings routinely run full-size or oversized neutrals as the honest answer.

The second is the multiwire branch circuit. A shared neutral serving two or three ungrounded conductors is a working circuit conductor, sized with its circuit at full rating, not a service neutral eligible for a calculated reduction; its balance depends on how the receptacles happen to be loaded on a given day, and its failure mode, an open neutral putting 240 V across 120 V loads, is unforgiving. The third is any installation where the load is mostly single-leg 120 V load with little on the other leg to balance it: the unbalanced load then approaches the full leg current, and the calculation itself, honestly done, produces a full-size neutral. In all three cases the rule and the physics agree, which is usually a sign the rule was written from the physics.

Feeder neutrals, and the same-size question

A feeder neutral, to a sub-panel or a second building, is sized by the same 220.61 rule applied to the feeder's own unbalanced load, and it carries its own floor: under 215.2(B) it may not be smaller than the equipment grounding conductor that 250.122 requires for the feeder's breaker. It travels insulated with the feeder conductors and lands on the isolated neutral bar at the sub-panel, never bonded there. In practice, feeder neutrals are frequently run at the same size as the ungrounded conductors anyway, because the calculated saving is one conductor size against the cost of certainty about future loads, and because a panel whose loads change over decades should not have its neutral frozen at today's balance.

Which answers the question this guide is titled with, in its commonest form: does the neutral need to be the same size as the other conductors? No, not by default, and not by habit either. It needs to be the size produced by three checks in order: the calculated unbalanced load under 220.61, the floor that applies to its position in the system, and the prohibitions that withdraw the reduction for nonlinear and multiwire cases. Run all three and the answer is sometimes full size, often one size down at a dwelling service, and never a guess. The wire size calculator checks the ampacity side of whichever size the three checks produce, with derating and voltage drop applied to the run.

Planning only

This guide is a planning reference built on NEC 2023 methods. Neutral sizing on a real service or feeder follows the calculated load for that installation, local amendments vary, and the final selection belongs with your authority having jurisdiction and your licensed electrician.