Neutral Current Exceeding Phase Current in an Office Submain
A three-phase four-wire TN-S submain feeds thirty workstation clusters in an open-plan office at 400/230 V. Both the line conductors and the neutral are 70 mm² copper, protected upstream by a 160 A moulded-case circuit breaker. The load is almost entirely switched-mode power supplies, LED drivers and desktop IT. A power quality survey at the board records true RMS phase currents of 62 A, 58 A and 65 A, and a neutral current of 118 A. The neutral bar shows discoloration and the insulation on the neutral tail at the lug is brittle. A clamp on the main earthing conductor reads 0.4 A. The building operator says the phases are balanced within 10% and concludes that the neutral reading must be a meter error.
1. Explain how the neutral current can be almost twice the highest phase current, and why the balance argument is misleading.
Read the worked answer
Why the fundamentals cancel but the harmonics do not
- In a balanced four-wire system the fundamental currents of the three phases are displaced by 120° and sum to zero in the neutral. That is the reasoning the operator is applying, and for a linear load it would be correct.
- Switched-mode power supplies draw a quasi-square input current rich in odd harmonics. The triplen harmonics - 3rd, 9th, 15th and so on - are in phase with each other in all three phases, so instead of cancelling they add arithmetically in the neutral.
- With roughly 60 A of fundamental per phase and around 40% third-harmonic content, each phase contributes about 25 A of third harmonic; three of them sum to about 75 A in the neutral, before the ninth and higher triplens and any residual fundamental unbalance are added. That accounts for a measured 118 A.
Consequences
- The neutral is the same 70 mm² as the phases and is not protected by the 160 A breaker against overload. It carries more current than any phase, heats up, and cooks its terminals - exactly the discoloured bar and brittle tail insulation observed.
- 'Balanced within 10%' describes the fundamentals only and says nothing about the neutral, which is driven by waveform shape rather than by phase balance.
- The 0.4 A on the earthing conductor is consistent with normal capacitive leakage and does not indicate a neutral-earth fault, but it should still be excluded formally.
2. What measurements would confirm the diagnosis, and what checks would you use to rule out a fault rather than a load characteristic?
Read the worked answer
Confirming the harmonic explanation
- Harmonic analysis on all four conductors to at least the 25th order, recording individual harmonic magnitudes, phase angles and THD per phase and on the neutral. The decisive evidence is a large third-harmonic component on each phase, in phase with each other, summing in the neutral.
- Capture the neutral current waveform rather than only its RMS value; a peaked waveform with three peaks per cycle is the signature of triplen summation.
- Log over a full working day to show that the neutral current tracks occupancy and IT load, which is characteristic of the load and not of a defect.
Ruling out a fault
- Insulation resistance neutral to earth with the main switch off, and observation of the earth conductor current with all loads off and then on; a step change with load points to a neutral-earth contact rather than to harmonics.
- Check for parallel neutral paths, borrowed neutrals and any PEN arrangement in the building, which can make a clamp read a current that is not the circuit's own.
- Verify neutral continuity and terminal tightness; a loose neutral produces voltage unbalance and heat, and can mimic or worsen the reading.
- Thermal imaging of the neutral bar and lugs under load to quantify the overheating and to record the condition before remedial work.
- Confirm the instrument's true-RMS capability and that the current probe is rated for the harmonic frequencies present before accepting any number as evidence.
3. Recommend the remedial measures and the design rules to apply when the submain is replaced or upgraded.
Read the worked answer
Immediate measures
- Treat the discoloured neutral bar and brittle tail as a fire risk: reduce or redistribute load, re-terminate with correctly rated lugs, and monitor with thermal imaging until the permanent fix is in.
- Do not simply increase the 160 A breaker; the neutral has no overload protection and the problem is thermal, not overcurrent.
Design rules for the replacement
- Where triplen harmonic content is significant, the neutral must be treated as a current-carrying conductor. It should be sized at least equal to the phase conductors, and in practice a 200% neutral is the robust answer for an IT-heavy submain of this kind.
- Apply a harmonic derating factor to the phase conductors as well, since the tabulated capacities assume sinusoidal current, and reflect it in the Ib ≤ In ≤ Iz check.
- Avoid shared neutrals on multiwire branch circuits serving nonlinear loads; give each cluster its own neutral so triplen currents cannot accumulate in a common conductor.
- Use four-pole or switched-neutral devices so the neutral is opened and isolated for maintenance.
- At the source, specify power supplies with active power factor correction, and where a transformer serves this load, specify one rated for harmonic duty (K-rated) rather than a standard unit.
- Provide neutral overcurrent detection or thermal monitoring on the board, and record the harmonic survey and the design basis in the as-built documentation so the next designer does not repeat the original assumption.