One thing I don't understand is it's said that places with lots of electronic SMPSUs like data centres and that type of thing can have higher currents in the neutral than the phases how is this possible? I can't really understand it hope someone can explain it
only a problem in systms with more than one phase.
in a 3 phase system with 3 similar resistive loads (orr indutive or capacitive, so long as the current waveforms are sinewaves), the 3 currents in the neutral sum more or less to zero.
But if we have anything with a bridge rectifier feeding a capacitor or a battery, the current waveform is far from sinusoidal, but takes the form of no current at all while the mains is lower than that voltage on the capacitor, perhaps until 3/4 of the way up to peak, then a large current for the top bit of the waveform, while the wave is forward biassing the diodes, and then falling back to no current at all once the voltage of the incoming wave is less than the DC again.
3 such waveforms of short duration pulses offset by 1/3 of a cycle do not even overlap in time, let alone cancel, so when each live carries 2 pulses per cycle period, and the neutrral carries 6. = 3 times as much heat.
In reality it is rarely quite as bad as this, as some efforts are made to make the current waveform a bit more sinusoidal ('power factor correction' is not always just about phase shifts, it can be about stretching out short pulses as well)
The offender in this case is generally switched mode power supplies introducing 3rd order harmonics (i.e.150Hz)**. As Mike says the three phase currents are (approximately) balanced and sum to zero in the neutral but if you look at three phases of 150Hz offset by 8.33 ms in each phase then the peaks always coincide so they are additive in the neutral. The problem was so bad in the early days of office blocks (less filtering of the harmonics combined with computers requiring more power) that Terasaki actually produced a four pole circuit breaker with the neutral oversized compared to the phases, specifically to cater for this problem. These days the problem is better understood and such measures are unlikely to be needed.
Alasdair
**The order of the main harmonic is 2*n ± 1 times the fundamental frequency so for a single phase switched mode supply the biggest harmonic will be at 3 times, while for a three phase it will be at 5 and 7 times.
To add to mapj1 explanation, the worst case for 3rd harmonics adding is when the SMPSUs are all running at about 60% of rated load - which is quite normal.
I once worked on a project for a data centre where every Distribution Board had a 1:1 delta / star transformer built in to remove the third harmonic and then had a oversized neutral in the generator as well.
