Hi Folks,

So I raised this problem quite some time ago with regard to a caravan camp site where the owner was having the local pitch MCB being replaced at random due to over heating of the neutral connection on the MCB.

At the time I could not explain what was going on.

Since then I have been involved in a more serious example of the same but devices were being burnt out, cables were melting and protective device were tripping randomly.

The answer/ cause.............................................Harmonics.......................yep the one every body hears about but sometimes do not realise its a real world issue.

My conclusion is that modern mobile caravan/ campers etc. use numerous dc/ac inverters and converters for their vehicle and camping needs and often LED source illumination and other devices.

I suspect that these items vary in quality and may well be producing harmonics which are resulting in sub distribution harmonics and excessive neutral current well beyond the rated values of the devices specified for the "normal" expected running loads.

There are some good inline filters that could be fitted which prevent such harmonic distortion and effects and I for one will be giving this aspect of electrical design more consideration in future.

I have kept it short but believe thats the problem.

Cheers

DMB

if so then clamp metering over the 3 phases in turn and then over the neutral will reveal all. 
The electronics problem is perhaps easier to visualize in the time domain (thinking voltage against time like a scope trace) than the frequency domain (spectral analysis).
Rectifiers feeding batteries and capacitors only conduct for the part of the time the AC waveform exceeds the battery voltage (or capacitor voltage), leading to a very discontinuous current concentrated near the crest of the sinewave.
This leads to a trapezoidal waveform (flat topped 'acme thread') as the voltage drop on the line only occurs when the current is flowing.

Now it is clear that 3 such pulses on the same wire cannot cancel in the way that 3 sinewaves offset 120 degrees in time would have done, as they are simply not overlapping, rather the neutral cable just sees current pulses like the live wires but 3 times more frequently with less time to cool in betwwen.

In extremis, with no resistive loads to cancel, the current on the neutral current can be the sum of the 3 individual phase currents.

Spectrally such current blips are composed of 3rd 9th harmonics etc but to describe it like that does not make the cause of problems quite so obvious.
Low pass filtering effectively  smears the short pulses into longer ones, so there is some time overlap, but it is not a perfect solution, and gets bulky at high currents (after all the pulse energy has to be stored in the Ls and Cs of the filter and released slowly over the cycle period - more energy needs a bigger bucket)

A better approach once you get beyond the main breakers is perhaps to treat the supplies as three single phases, each with a neutral the same size as the associated phase conductor  if you can.

If harmonic currents are flowing in a single phase circuit then there is no more risk to the neutral than there is to the live conductor,.

Mike

The answer/ cause.............................................Harmonics

But aren't the pitch hookups single phase?  I can see that in a 3-phase part of the system N currents fail to cancel out properly and so N can be higher than expected (max 1x the line current) - but in the single phase part of the system aren't the N currents going to equal the L current whatever it is (with harmonics or not). If the Ls aren't burning out, why should the Ns?

  - Andy.

The answer/ cause.............................................Harmonics

But aren't the pitch hookups single phase?  I can see that in a 3-phase part of the system N currents fail to cancel out properly and so N can be higher than expected (max 1x the line current) - but in the single phase part of the system aren't the N currents going to equal the L current whatever it is (with harmonics or not). If the Ls aren't burning out, why should the Ns?

  - Andy.