|> And now for the tough question. What if one side has a very inductive |> load of 20A (power factor 0.25) and the other side has a very capacitive |> load of 20A (power factor 0.25) ... now what is the current on the common |> wire? |>
| | Well..... Since you asked.... | | For the inductive load, 20A @ 0.25 means there is 5A of 'real' current and | sqrt(20^2-5^2)=19.4 A of lagging reactive current. Similarly, for the | capacitive load you would have 5A of 'real' current and sqrt(20^2-5^2)=19.4A | of leading reactive current. | | But the two supplies are '180-out' from each other. So the 5A real current | in one load will exactly cancel the 5A real current in the other load. | Unfortunately, because of the 180 shift, the lagging reactive current will | not cancel the leading reactive current, instead they will actually add. So | you'll end up with about 38.8 A of imaginary current in the neutral, | possibly overloading it.
Bingo.
| While this does overload the neutral, it's such a remote/rare possibility,
Of course. Anyone stupid enough to hook up "reatively unbalanced" loads like that deserves what they get :-)
| that it is seldom worried about. Another 'neutral danger' is when running | high harmonic loads on each phase of a wye-connected three-phase system. | The harmonics from each single phase load will not cancel perfectly at the | neutral and can lead to more current in the neutral than in any of the three | phase conductors.
Delta can even potentially see a 15% current increase from such loads. That's not something you'd likely worry too much about, though.
And even single phase is not completely immune. If the loads on each side of the neutral are designed enough differently that they conduct at different times, such as one conducting before peak and the other conducting after peak, you can see this kind of overload even for single phase.
And then there is a potential issue of current peaking causing extra heating on conductors due to the I^2*R losses. Drawing 400 amps at 50% duty cycle heats the wiring more than drawing 200 amps at 100% duty cycle. Where harmonics raise the current peaks beyond what they would normally be in a sinusoidal current, there will be more heating in all the conductors. I would not expect it to be that big a deal, but it could crop up in an unexpected place some day (like somone thinking that
48 amps of lights at 25% on time for a big sign averages out to work on a 15 amp circuit and wonders why the breaker trips eventually).