Bjarney,
I haven't actually found any cases where the neutral current was high enough to cause any problems. I haven't even talked with anyone who has personally observed such conditions. I spoke at length with a design engineer for a firm that does mostly commercial office buildings, and he told me that he never uses any of the oversized neutrals or K rated transformers and has never found any problems in the system after the client has moved in and went into full operation. Some of these installations that he told me about had over 300 computers plus electronic ballast all on 120/208 wye systems. The was an article 3 or 4 years ago that suggested that where there are very large numbers of switching power supplies on the system that the neutral current actually goes down. The reason that was cited was that the harmonics are created randomly and where there are large numbers of generators, the random timing actually results in harmonics that do cancel.
But back to your installation, isn't the neutral current in a 3 wire system fed from a 4 wire system always at least equal to the highest ungrounded conductor current? The formula is the square root of (A^2 + B^2 + C^2)-((A x B)+ (A x C) + (B x C)). A^2 = A squared. If you use a 10 amp load for any 2 hots and solve you will have 10 amps on the grounded conductor.
Don
