It sounds like we are converging on saying the same thing: you have to do the vector math. But either I am misunderstanding you, or you are not correct about the phasing of the current being supplied to the loads being discussed.

Consider a three phase wye system.

Call the neutral point our voltage zero reference.

Represent the voltage and phase of leg A relative to the neutral point by a vector of unit length and angle 0 degrees.

The voltage and phase of leg B relative to the neutral point will then be a vector of unit length and angle 120 degrees.

The voltage and phase of leg B relative to _leg A_ will then be represented by a vector of length sqrt(3) and angle 150 degrees.

For a resistive load connected between leg B and leg A, the current will follow the applied voltage. Thus the current flowing through this resistive load, from phase B to phase A will be sinusiodal and at phase 150 degrees relative to our reference phase voltage. The current flowing _out_ of phase B is at phase angle 150 degrees, the current flowing out of phase A is at 330 degrees relative to our reference zero.

The current flowing through the resistive load is in phase with the voltage difference between leg B and A, _not_ in phase with the phase to neutral voltage of either phase A or phase B.

If you look at this from the other side: consider a 120/208V wye system, and place a single phase 16 ohm resistive load between phase A and phase B. 13 amps will flow through this resistive load, and 2704W will be delivered to the load. But for 13A to flow through the load, 13A has to flow through both phase A and phase B. This means that 1560VA is being supplied by each phase. If 3120VA is being supplied, but only 2704W delivered to the load, then there has to be a power factor somewhere [Linked Image]

-Jon