If you start looking at the physics of ungrounded systems you might find the L-L voltage might actually be a too low number, the maximum value of some could be 480*1.732 or 832 volts or even more.I understand that the NEC defines the "Voltage to Ground" to be equal to the L-L voltage for ungrounded systems, but physics-wise it's undefined.
Look at the eaton paper I linked above in post 37 and this one:
