The driving waveform suitable for a motor is determined by the shape of the back emf generated by it. For example, the methods shown in this thread are meant for true synchronous (open loop) operation of motors with sinusoidally-distributed stator windings, that often result in sinusoidal back-emfs.
http://www.vias.org/feee/c13_motors_08_02.html
http://www.ece.umn.edu/users/riaz/animations/3phasewdgs.html
However, the term BLDC has become very common these days and some motors labelled 'BLDC' might actually have sinusoidal back-emf like PMSMs. This maybe easily verified by running the machine as a generator at constant speed and checking its generated voltage waveform on an oscilloscope.
I've just tried testing the back emf of one of my Nidec motors, here's the result.
Combined with 'oilite' sleeve bearings, these motors do look pretty good.
Combined with 'oilite' sleeve bearings, these motors do look pretty good.
The phase currents have harmonics that originate from either the stator voltages or the back-emf. Therefore, if both these quantities are sinusoidal, the phase currents would also be sinusoidal (even without current control) and the torque ripple and (hence) the speed ripple would be negligible. The above motor is a good example of this situation.
However, many BLDCs have trapezoidal back-emfs and such motors are unlikely to significantly benefit from sinusoidal drive as the currents would still contain harmonics from the back-emf even if the stator voltages are made clean and perfectly sinusoidal.
However, many BLDCs have trapezoidal back-emfs and such motors are unlikely to significantly benefit from sinusoidal drive as the currents would still contain harmonics from the back-emf even if the stator voltages are made clean and perfectly sinusoidal.