I've researched and have a working prototype of a controller witch drives a three phase motor with Field Oriented Controll. FOC allways keeps the magnetic field in the motor 90degrees compared to the magnets - to ensure that all force go to make spinning force and not waste any on linear. FOC reduces the torque ripple to virtually none and also reduces the decoupling needed, as the current ripple is reduced.
Recently I teached our newest group member about FOC, and he basically said; "Why not apply this to two phases instead of three?" And so my head started thinking, a lot. I've tried to compare the techonogies, and this is my conclusion:
A three phase motor requires six steps to complete one electical rotation. A two phase motor requires four steps. If you try to fit these controll schemes, square and six edge, into a circle - the six edge fits best and so gives the best "circle follower". Mix FOC into the picture and everything changes. With FOC you can look at the controll scheme as a circle - for both the three phase and two phase motor.
I tried thinking about inverters and realized that I could run the motor with two class D amplifiers, based on one half bridge each - with negative bias. Four fets instead of six could be used (ofc with comparably higher current ratings). Less gate drivers, less current sensors, less processing power to do the vector controll. Basically we could slash the controller cost by 20-30%, possibly more. The motor would still need three phase wires tho; Gnd, phase A and Phase B.
Two phase BLDC is very new to me and I would appreciate it if someone could point out the, not so to me, obvious flaw I've omitted.
Recently I teached our newest group member about FOC, and he basically said; "Why not apply this to two phases instead of three?" And so my head started thinking, a lot. I've tried to compare the techonogies, and this is my conclusion:
A three phase motor requires six steps to complete one electical rotation. A two phase motor requires four steps. If you try to fit these controll schemes, square and six edge, into a circle - the six edge fits best and so gives the best "circle follower". Mix FOC into the picture and everything changes. With FOC you can look at the controll scheme as a circle - for both the three phase and two phase motor.
I tried thinking about inverters and realized that I could run the motor with two class D amplifiers, based on one half bridge each - with negative bias. Four fets instead of six could be used (ofc with comparably higher current ratings). Less gate drivers, less current sensors, less processing power to do the vector controll. Basically we could slash the controller cost by 20-30%, possibly more. The motor would still need three phase wires tho; Gnd, phase A and Phase B.
Two phase BLDC is very new to me and I would appreciate it if someone could point out the, not so to me, obvious flaw I've omitted.
I don't quite know what to think of the idea myself yet. The most tempting thing is actually the lessened strain on the uC having x-y as the two directions.
