2-phase motor / controller idea any good?

[Lebowski]

I've been wondering about a new motor/controller idea I have...
Most (all?) of our BLDC motors are 3 phase motors. 3-phase is used in the mains power
distribution network because you don't need a return wire for the ground, in essence
the currents in the 3 phases cancel each other out such that no current is flowing
in the ground wire... which is therefore omitted.
But for our motors this doesn't really matter.... right ?
Would a two-phase motor not make more sence ? Ok, is has 4 wires instead of 3 but
this is no biggie. More interesting is that one phase follows a sine while the other
phase follows a cosine funtion (versus time). If you power the motor by a controller
that implements a negative resistor the currents will also be sine and cosine.
Power from 1 winding will be sin^2 while from the other will be cos^2. Which means
power delivery is constant over time without any ripple.
Also, even more interesting, power consumption will also be constant meaning that the
engine/controller will drawn a constant current (assuming constant battery voltage) from
the battery. Lithium batteries like a constant current much better than a pulsed one. Also
there would be less influence from wiring inductance...

what do you guys think ?

[rhitee05]

The main reason most motors are 3-phase rather than 2-phase is self-starting. A 2-phase motor can end up in a situation where it will either not self-start or will rotate in the wrong direction, whereas neither of those can happen to a 3-phase. If you look at a something with a small 2-phase AC motor, like a window fan, the motor usually has a second set of poles with a "shaded pole" configuration or sometimes a small capacitor. This has the effect of making the motor start like a 4-phase motor but run as a 2-phase. I think compressors (AC, refrigerator, etc) usually use capacitor-start motors.

There's nothing sacred about only 3 phases. The 4-wire motor you're speaking of would be more properly considered a 4-phase motor, I think. I've heard of 5- and 6-phase motors before. They tend to have lower torque ripple, not sure if there are other advantages. A sinusoidally-wound 3-phase motor with sinusoidal excitation also has zero ripple, however. The wiring and power electronics are more complex for more phases, which is why most stick to the sweet-spot of 3-phase.

[Lebowski]

I'm thinking about a 4-pole 2 windings 90 degrees out of phase design, so the 4-phase you're talking about.. bit of confusion
about nr of connections versus nr of coil, sorry 'bout that

[rhitee05]

You could debate what the best name for configuration you're describing is - 2 phase, 4 phase, or something else. However, to control that configuration you would need an H-bridge for each coil pair. That's 4 FET pairs as opposed to the 3 pairs you'd need for a regular 3-phase design. What's the advantage?

[Lebowski]

You could debate what the best name for configuration you're describing is - 2 phase, 4 phase, or something else. However, to control that configuration you would need an H-bridge for each coil pair. That's 4 FET pairs as opposed to the 3 pairs you'd need for a regular 3-phase design. What's the advantage?

Well I think it'd be a cool thing to try, it would need a large and complex algorithm
inside a PIC to implement this which is fun . Also I think constant current consumption
would be an advantage as Lithium batteries prefer this (larger Ah for constant
current than for pulsed). With FETs costing a few $ 4 instead of 3 FET pairs
is not an issue...

[rhitee05]

If you want an interesting and more complex problem for a PIC, I'd suggest you try to implement sinusoidal commutation for a standard 3-phase motor. This should in general give you less torque ripple. If you want to make it more complicated still, add a current sensor on 2 phases and then you can implement real-time adjustment for the motor timing to keep the current aligned. Take it one step further and you could even play with field-weakening modes for high-speed operation... Plenty of things you could try without the hassle of needing a custom controller and custom motor. I'm not even sure how much theory is available for 4-phase motors for you to go off of. If you want to try it, though, go have fun I'm not one to criticize!

BTW, reducing battery current ripple isn't really a function of commutation patterns or the number of phases. Current ripple is mostly due to the PWM, so reducing it is mostly a matter of larger and/or lower ESR bus caps.

[Lebowski]

Building a custom motor is no problem I also build my current 3-phase motor from scratch.
Am looking into AC induction type motor.....

What is the whole idea behind sinusoidal PWM, isn't it fo have a constant phase current ? Wouldn't
it be easier to measure the motor current with a high-bandwidth loop and control it to a constant
value ?

[rhitee05]

Obviously doing sinusoidal PWM is harder than constant current. The advantages are constant power delivery to the load, less torque ripple, and lower harmonic content (more useful for line-powered equipment). It's also more efficient for motors with sinusoidal BEMF (distributed winding). I know most induction motors are wound this way, and I think some of the large pole-count hub motors are as well. The idea is that if the motor has a sinusoidal BEMF and you drive it with a square wave, the current near the beginning and end of the commutation is causing loss but not generating much useful torque.

[jdb]

You could debate what the best name for configuration you're describing is - 2 phase, 4 phase, or something else. However, to control that configuration you would need an H-bridge for each coil pair. That's 4 FET pairs as opposed to the 3 pairs you'd need for a regular 3-phase design. What's the advantage?

Not saying its a great idea, but you can drive 2-phase AC power with three pole pairs (and therefore, only three wires). Provide A, B, and Neutral outputs from the controller. Connect the two windings A-N and B-N. Drive N with 50% constant pwm, A with sine, and B with cosine pwm. The current through the neutral wire is nonzero, but it does provide two-phase power if you really want to play with it.