6 phase motors - are they all that?

[macribs]

Trying to google and see if I could find a clear answer to this I realize there is very little written about 6 phase motors.
And it is hard to find anything written in a language suitable for the average layman.

From what I have read about 6 phase motors I can't help to wonder why people still build 3 phase motors. And even in that Colossus thread there was posts asking for a special 6 phase version of that motor. I saw on youtube 6 phase was said to be in fact two 3 phase motors in just the one housing. Delivering more power, not double up but not that far from double the power vs 3 phase .

It seems 6 phase motors run smoother then 3 phase. Have more torque and put less strain on the controller as a 6 phase uses 2 controllers rather then 1. This also gives you a chance to get home in the event one controller fails. You can still run the motor as a 3 phase and get home.

When choosing controllers you will have an even wider range of controllers to choose from because you might do with out the top spec'd controllers as you split the load. I've seen threads here that show 6 phase motors with great acceleration and top speed more then 100 miles an hour. And all that delivered without access heat build up, with less noise then from a 3 phase motor - in fact the 6 phase was almost silent and with smooth throttle response and riding.

If 6 phase is better then 3 phase why so few 6 phase motors available?

And will 6 phase work well for outrunners RC motors as well as for DD hub motors? Imagine them RC motors being dead silent and what a nice ride it would be.....rather then high pitched noise we here today like a high power blender making drinks.

What situation and what motors would not work well as a 6 phase?

 

[kenkad]

macribs,
You might want to actually refer to them as multiple (multi) three-phase motors to differentiate from 5 phase or 7 phase (in the literature) A multi three-phase motor uses multiple three-phase slave controllers. I really doubt you get twice the power. Smoother running (less cogging torque) should be true. What you would have available is to shut down one of the three-phase slave controllers if you are cruising and do not need the additional torque. One important fact is that each of the slave controllers would be drawing less amps, so less heating, both in the motor and the slave controller.

I have attached a picture of an Apex motor that has two magnet rings on the rotor and 12 coils on each stator ring. In theory, you could have four three-phase groups on each ring (each sequence of three coils can be a three-phase group). There ends up being a great great deal of wiring if you go this route. You better make the coils easily replaceable if you go this route.

So many options, so little time to exercise them.
kenkad

 

[Punx0r]

I thought a 6 phase motor was just like a 3 phase, but with twice the number of coils and spaced at 60 rather than 120 degrees? :|

 

[kenkad]

PunxOr,
I showed that Apex motor as an example. Best design practice is to have each three-phase group to have six coils so each sequential group of three coils are 180 degrees apart, just to balance the forces on the motor bearings. Keep in mind my example is an AF BLDC motor. The design we have simulated has 24 coils so we have divided that into four three-phase groups. That is on an outer ring like the Apex design. Our inner ring has 12 coils so we have divided that into two three-phase groups. In total, the motor has six three-phase groups. The controller has six slave controllers, each identical in design, using an IPM and a microcontroller. Incidentally, they do not have to be in timing sync relative to the driving waveform. Visualize the entire controller as a hexagonal can, sides being the slave controllers, IPM to the outside with a heat sink. Top of the can is the master controller managing the slaves. Bottom of the can is the wiring connectivity to the coils, battery power in and three-phase connectivity out.

As I said earlier, you should be able to operate as many of the three-phase groups as you need torque. More three-phase groups means less current per three-phase group. Less three-phase groups operating, less heat dissipation. Cruising should use less torque than starting from a dead stop to get up to speed. This work is directed toward single passenger urban commuting vehicle design concepts and is strictly a technology demonstration. That is all I can say at the moment.
kenkad

 

[macribs]

I have seen people do modify their motors in various ways. Can any electrical motor be modified into 6 phase to enhance the torque, smoothness and split controller load? And will such a modification lead to any negative issues worth considering before modifying the motor? What about strain on axles and bearings? Will 6 phase put more stress on those part or less? (smoother operation might wear less?) Will a 6 phase make for any less longevity then a 3 phase?

Are there any downsides to 6 phases for use in e-bikes and ULEV?

 

[kenkad]

macribs,
Multi three-phase motors should be designed that way at the start. I personally would never try to rewind a inrunner or outrunner as a multi three-phase motor. Others may be braver. Bearing stress is very very important. The number of coils and magnet poles is also very important. That is what I tried to point out. It is also not a simple task to take two three-phase controllers and run them as two slave three-phase controllers (how do you tightly slave them together and who is going to be the master?). The world is not that simple. We are using a very high speed 14-bit positional encoder in our design effort. Using just three Hall sensors seems kind of crude for this task. Just my opinion. I am too old to waste my time on 'stupid stuff' (memorable statement by a current political figure). You are asking good questions, but, the answers to your questions are not so simple. In spite of what some E-S members are stating, I would question everything. BLDC motor design technology is not rocket science, but, maybe it should be. Lets hear laughter now!

I am surprised you are not getting more comments on this question.
kenkad

 

[Miles]

As I said earlier, you should be able to operate as many of the three-phase groups as you need torque. More three-phase groups means less current per three-phase group. Less three-phase groups operating, less heat dissipation.

I still don't understand this, Ken. You can switch off a phase group but the parasitic torque remains the same....

As for the arguments for smoothness, it should be possible to get torque ripple of less than 1% with 3 phases.
Cogging depends on the LCM of the slot pole combination chosen and the geometry of the teeth and magnets.

I just don't see a significant advantage to the motor in running the phase groups separately.

 

[macribs]

I still don't understand this, Ken. You can switch off a phase group but the parasitic torque remains the same....

As for the arguments for smoothness, it should be possible to get torque ripple of less than 1% with 3 phases.
Cogging depends on the LCM of the slot pole combination chosen and the geometry of the teeth and magnets.

I just don't see a significant advantage to the motor in running the phase groups separately.

No I don't think the amount of output torque remains the same when one or more phase group or group of three is switched off.
But then again you won't need it to be, acceleration takes more torque then remaining a given speed. So switch off according to your programmed desire and use for this example 6 phase in acceleration, or lower speed climbs. When the need for more torque is biggest. However switching off a phase group might come with a slightly penalty, magnetic drag?

Well advantages for a 6 phase would be more torque when needed the most. In return you will accelerate faster, climb steeper and all this producing lees heat, aka higher efficiencies. Your controller load will be split between multiple controllers, each controller running an easier load. In return each controller runs cooler, which also means less resistance and even here efficiencies gain. More efficient operation, less heat and more power converted to torque in the wheel. As many or most motors already runs at 88-94% efficiencies any chance we get to gain some more efficiencies should be grasped with both hands.

And to top it all, you even get some point of redundancy, if you have failure in one controller you are not stranded roadside.

If yo can manage all that running one single phase motor, I guess you are right. What is the point of 6 phase.
If you can't manage all that I think we just got a winner - 6 phase.

 

[Miles]

So switch off according to your programmed desire and use for this example 6 phase in acceleration, or lower speed climbs. When the need for more torque is biggest. However switching off a phase group might come with a slightly penalty, magnetic drag?

For the motor, I don't see any advantage from switching off....... As for the controller losses, perhaps Ken can explain that bit....

 

[Punx0r]

Kenkad, the multi-3 phase arrangement shown in your example is interesting and not something I realised existed. It's clear you have a large knowledge of these motors.

My initial thought is that the inner set is at a significantly reduced radius compared to the outer set. Considering the motor has to be large and heavy enough to accommodate the outer set, is this a disadvantage for torque density? I appreciate this may only be a issue for lightweight transport scenarios.

I recall a recentish thread involving discussion of a >3 phase motor from a premium ebike manufacturer (they were claiming no motor drag when pedalling) where comparisons were made between 3 and 5+ phase motors. I recall a persuasive argument that you can get very close to a perfect motor with only 3 phases.

It would be good to hear from JohninCR as I believe he has experience running 6 phase hubmotors on 2off inexpensive 3 phase controllers.

 

[Miles]

My initial thought is that the inner set is at a significantly reduced radius compared to the outer set. Considering the motor has to be large and heavy enough to accommodate the outer set, is this a disadvantage for torque density? I appreciate this may only be a issue for lightweight transport scenarios.

No more than the turning moment spread on an axial flux motor with more conventional topology? I'm not sure what the advantage of the segmented "U" coils is, though...

 

[macribs]

macribs,
Multi three-phase motors should be designed that way at the start. I personally would never try to rewind a inrunner or outrunner as a multi three-phase motor. Others may be braver. Bearing stress is very very important. The number of coils and magnet poles is also very important. That is what I tried to point out. It is also not a simple task to take two three-phase controllers and run them as two slave three-phase controllers (how do you tightly slave them together and who is going to be the master?). The world is not that simple. We are using a very high speed 14-bit positional encoder in our design effort. Using just three Hall sensors seems kind of crude for this task. Just my opinion. I am too old to waste my time on 'stupid stuff' (memorable statement by a current political figure). You are asking good questions, but, the answers to your questions are not so simple. In spite of what some E-S members are stating, I would question everything. BLDC motor design technology is not rocket science, but, maybe it should be. Lets hear laughter now!

I am surprised you are not getting more comments on this question.
kenkad

Well there seems to me many people have already sat their mind on the topic of 6 phase vs 3 phase. I don't know what made them decide which to fancy or if they put in research beforehand. But the select few can often carry a surprisingly large amount of "common knowledge" and swing things into consensus. I mean if heavy weighter's one time pitched in to a thread like this, a commonly feeling of correctness might spread based on the power of suggestions alone. You read something or references to something often enough and eventually that will become the truth, the whole truth and no questions asked. And soon it might turn into a public view - consensus sat no point of arguing anymore.

Right or wrong? Who knows anymore. But seeing as so many have read something along those lines so many times it must be correct, right?

We can chase grams and lbs all we like. The payout is great. The initial buy in expensive. Less weight gives a more agile and faster bike. People often talk about how bad it is to run mx rims and tires due to weight. Don't be surprised to learn the same folks might even carry a 15-20 lbs something in their rear wheel themselves. How was that again, weight penalty? And you put all the added weight where it hurts the most? Coherent anyone?

While there is nothing wrong with running hubs in the rim, in fact for many people it will make perfect sense according to their bikes actual use. Yet we can not argue that putting that motor in the triangle will not give you vastly enhanced handling. But do you really need that extra handling? If all you do is riding to work and back, with the occasional run to the store or the pub.

Maybe that can be part of why this topic has not gotten more attention. Most people feel what they got is more then enough.
Why chase the grams when things are working just great? Often we see arguments of simplicity thrown in for good measure. Simplicity? Well if simplicity was really to be targeted at all costs, then we should all ride hard tails with no suspended fork. Or at the very least just run simple old school shocks and forks with plain oil - and progressive springs? why even bother. Who needs the complexity of an air ride? Broken seals and more maintenance? Oh wait - the benefits of air ride out weigh the added complexity for most people.

Is a 6 phase more complex then a three phase or less complex? It is all in the eye of the beholder. Yes it is maybe more complex to set up initially. But is really prolonged life and easier and more efficient runs more complex after all? If it in fact leads you to less maintenance because you don't need to push your parts to and often beyond the limits?

 

[Thud]

For what its worth:
http://endless-sphere.com/forums/viewtopic.php?f=30&t=22091#p321635

I would be more interested in a discusion about 5 phase motor design & any theory's regarding potential benifits.

 

[macribs]

For what its worth:
http://endless-sphere.com/forums/viewtopic.php?f=30&t=22091#p321635

I would be more interested in a discusion about 5 phase motor design & any theory's regarding potential benifits.

Well feel free to join in - we don't need to constrain the thread to only 6 phase vs 3 phase if 5 phases brings stuff to the table too.
There was some papers posted on 5 phase motors in a thread a while back but I didn't take the thread serious as it seemed to target one specific vendor and praise that vendor rather then to discuss the fundamentals of the design theories behind it. At the time it kind of look like a marketing thread or a for sale thread hidden in a none sale part of the forum.

I do remember it being said that the company behind the motor in question in the thread was the world only manufacturer of 5 phase motors, and up until they started 5 phase motors had been only an academic exercise with no real life head to head testing to other design theories.

Maybe the 5 phase is all that?

I might have jumped the gun on that thread cos I seen it is still active, yet I have not re visited.

 

[Punx0r]

My initial thought is that the inner set is at a significantly reduced radius compared to the outer set. Considering the motor has to be large and heavy enough to accommodate the outer set, is this a disadvantage for torque density? I appreciate this may only be a issue for lightweight transport scenarios.

No more than the turning moment spread on an axial flux motor with more conventional topology? I'm not sure what the advantage of the segmented "U" coils is, though...

Good point.

I wonder how two sets of coils in an axial flux arrangement compares to a single larger coil covering the same total area. The single coil could have more iron, the double coil more copper...

 

[Miles]

I wonder how two sets of coils in an axial flux arrangement compares to a single larger coil covering the same total area. The single coil could have more iron, the double coil more copper...

I'd say that you have a free choice in how to apportion the area, for either case.... Probably, the double coil has the potential to waste more copper... Anyway, there's quite a lot of wasted area in that design, for sure....

 

[kenkad]

Miles,
I never said that parasitic torque was not present. That is a characteristic present in every permanent magnet based motor regardless of how many three-phase groups the motor has. As you said, reduce the parasitic torque by virtue of the motor design. The more magnetic poles and the smaller the coil face, the lower the parasitic torque. For me, that is a non factor in this discussion.

I have said in other threads, designing motors and controllers as separate entities is a big mistake. I contend that you must design the motor and controller concurrently. Everyone is entitled to their opinion.

Macribs,
Regarding redundancy. If you have distributed uCs in the controller, you do not necessarily gain redundancy. In our case, if the master uC dies, the controller dies. If a slave uC dies, then you can still go on provided the IPM is shut down on the dead slave controller.

The work we are doing is for technology research only. It is hard enough to find academic institutions to participate. If they quit any part of the project, we still have to go on with the effort ourselves or find another institution. Whether this effort is/is not successful, the world will go on. This is simply a challenge to find better ways of implementing known technology.

With regard to controller losses, my/our interest is only in IPMs. Higher power Mosfet based IPMs are in the works. Here is one of those never ending struggles. One uC in a complete controller ends up having to run faster, thus probably requires a lower operating voltage (3 or 3.3 VDC), then more susceptible to noise, more critical PCB design, etc., etc. Multiple IPMs allow reasonable PCB trace widths/copper thickness and the critical parts are inside the IPM module. Stop putting additional copper on the traces for increased current capability. Multiple uCs share the computational workload and thus the software is less complicated. How much more explanation is needed? I thought it was called KISS.
kenkad

 

[Miles]

Miles,
I never said that parasitic torque was not present. That is a characteristic present in every permanent magnet based motor regardless of how many three-phase groups the motor has. As you said, reduce the parasitic torque by virtue of the motor design. The more magnetic poles and the smaller the coil face, the lower the parasitic torque. For me, that is a non factor in this discussion.

True, you haven't. But where are the gains in switching off coils? You have the same parasitic losses and you are reducing the active copper..... There may well be gains in the controllers but it only looks like bad news for the motor itself....

 

[kenkad]

Why would anyone want to reduce the active copper? That one is easy. If you have a single three-phase group, and smaller diameter wire, then the inductance goes up. Go to larger wire? That is good except it is harder to wind and more difficult to make interconnections. More three-phase groups mean smaller diameter wire, consequently lower inductance, maybe easily replaceable coils, etc. In a previous thread, the discussion was how inductance is a two edged sword. Too much is very bad (waveform distortion, need dynamic timing control, etc.) and too little has its problems as well. I prefer to try to explore how to make motor designs less expensive, not more expensive by going to some kind of special wire forming/connecting, etc. That is the philosophy we are pursuing and trying to initiate in the academic environment. We will either succeed or not.
kenkad

 

[Miles]

You're not addressing the fundamental point. Not utilising all of the available copper means a less efficient motor.

 

[kenkad]

Miles,
You are welcome to your stated position. I prefer to say that we are using the necessary copper to create the torque required at that moment of time in the most efficient manner. Tongue in cheek, when crossing a desert, should I have one very large bottle of water and take little sips or have a dozen small bottles (same total quantity) and only use one at a time to take a sip, unless I need to take a bath. Oh well, I thought it was a good analogy. When you get old, you begin to realize that nature expresses itself in very similar ways across many areas of science. One large ant pulling a grasshopper or many small ants pulling the same grasshopper. Humans are a diverse population and not all think the same way.
kenkad

 

[Miles]

Miles,
You are welcome to your stated position. I prefer to say that we are using the necessary copper to create the torque required at that moment of time in the most efficient manner.

That doesn't make any sense, to me. The analogy is very misleading. Oh well, I don't expect to extract energy from the magnets, either...

 

[kenkad]

Attached is a paper by ORNL on stacking multiple three-phase AF BLDC motors on a COMMON shaft. THIS IS ANALOGOUS to multiple three-phase groups in a single motor design. Why would they do this research; max current through any three-phase group, inductance issues, modularity, etc.? You can read it and determine your own conclusions.
kenkad

 

[Miles]

Thanks Ken.

The paper describes a modular, stacked, 3 phase system. The motivation seems to be the ease of power scale-ability between different applications thus saving development and production costs. Nowhere does it suggest power scale-ability within an application by switching of the modules... It does discuss increasing torque ripple frequency by introducing a phase shift into the individual PWM carriers.

 

[Lebowski]

Attached is a paper by ORNL on stacking multiple three-phase AF BLDC motors on a COMMON shaft. THIS IS ANALOGOUS to multiple three-phase groups in a single motor design. Why would they do this research; max current through any three-phase group, inductance issues, modularity, etc.? You can read it and determine your own conclusions.
kenkad

Funny how stuff like this is discussed in papers in the official literature and such, while us here at ES just build and run that sort of thing. I agree that multiple motors on a single shaft is the same as a 6 phase motor. I've run 6 phase motors using two independent controllers without any problems, the only info shared between the controllers was the throttle info. There are videos to be found somewhere in one of my threads...

Regarding 6 phase versus 3 phase, there is a small advantage wrt winding the same motor as a 3 phase, due to the winding factor of 1 (instead of .965). The copper losses will be 7% less assuming the same torque. I'm assuming Johns 6-phase motor with 30 degrees between the two 3-phase motors. Another advantage is, for the same power you have for instance 2 6 FETs instead of one 12 FET, so less problems with non-equal current sharing between FETs.

5 phase over 3 phase, well, my guess would be 5 phase is noisier and will have more torque ripple (this is what is causing the noise, so actually the same point). A motor winding will have non-sinusoidal back-emf, so next to the fundamental sine wave it will also make 3rd and 5th order harmonics. In a 3 phase WYE the 3rd order will cancel out and you will have power ripple coming from the fundamental sine of the current and the 5th order sine from the back-emf voltage.
In a 5 phase motor the 5th will probable cancel when wound in WYE, with power ripple coming from the 3rd order back-emf component.

Nown typically in the backemf waveform the 3rd component is higher than the 5th order component. So, the 5 phase motor I expect will make more noise as its power fluctuations come from the 3rd order harmonic...

This all for WYE wound, Delta wound your efficiency is bad no matter what.