Halbach Axial Flux Motor

Discussion in 'Electric Bicycles' started by safe, Oct 31, 2009.

  1. safe

    safe Active Member

    Halbach Axial Flux Motor


    LaunchPoint has completed optimization and initial prototyping for a high efficiency, high power density, Halbach array electric motor. With a total weight of only 1.4 pounds, this motor produces 7 horsepower at 8400 rpm with 95 percent efficiency. At 5 horsepower per pound, this motor has a higher power density than any other motor on the market. The brushless, axial flux permanent magnet design is highly tolerant of temperature and centrifugal effects, and the ironless rotor and stator eliminate iron eddy current and hysteresis losses.

    http://www.geminielectricmotor.com/A Solution for the Future.htm




    Last edited: Oct 31, 2009

  2. safe

    safe Active Member


    Apparently this is a (relatvely) new way to order the magnets so that you do not need an iron core.

    This brings up an idea... :idea:

    What if you created a bicycle chainring that held the Halbach array attached to them and then mount onto the frame the coils to create the magnetic field. Then by applying multiphase (three or more) you can control the chainring like a motor... the chainring in effect "becomes" the motor.

    No geardowns.

    No extra chains.

    Very high power to weight ratio.

    ...the only thing I'm not sure about is the torque at low rpms.

    (will require some more study)

    Will require multiphase controller.
    Last edited: Oct 31, 2009
  3. safe

    safe Active Member

    Halbach Torque and Power

    This is very good news...

    The AC Induction motor has behavior that does not cut out abruptly at increasing rpm (frequency) because the slip of the rotor is without cogging. An AC Induction motors iron rotor has no orientation to it's magnetism.

    The Halbach motor design is actually in some ways like an AC Induction motor because it too has no cogging. The Halbach array has a flat magnetic field on one side and nothing on the other. The Halbach motor then is set to motion entirely by the frequency of the magnetic coils.... so just like the AC Induction motor the behavior is that of:

    "Field Weakening"

    ...this almost feels like finding the "Holy Grail". :cool:

    This seems to have all the advantages of the brushless permanent magnet motor (lightweight) AND the AC induction motor (wide powerband).

    The real genius of this motor design is that you eliminate the core since it's no longer a part of how magnetic fields are created. The motor uses a sort of "closed loop" from coil to coil... so many of the issues of other motors sort of vanish.

    I'm somewhat stunned.... :eek:


    Drop the idea about a chainring mounted motor, the needed rpm looks to be too low (100 rpm) for that to be efficient. However, at 500 rpm (multiple pole design) a simple geardown using the chain would be enough. It should be possible to rig something up that uses the stock bicycle chain. (11 tooth motor, 55 tooth chainring, ~30 tooth rear wheel sprocket)


    Attached Files:

    Last edited: Oct 31, 2009
  4. safe

    safe Active Member

    Last edited: Oct 31, 2009
  5. AussieJester

    AussieJester Member

    Still trolling threads at Endless Sphere i see Safe... If CNCAdict from ES gets interested we will see some on ebikes within 6 months. That said he has been AWOL for awhile was meant to have the RC motor design concept ramped up for ebike use and in production by now...Tiz definately an interesting motor though.

    I like the size too


  6. safe

    safe Active Member

    It's A Huge Breakthrough !!! (no joke)

    Don't you realize the significance of this?

    What this means is that you can get the lightweight high power-to-weight performance of the standard permanent magnet motor and also get the field weakening of the AC induction motor.

    You can literally get it all and not just parts in one motor.


    There is no core to these things so you have no cogging, no hysteresis, no eddy currents to deal with.

    This was really only invented at Lawrence Livermore Labs just recently (1980's), which explains why it's not known about much by the rest of the world. Most people will continue to use the standard permanent magnet motor configuration until they wake up to this idea.

    What is surprising is how relatively simply the invention is...


    Not only that, but since there are no return paths for the magnetic flux lines you can INCREASE the pole count significantly without suffering the magnitude of losses that an AC induction motor endures. This means that you can get LOW RPM motors to work well... which means the "Holy Grail" of motors that do not require complex geardowns is possible.

    If the base speed of a 60 Hz 2 Pole motor is 3600 rpm,

    then the base speed of a 60 Hz 8 Pole motor is 3600 / 4 = 900 rpm,

    and the the base speed of a 60 Hz 12 Pole motor is 3600 / 6 = 600 rpm. :party:

    Last edited: Nov 1, 2009
  7. safe

    safe Active Member

    Harmonics "Need Not Apply"

    It becomes obvious that using Harmonics like the Third Harmonic Injection for AC Induction motors becomes no longer necessary or even desired.

    Why did we want harmonics?

    Well, because the iron rotor in an AC Induction machine creates this constant "echo chamber" as the magnetic flux passes through it. It's like shouting in a cave.... the sounds echo back and forth within the rotor of the AC Induction motor until they do something. The concept of harmonics is that you send specific waves into the rotor so that they bounce out at just the right time to cause torque at a lower rpm than the fundamental.


    With the Halbach configuration the permanent magnetic field is "flat". Also, the magnetic field lines now go in circles between the field coils so there is no core to go through and no possible reason for using harmonics.

    So a simple Sine wave controller (Three Phase) would be all you need for this... and that means you can use less expensive components.


    Note: RC motor controllers (ESC) should not work with these motors because there wouldn't be any kind of backEMF pulses to watch. As I understand it the RC ESC's in effect watch for the cogging behavior of the standard permanent magnet design and do their timing based on that. The Halbach design needs a controller more like the AC Induction motor because there is no cogging. However... many of the AC Induction motor controllers use sophisticated algorithms that figure out what is going on inside the iron rotors and you DON'T WANT THAT, so you might have to take a step back to regular Voltage vs Frequency control. It also might be possible to think up some new and better controller schemes for the Halbach motor.

    The controller will be another interesting thing to watch evolve.

    However, it also might just be as simple as Voltage vs Frequency with some kind of current control thrown in. From what I can tell these motors have very, very simple mathematics to them. This is actually making things less complicated and "better".

    Obviously I'm very excited about this.... it's like searching for something knowing that the right answer has not been found and then finally finding it. At last we (the ebike community) might have found what we needed from the beginning.
    Last edited: Nov 1, 2009
  8. AussieJester

    AussieJester Member

    No safe im a total idiot i don't understand anything about it .... :: shakes head :: Im surprised you can though, took you long enough to discover brush tension and electric motor timing :p

    No Safe, i simply don't get a hard on and feel the need to make a dozen unanswered posts expressing how great electric motors are...When they are available off the shelf for a reasonable price ill consider getting one, until then im more than happy Like 95% of the other people in this world will be, to use what is available now. Do carry on though what ever floats your boat...

  9. safe

    safe Active Member

    AC Induction Motor Project "Cancelled"

    Had I rushed forward into the AC Induction motor project without thinking about it very much (just buying a bunch of stuff) it's very likely that I would not have had the freedom to cancel the AC Induction motor project because I would have had an investment to protect. So far I've only invested some time (drilling the stator) and a little money ($20) for an old dishwasher motor.

    Unless something emerges that says the Halbach design does not live up to their performance charts (unlikely) it only makes sense to switch to this design over the other.


    What do we know, what have we learned?

    :D The journey began with permanent magnet motors of the brushed and brushless kind. They have in effect a couple of design flaws. The first flaw is that they have a fixed speed that they can attain before the backEMF (caused by cogging) prevents more rpms. This means in practice that you need to use multispeed gearing to get useful performance. The second flaw is just with the brushed motors and that is that the brush/commutators tend to fail if you push them too hard.

    :D The "next step" was to realize (by watching the Tesla roadster and others) that with the AC Induction motor you can expand the powerband indefinitely. This is because the iron rotor has fields created in them and they are "smooth" in the sense that there is no cogging behavior. No cogging, no speed limits.

    :D Finally... with the Halbach... we eliminate the cogging and even throw away the iron core itself dropping the weight of the motor significantly. You now have effectively an AC Induction motor in that the movement has no cogging elements to it and yet you use permanent magnets.

    The Halbach makes the permanent magnet motor "appear" like an AC Induction motor.

    ...and the controllers will be more AC Induction than Brushless motor in logic.
    Last edited: Nov 1, 2009
  10. safe

    safe Active Member

    It vindicates my intuition about the direction of ebike motors.

    The whole permanent magnet motor thing with it's geardowns and multispeed complexity was just...... wrong.... in my opinion. The maintenance on all that stuff was going to make operating these things a nightmare.

    You are trying to be "Mr Practical" and all... and I get that... but have some excitement that the answer has been found.... "finally".


    All these Halbach motors are placing efficiency numbers in the middle to upper 90% range... that's simply incredible !!!

    If I'm going to build a motor from scratch I might as well do this type.
    Last edited: Nov 1, 2009
  11. safe

    safe Active Member

    A Motor Made With Fiberglass

    What I'm thinking is that with a steel frame underneath to provide some basic strength and then fiberglass on top you could build a Halbach motor with just basic tools.

    The idea would be to wire up all the magnet wires for the stator side and then just embed everything in fiberglass. For the rotor side (where you place the permanent magnets) you build something with steel that will be where the bearings are attached and then embed that steel frame with the magnets so that everything gets glued together. Add fiberglass cloth as you do the embedding and that will reduce the chance of things cracking.

    I've found that with the brushed motors that fiberglass can handle the heat very well as long as it's not directly in contact with current. (like as a commutator which is just too much) For things like the magnet wires the fiberglass is able to deal with that level of heating.

    Anyway... it might be possible...


    ...it might be a good idea to buy some carbon fiber cloth.
    Last edited: Nov 1, 2009
  12. safe

    safe Active Member

    What It Is Not...

    The Halbach motor design gives us the ability to widen the motors powerband and make it "act" like an AC Induction motor.

    What it doesn't do is signfiicantly increase performance compared to a regular permanent magnet motor operating within it's "sweet spot" of efficiency.

    If you troll the RC forums you get a lot of dissapointment with the Halbach motor because it doesn't deliver any improvement for their airplanes that run at a singular (high) rpm most of the time.

    The reason for this motor for ebike use is to have powerband widening.... it's because of the lack of a speed limit of the motor that it's of value.

    I'm sure a lot of people are going to think that and bring it up.

    (so I figured that I'd explain that beforehand)


    Also I'm realizing that when you do this it doubles the number of magnets needed for any particular pole count you desire to make. This means that you end up having to use lot's of smaller magnets rather than just a few bigger ones. It might be hard to build a Halbach motor that actually achieves 12 poles with Three Phase power because the magnet count gets too high. 12 poles times 3 phases times the Halbach doubling means 72 magnets :ack2: A more realistic idea would be 4 poles, Three Phase which means 24 magnets. So the motor base speed becomes 3600 * 2/4 = 1800 rpm. Maybe 6 poles, Three Phase is possible which means 6*3*2 = 36 magnets and a base speed of 3600 * 2/6 = 1200 rpm?

    Basically the Halbach design filters out the cogging aspect of the backEMF and keeps the smooth part, so you still get "Field Weakening" as the frequency increases (just like an AC Induction motor) but that still means that you can use fewer to no gears compared to the standard design.
    Last edited: Nov 1, 2009
  13. safe

    safe Active Member

    The concept of "Base Speed" goes out the window...

    The very idea of a "base speed" that is used with AC Induction motors sort of dissolves or at least mutates to something that I don't understand fully yet. With AC Induction motors the "base speed" is the speed where the stator and the rotor reach their optimal behavior. This optimal behavior has much to do with the way iron behaves on both ends.

    With the Halbach motor there is no iron anymore!!!

    ...so the very notion of "base speed" seems to become meaningless.

    Basically what you are dealing with is a fairly linear Voltage vs Frequency relationship where the magnetism of the Halbach rotor creates a constant "smooth" backEMF that will correspond to the point when the Field Weakening begins. So strong magnets are going to change the point at which it happens and not the iron because there is none. The reason for doing the Voltage vs Frequency in the first place is to protect the magnet wires... so you can play with that too using thicker wire or Litz wire. (multiple winds)

    It will be interesting to investigate this behavior more... hopefully I find a pdf that covers it well. This Halbach design creates a new landscape to think about and understand. (it warps the mind a little I guess)

    There is a possiblity that you can use lower pole counts and then stronger magnets to arrive at lower frequency "useful" power. The V/F rate can be modified in either way I'm "guessing". (will report back when I'm certain) I have a "feeling" that there is going to be a "sweet spot" where the number of magnets (meaning higher pole counts) verses the magnet strength will produce a tradeoff. Bigger magnets are stronger, so you use less, have lower pole counts and that's equal in effect to having more smaller, weaker magnets with higher pole counts.

    Magnet Strength * Magnet Count * Pole Count "roughly equals" a Constant?

    ...the analogy being something like a gasoline motor where you decide on either a big single cylinder or many smaller cylinders. If the overall volume is the same, then there is a "sweet spot" of the ideal cylinder number.
    Last edited: Nov 1, 2009
  14. AussieJester

    AussieJester Member

    LOL.. your pretty slow on the uptake arent you pal...as i said i don't get excited over things like electric motors Safe, great its a new design yet too be proven for this application, on paper it looks promising, you can spout off about how its the 'answer' all you like, but IMO we have excellent options available now your just too stingy to try them and want too do EVERYTHING on the cheap, this new Halbach design is no exception, I have nothing against frugal means Safe but sometimes you need to open the wallet let the moths out and pay for some things. If you had simply bought a cyclone 1000watt motor you would have negated months ramblings and possibly have your bike in production, the rate your going you will be dead and the bike will still be motorless. i have seen dozens of ebikes that work brilliantly with the technology available just because you haven't had the nous to try it doesn't mean its not good. You have only ever used brushed motors Safe you have no idea how anything else performs in practical applications.

    And yeah... the whole gears thing is totally wrong, doesn't work well at all does it, best scrap every motorcycle ever made and start from scratch :-S

    No.. Safe...Im not being practical its called being 'realistic' Halbach design motors aren't available for ebike use yet period, so no point me getting all "excited" over them now is it! you try and make one of these motors you will fail, you quite simply dont have the machinery at your disposal to accomplish the task at hand...im questioning your skills too having seen your brush plate 'catastrophe' i'm guessing your eyesight lets you down when intricate work involving exact clearances and measurements are involved? as larger jobs you excel at.....You need a lathe and a mill for this type of work PREFERABLY CNC..Why don't you make one Safe? If you had a cnc machine you could design a Halbach motor motor in AutoCad **** ill even give you a copy of AutoCAD FREE!! Then you could mill it out of alloy with extreme accuracy. You Can build a cnc mill for well under 1000 bucks checkout CNCZone FREE plans available. Only thing i have trouble with is the electronics and the code which you wouldnt. Makes ALOT more sense than the method you intend too use.

    Shall follow on ES the progress makes more sense than your repeated posts stating the same thing over and over again LoL..

    Best of luck with your Halbach design motor build though wish you luck, you will need it...

  15. safe

    safe Active Member

    Safe Considers "Hub Motor" Project

    After all the years where I "bashed" hub motors I dare to say that finally there might be a valid situation to own one. :eek:

    Here's the idea....

    I was thinking of placing the Halbach motor onto the chainring, but normally the chainring runs about three times slower than the rear wheel. This is sure to be common knowledge within the hub motor world. But the problem with hub motors has been that in order to get any power you needed the "Big Iron" in the rear hub and all that weight is a problem for handling.

    Along comes the Halbach motor concept... no iron to worry about and a weight of 1.4 lbs for a full 7 hp in the best case. But let's assume that for 5 lbs I'm able to get a 1-2 hp motor and here's how it might be done.

    First you start with a double sided hub that has disc brake bolts on each side. You can buy them here:


    ...next you get either a disc brake rotor or an aluminum chainring of about the same size. Since aluminum would be less likely to effect the magnetic field that's probably the way to go, but in the picture I show both. Anyway, on the disc you place 48 small but strong magnets so that you can put together a Three Phase, 8 Pole motor. (3*8*2 = 48) With an 8 pole motor you will have a speed of 800 rpm at 60 hz and the amount of power you get all depends on the Voltage vs Frequency ratio. That is in turn determined by the thickness of the wire and the number of turns. More is better. You would end up wth a total of 24 sets of coils, but only three actual strands of wire. You can then control the motor using simple Voltage vs Frequency Three Phase control. (which is possible by buying an $8 chip)

    Anyway... now what you end up with is a rear wheel with a disc brake on the left and a motor made up of two discs on the right. The stator disc would be hollowed out and mounts would be attached to the frame that would extend bolts out to the stator so that you could adjust it in three dimensions. This is needed so that you can get the air gap reduced to as little as possible. A small fixed sprocket would be nestled in the middle and everything would bolt onto the six bolt disc mounts. A front freewheel separates the pedals from the rear wheel. (so that you can let the motor run while you coast) Ideally the freewheel would be on the rear wheel and I'll see about that idea. (depends on how hard it is to mount it) If it can be done easily then the front freewheel can be made into direct drive. This will reduce friction, so if it can be done I'll do it.

    Hopefully that's a picture that people can grasp... as time goes on it will become more clear.

    Attached Files:

    Last edited: Nov 1, 2009
  16. luka69

    luka69 New Member

    Attached Files:

  17. AussieJester

    AussieJester Member

    Hub motor boys are now got the delta and wye wiring switching on the fly have you seen this safe? effectively gives them two 'gears' to use by simply flicking the switch. GGoodrum is also working on similar device for RC motors too do the same thing i believe?

    With your design concept i see a minor problem with fitting it all between the chain stays and still being able to incorporate the sprocket for manual pedal abilities...


    p.s i use one of those hubs from ChoppersUS on my bike they are very solid, i actually cut one up too use the disk brake 'adapter' on them for the rear of my trike, you wont break one in a hurry all steel. :)
    Last edited: Nov 1, 2009
  18. safe

    safe Active Member

    This is a more "doable" project than brushed motors. This is no electrical current to deal with that cuts across a boundry that is fragile. Many people have been able to build stators with a board, some nails and then fiberglass resin. My brushed motor (which I just checked today) still has a strong fiberglass covering all over the magnet wires and it's strong as can be. There have been no problems with the wire side of rewinding.

    This project actually looks fairly easy to do... we will see...


    The weather has gotten better and I've added another 120 miles to the last motor (the old bike) so it's up to 7,120 miles now. (620 miles on that motor)

    AussieJester... since you are always bragging about "actually riding" being the most important thing, just how many miles have you ridden?

    I mean if I've gone 7,120 miles doesn't that grant me the right to dream a little? If all I did was talk I can see you complaining, but I'm probably doing 5 times the miles you are... if comparing miles is so important that is. :whistling:
  19. luka69

    luka69 New Member

    Anyone here interested in getting back to the original Topic "Halbach Axial Flux Motor" ????????????????????????????????????????????????????
  20. safe

    safe Active Member

    Unless they use a 8 pole count motor they still will need a serious geardown unit. Wye is lower rpm, delta is higher.

    Still... the point of the Halbach design is that you get the field weakening without the need for a Delta - Wye switch and also you get to save weight because there is no iron.

    I understand all the concepts... forgive me for wanting to go after what I think is the right thing to do... (never been much for following the herd)