Halbach Axial Flux Motor

[safe]

Where did you pull up the field weakining idea on the hallbach? They are synchronous motors.

This was answered once already a few posts back...

http://www.motoredbikes.com/showpost.php?p=242423&postcount=12

Yes, from a "peak efficiency" and "peak power" perspective the Halbach does NOT offer much. I've seen all the disappointment on the RC forums about the Halbach not delivering any "more" than before.

The big advantage is the powerband width. (and that's my goal... not "peak" which is limited by the law)

So for ebikes that are "legal" this is of value, but for RC motors it's not.

This is actually an area where what is good for RC is bad for ebikes.

But don't take my word for it... this is the actual torque curve of a Halbach motor:

...while the Halbach is in fact synchronous it demonstrates "AC Induction motor-like" behavior.

The Halbach is everything you would want in an AC Induction motor, but without the "slip".

 

[safe]

Good News Continues To Roll In...

Thanks again to Luka69 for finding this zip file and pdfs.

http://www.motoredbikes.com/showpost.php?p=242479&postcount=16

In the "SolarCarMotorPaperAsPublishedSml.pdf" file they discuss pole counts and they arrived at the number of 40 poles as the ideal for their motor.

I was worried that (8*3=24) 24 poles was too much.

Also, the Halbach has extremely high efficiency.

---------------------------------

Without the need for a core of any type the Halbach allows the motor designer to stretch out the pancake as much as you want. This seems perfect for matching things like discs and chainrings on bicycles because their shape is more like a big thin disc than the motors we are used to.

We ebikers like big wide flat discs because they are easier to design around. However, for something like an RC airplane the big flat disc is going to be hard to fit anywhere without causing increased wind resistance. So (again) the form factor of the Halbach suits the ebike and yet does not suit the RC world.

(the RC world will likely never adopt the Halbach)

 

[safe]

Sensors May Be Achilles Heal...

Unlike the AC Induction motor that relies on "slip" the Halbach motor is more along the lines of a regular brushless motor and needs sensors.

That's a major bummer.

From the "Solar_motor_ kit_ notes.pdf":

5.2. Rotor position sensor and adjustments

The complete motor comes with a tested and aligned position sensor, for the frameless motor a tested PCB is supplied. The rest of this section is only directly relevant to the frameless motor. The rotor position sensor contains three hall effect devices (Item 1 in Table 2) used to sense the fringing magnetic field on the inside of one of the motor magnet rings. The top of the hall sensors should be located no more than 1 mm from the inside edge of the magnets and around about the centre of the magnet (preferably slightly closer to the backing iron).

 

[johnrobholmes]

You can run it sensorless with RC equipment.


I'm still not sure why you think it can run any faster than another PM motor with identical pole/magnet counts. The only difference is that the Hallbach array has a directed flux with no back iron.

 

[safe]

Field Weakening... Is It Real?

I've based my understanding mostly from this pdf file. (see below)

In it they say:

In the constant torque mode below base-speed, maximum torque per ampere control is used, which results in zero -axis current since the rotor does not exhibit saliency. Above base-speed, when the motor operates under both supply voltage and current limitations, flux-weakening control is employed to maximize the power [4], [5]. Maximum flux-weakening performance is achieved [6] by designing the machine such that (V=LI) where (I) is the rated phase current, (L) is the -axis inductance and (V) is the stator winding flux-linkage produced by the rotor magnet.

...Hence, a high level of flux-weakening and demagnetization withstand capability can be achieved, as will be demonstrated later.

...it's possible that I've completely gotten the theory wrong based on this pdf sending me astray.

It was that paragraph and the chart (above) that came with it that made me think "AC Induction Motor".

----------------------------------

Efficiency seems to not be related to anything but current through the magnet wires... so the performance seems to not follow the typical permanent magnet curves.

Note: It's possible that most people are ignoring the Field Weakening area because they see it as a waste of resources. Other curves very well might account for frequencies equating of from zero to the base speed only.

----------------------------------

They might have even designed a unique motor specifically for field weakening. Look at the image, it seems to suggest two separate sets of windings.

 

[johnrobholmes]

They really don't give enough about the rotor. My guess is that either it is an induction motor with added halbach rings, or it is a normal PM motor and the "field weakening" term is being used instead of timing advance. The torque/rpm graph will look just like the one they posted when you use zero advance at lower rpms and higher advance as the motor speeds up.

 

[safe]

Hmmmmmm.....

This is interesting:

HALBACH magnetized brushless machines exhibit an essentially
sinusoidal airgap field distribution and a sinusoidal
back-emfwaveform, as well as negligible cogging torque,
without employing skew or a distributed winding. Thus, they
are eminently suitable for brushless ac operation. While Halbach
machines often employ segmented magnets to realize an
approximate Halbach magnetization [1], [2], in this paper an
anisotropic bonded NdFeB Halbach magnetized ring magnet is
employed. The magnet is produced by orienting the NdFeB injection
moulding compound in a powder alignment system [3],
and subsequently impulse magnetizing it in a Halbach field. The
features of such a Halbach magnetized machine, in terms of
its airgap field distribution and back-emf and cogging torque
waveforms, have been compared with those which result with a
Halbach magnet fabricated from discrete magnet segments [2].
However, the control performance of Halbach magnetized machines
has not been reported. Thus, this paper investigates the
steady-state performance of a Halbach magnetized brushless ac
machine in constant torque and flux-weakening modes, both
theoretical and experimentally. It also investigates the influence
of cross-coupling magnetic saturation on the torque capability.

So this rotor was built in some injection molding technique rather than using descrete magnets.

Hard to figure out what is really going on here.

-----------------------------------

Back with the solar car they actually place additional inductors into the windings so that they get enough inductance to make the motor work without burning itself up. Maybe the trick is in using variable inductance in order to control the motor in a Field Weakening type way.

My basic attitude is either Halbach can either do Field Weakening or it can't and if it can't (is limited to the same "voltage barrier" as the standard permanent magnet motors) then it takes away from the attractiveness of the idea.

----------------------------------

This is going to take a lot more reading...

 

[safe]

Current Sensor Rather Than Speed Sensors?

The concept of the speed sensor is to limit the advance of the frequency so that you don't get ahead of the speed of the actual motor. Since current demand is related to this difference between "actual" motor speed and "desired" speed (throttle setting) it seems to me that you could skip the speed sensors and just use a current sensor instead. With Three Phase power the rotation is always in the right direction.

Thoughts?

Soft starting seems to be needed to gently get things going in the right direction, but after that it should be possible to rely on the current sensor to limit the rate of advance.

A lot of this stuff is similar (if not the same) as brushless controllers, but they tend to not be too interested in the current drain. For the legal limit of power for ebikes you need to be "current centric".

 

[safe]

The Size Observation

Probably the main thing I'm seeing here is that with an iron core (whether you are dealing with induction type or permanent magnet type) you are always dealing with the need for intense magnetic flux density in order to achieve peak performance.

It's well known that for induction motors that small sized motors do NOT perform well when you increase the pole count. This is because as you increase the pole count you also increase the number of magnetic flux return paths. This then forces you to focus on harmonics to "simulate" higher poles and get better low end performance.

With permanet magnet motors in the typical configuration you can increase the pole count, but that basically lowers the effective gear ratio of the motor (the kv ratio) and that's going to hurt you one way or the other in that you need to deal with that in gearing somehow. It's possible to use Delta-Wye to adapt, but it's not really what you want to do. (it gets complicated and the transition can pose problems)

With the Halbach motor design I see the ability to spread out the motor into a large disc and then increase the pole count to a very large number. If you select the right amount of copper to allow a high enough top speed for the motor you can then drive the motor high enough for the top speed and also still get enough low end when you need it if you carefully control the current at low frequency.

There is no need for an iron core, so it kind of opens up the design to something that no longer looks like a motor we are used to seeing.

(more like a flying saucer than a motor )

...this just begs to be placed into the chainring or the disc brake areas.

Even if the Field Weakening is a "misdirection" based on that one pdf a linear powerband motor that can be tweeked right could still be a desireable goal. (it just means you need more copper)

The Halbach motor idea has already taken a few "hits" but it still looks good... will have to look at it more...

 

[johnrobholmes]

I think you are referring to an axial motor as a Halbach. Halbach only refers to a positioning of magnets to get a directional flux.


Axial flux motors do beg to be used on bikes. The form factor is perfect, and as you state they can be very easily scaled up to get proper torque and speed characteristics suitable for single stage drivelines. Made large enough it could be used on/ in a hub.