New kid in town. They develop axial flux PM motors for EV (like YASA). 
https://www.magnax.com/electric-transport
https://www.magnax.com/electric-transport
Axial flux electric motor
1 - 20 of 28 Posts
Oh no, another hub-motor.
This is, of course, not the definition of insanity, but it's an example of nonsensical behaviour.
There are the usual claims of superiority by every measure, including a wild claim of 98% efficiency. To their credit, they do have a little better description of their design than most, rather than just claims... but nothing directly on the website explains why their design is more efficient or more power-dense.
Bizarrely, although the application featured in the website images is a hub-motor in a car, the only specs are for an 850 kilogram and 1.6 metre diameter machine putting out 100 kW at 60 rpm... apparently for some sort of heavy industrial equipment. Reading through their site and attached material, it appears that the have built only this size of unit, and they have worked only on industrial applications.
The downloadable "whitepaper" does provide more technical background, but is mostly a survey of some variants of existing industrial motor technology, and provides no real explanation of their unique feature. It does explain clearly that they are the same as YASA motors, which strangely don't have incredible efficiency or power density, while using the same radial flux and yokeless stator design.
“The definition of insanity is doing the same thing over and over again, but expecting different results”
This is, of course, not the definition of insanity, but it's an example of nonsensical behaviour.
There are the usual claims of superiority by every measure, including a wild claim of 98% efficiency. To their credit, they do have a little better description of their design than most, rather than just claims... but nothing directly on the website explains why their design is more efficient or more power-dense.
Bizarrely, although the application featured in the website images is a hub-motor in a car, the only specs are for an 850 kilogram and 1.6 metre diameter machine putting out 100 kW at 60 rpm... apparently for some sort of heavy industrial equipment. Reading through their site and attached material, it appears that the have built only this size of unit, and they have worked only on industrial applications.
The downloadable "whitepaper" does provide more technical background, but is mostly a survey of some variants of existing industrial motor technology, and provides no real explanation of their unique feature. It does explain clearly that they are the same as YASA motors, which strangely don't have incredible efficiency or power density, while using the same radial flux and yokeless stator design.
As I see it, it seems to be a concept which can be used for both bigger machines (high torque - low rpm) as smaller hi-speed motors.
They also say that the motor can be mounted on the chassis.
In any case, 15 kW/kg peak sounds great for my e-bike project. I'm gonna check it out.
They also say that the motor can be mounted on the chassis.
In any case, 15 kW/kg peak sounds great for my e-bike project. I'm gonna check it out.
Sure, check it out, but don't expect too much: aside from the issue that Magnax has probably never built an e-bike sized motor, keep in mind that nothing scales that easily, so the power density and efficiency in a small motor will not be the same as in a large motor... and the one example they have is very large.
Also, while they claim 15 kW/kg, the one motor spec they publish is for an 850 kg and 100 kW motor - that's only 0.12 kW/kg!
I don't see any reason to expect better than YASA's product, which run up to 6.7 kW/kg in the P400 series.
Also, while they claim 15 kW/kg, the one motor spec they publish is for an 850 kg and 100 kW motor - that's only 0.12 kW/kg!
I don't see any reason to expect better than YASA's product, which run up to 6.7 kW/kg in the P400 series.
Yeah, but the 0,12 kW/kg is for a 16 kNm torque generator machine running at 60 rpm which is a pretty low RPM.
Smaller motors run in the x thousands RPM and according the formula kW goes up linearly with RPM...
I'll give you a sign when I know more details.
Cheers
I agree that the power density problem in this case is low speed, but the only real motor they have (allegedly) built doesn't demonstrate what they claim. I say "allegedly" because the torque and power output graph published in the spec sheet is an obvious fake (perfectly straight-line constant torque), rather than an actual test result.
Power density will not scale up linearly with speed in the real world. If the same motor ran ten times faster at the same torque it would have ten times the power density; however, at ten times faster it is unlikely to maintain the same torque. If constructed for ten times higher rotational speed, it will also likely be heavier.
Can you imagine that 1.6 m diameter motor running at 6,000 rpm, putting out 16 kNm and 10 MW? I don't think so.
Of course, you're looking for something much smaller, running at slightly higher speed and much lower torque, but there's nothing provided to suggest that anything about the one published motor can be extrapolated to 15 kW/kg, or 98% efficiency, or anything else.
Joined
·
1,952 Posts
It's an interesting concept. One possible advantage i can see is the use of one set of stator windings with two sets of magnets on rotors at both ends.
All the other axial flux motors seem to have a single set of magnets on a rotor located between stator windings on each end.
My experience with axial flux linear motors is that generally these machines scale up well, but not so much scaling down. These sort of large diameter machines would work great for drilling oil wells.
All the other axial flux motors seem to have a single set of magnets on a rotor located between stator windings on each end.
My experience with axial flux linear motors is that generally these machines scale up well, but not so much scaling down. These sort of large diameter machines would work great for drilling oil wells.
That makes sense.
An example they mention in the whitepaper is a wind turbine, which is even slower and larger. Direct-drive wind turbines have been tried (probably for decades), but have not yet shown a clear advantage over higher-speed machines with gear drives. 100 kW is very small by commercial wind turbine standards and used only in remote locations; to produce the 2 MW at 15 rpm of a mainstream turbine, a stack of 80 of the featured Magnax units (weighing an unworkable 68 tonnes) would required! Obviously they could scale up - they would need to. It's not apparent to me how a Magnax machine would have any advantage over the Siemens and General Electric direct-drive machines (which appear to be radial-flux) that have been doing this job for several years.
Hmmm??.... Sounds like a small drive unit for a pumped Hydro dam turbine !
....but they would normally be a few hundred MW !
Joined
·
4,329 Posts
The biggest downsides are that a) you are placing the motor in the place on the vehicle with the most intense stresses and b) if you have any kind of suspension, you have increased the weight of the wheel making it harder to give a good ride. I would submit that even on a traditional bike with no suspension it would be worthwhile to somewhat isolate the motor for longevity, if only on rubber bushings, and use a belt to convey power.
Over on the endless-sphere site, member Lebowski hand made an axial flux bicycle motor, then, made a multiple stator version as a mid-mount. He uses it all the time.
Then, he made the controller and is responsible for helping many people build their own permanent magnet motor controller.
Then, he made the controller and is responsible for helping many people build their own permanent magnet motor controller.
Joined
·
7 Posts
Hi,
This is Daan from Magnax.
Interesting discussion here. Allow me to give some additional information.
For our axial flux topology, torque quadruples when the diameter doubles. So for a 2 [email protected] 15 RPM generator, we would recommend:
- A stack of 8 discs of 4m diameter (total generator weight: 16 tons). delivers 1243 kNm
or
- A stack 5 discs of 5m diameter (total generator weight: 12,5 tons). delivers 1254 kNm
The weight of traditional direct-drive = at least 40 tons.
For a 2 MW turbine, I don't expect nacelles with larger diameters than that. But if so, the weight benefit increases even more.
Each disc has an axial length of 14cm.
Best regards,
Daan
Joined
·
7 Posts
Hi Brian,
True; our high-speed motors -for commercial purposes- are still in development, but we actually started with high-speed prototypes a few years ago. See below a photo of the last prototype which was installed back-to-back with an induction motor.
Then we created a high-torque prototype (16kNm) for a wind turbine manufacturer to check the scalability of the concept.
As an example for EV's; The 265mm diameter model has the following (preliminary) specs: 5500 rpm max, 150kW nominal, 300kW peak, 260 Nm nominal, 521 Nm peak, 22,5 kg.
We have compared the theoretical data (coming out of our software) with real prototypes and the numbers have proven to be very precise.
Our machines are also Yokeless but some reasons for the extreme power densities are:
- A new -patented- cooling concept which seems to be extremely effiective. Very effective cooling makes a huge difference in terms of performance. Feel free to reach out for details.
- We use electric steel cores and not SMC.
- Need to check with R&D if there are other reasons.
Kind regards,
Daan
https://cdn2.hubspot.net/hubfs/2795152/magnax%20axial%20flux%20high%20speed%20motor.jpg
Attachments
Hi Daan , welcome to the forum.
I see your focus is primarily the industrial/utility generator devices and understand why...its a big expanding market.
But you have any detail design or calculations for your motor , in a size suitable for a car size vehicle .?
IE; 100-200 kW, 400+ Nm, but in a size appropiate for car ?
Hub motors have not been viable to date, but maybe you have some ideas to change that , otherwise any direct drive train is preferable .
I Wish you success in your business .
I see your focus is primarily the industrial/utility generator devices and understand why...its a big expanding market.
But you have any detail design or calculations for your motor , in a size suitable for a car size vehicle .?
IE; 100-200 kW, 400+ Nm, but in a size appropiate for car ?
Hub motors have not been viable to date, but maybe you have some ideas to change that , otherwise any direct drive train is preferable .
I Wish you success in your business .
Daan, thanks for your response.
"Torque quadruples with the diameter" doesn't make sense, because quadruple just means to multiply by four (regardless of the amount of diameter increase?). My guess is that this is just a language issue; perhaps it was supposed to mean that torque varies as the fourth power of the diameter.
The working radius and the length of working flux gap both vary directly with the diameter, so the torque will rise as the square of the diameter for radial-flux motors of the same axial length... but they can be increased in axial length as desired without increasing complexity. In an axial-flux motor, if the magnet face radial width increases in proportion with the diameter (so it increases in radial dimension in proportion to diameter), that leads to a torque increasing as the cube of the diameter. I don't know how one would get to the fourth power, and I don't think it does.
- 4 m diameter is 2.5 times the diameter of the 100 kW unit, which suggests 39 times the torque of the 100 kW unit (by the fourth power of the diameter), which would be 625 kNm. At the cube of the diameter, it would be 15.6 times the 100 kW unit torque or 250 kNm. The 1243 kNm value for 8 discs is only 155 kNm each; that's 9.7 times the 16 kNm of the 1.6 m unit, so the torque is only rising by less than the cube of the diameter.
- 5 m diameter is 3.125 times the diameter of the 100 kW unit, which suggests 95 times the torque of the 100 kW unit (by the fourth power of the diameter), which would be 1526 kNm. At the cube of the diameter, it would be 31 times the 100 kW unit torque or 488 kNm. The 1254 kNm value for 5 discs is only 251 kNm each; that's 15.7 times the 16 kNm of the 1.6 m unit, so the torque is only rising by less than the cube of the diameter.
So, the design is predicted to scale up, but not as well as expected from the geometry. For me, this leads to two questions:
- How does this compare to existing radial-flux direct-drive products (not the geared units)?
- When the design is scaled down, how well does it work?
If you need to use a gearbox, it's just like any other motor... not a hub-motor.
Brian,
Daan's comment was " torque quadruples when the diameter doubles". .
..implying a second power (proportional to the square of the diameter) relationship.
And your comparison to the Leaf seems off also..
Leaf = 280 x 7.94 = 2223 Nm at the wheels
Magnax = 1000Nm at the wheels (potentially)....and if a hub motor it would be x2 wheels
So the Magnax hub drive car could have similar torque of the Leaf !
Daan's comment was " torque quadruples when the diameter doubles". .
..implying a second power (proportional to the square of the diameter) relationship.
And your comparison to the Leaf seems off also..
Leaf = 280 x 7.94 = 2223 Nm at the wheels
Magnax = 1000Nm at the wheels (potentially)....and if a hub motor it would be x2 wheels
So the Magnax hub drive car could have similar torque of the Leaf !
That does make sense; however, originally it didn't say that. I drafted my reply in email, from the emailed notification from the forum, and copied and pasted directly from his original text (shown below). I didn't notice that by the time I logged in and posted, Daan had edited his post. That part was my error.
Daan, thanks for the correction.
Joined
·
7 Posts
Hi guys,
Thanks for the comments.
I removed the in-wheel picture from the website since it gives people the impression that we do in-wheel powertrains. But we only deliver the AF motor. How customers implement the motor (in-wheel or chassis) is their choice.
The first motor we are going to release is the 265 mm version. This one will have the following specs:
- 5500 RPM
- 265 mm motor diameter
- 86 mm motor length
- Peak power: 300 kW
- Nominal power: 150 kW
- Peak Torque: 521 Nm
- Nominal Torque: 250
- Efficiency at nominal power: 91%, Peak eff. 98%
- Dry mass: 22,5 kg
- Cooling: water
So this gives a power density of 6,7 (nominal) and 13,3 (peak).
The reason why our power densities are so high is because of a new patented cooling system (which seems to be very effective and results in significant higher current density in the windings) and the use of grain oriented electric steel. (much higher flux density in the magnetic cores). And torque = proportional to current density x flux density.
Kind regards,
Daan
Thanks for the comments.
I removed the in-wheel picture from the website since it gives people the impression that we do in-wheel powertrains. But we only deliver the AF motor. How customers implement the motor (in-wheel or chassis) is their choice.
The first motor we are going to release is the 265 mm version. This one will have the following specs:
- 5500 RPM
- 265 mm motor diameter
- 86 mm motor length
- Peak power: 300 kW
- Nominal power: 150 kW
- Peak Torque: 521 Nm
- Nominal Torque: 250
- Efficiency at nominal power: 91%, Peak eff. 98%
- Dry mass: 22,5 kg
- Cooling: water
So this gives a power density of 6,7 (nominal) and 13,3 (peak).
The reason why our power densities are so high is because of a new patented cooling system (which seems to be very effective and results in significant higher current density in the windings) and the use of grain oriented electric steel. (much higher flux density in the magnetic cores). And torque = proportional to current density x flux density.
Kind regards,
Daan
Oh no I was thinking about this discussion and realized that when I calculated thrust I had multiplied by the tire radius - instead of dividing by it - thus understating thrust by a factor of about ten! That makes the red-highlighted values in this part of my post incorrect:
So, with corrected values, this example becomes:
Sorry about the confusion.
I was thinking about this discussion and realized that when I calculated thrust I had multiplied by the tire radius - instead of dividing by it - thus understating thrust by a factor of about ten! That makes the red-highlighted values in this part of my post incorrect:By the time I added the geared example, I had also reversed the torque and force values when I pasted them in! I blame distractions while I was doing this.
So, with corrected values, this example becomes:
If the torque varies as the cube of the diameter, and the largest possible hub motor for a car is about 400 mm in diameter, it would produce only 1.6% of the torque, or 250 Nm. With a minimum tire radius of about a third of a metre, that's 750 N of thrust per wheel... enough for a bicycle, but not a car. Maybe the torque reduction only varies as the square of the diameter (because hey, let's be optimistic): then you get 6.25% of 16 kN or 1000 Nm, for 3000 N of thrust. For comparison, a Nissan Leaf motor puts out 280 Nm, and that's multiplied by 7.9377:1 reduction gearing to produce 2223 Nm at the wheels, or 6668 N of thrust... which is what is needed for a couple of tons of car to accelerate acceptably.
A couple of 1000 Nm direct-drive motors actually are enough; however, as the followup post from Daan shows, a reasonable size of motor will not produce that much torque.Sorry about the confusion.
Good call.
The power and torque correspond to the maximum speed for both nominal and peak conditions (260 Nm at 5500 RPM is 150 kW; 521 Nm at 5500 RPM is 300 kW), which is like the published chart for the 1.6 m diameter motor, but unlike normal high-speed motor characteristics: normally, the peak current (and so peak torque) cannot be sustained past some point as speed rises, so torque drops off in proportion to speed and the power output is roughly constant for the remainder of the speed range.
So what is the form of the torque or power versus speed curve in this case? Are you saying that the peak torque is sustained for the entire speed range (due to cooling and magnetic path performance), which would explain the power density?
1 - 20 of 28 Posts
