speedmd
10 MW
Possible is to cast it once wound in high temp epoxy with a bit of long glass for structural integrity and some thermal compounds for heat transfer.
Dual rotor axial flux motor design
- Thread starter Miles
- Start date
speedmd
10 MW
Chevy is doing it on the sparks end turns. At least on the pre production run. They must have had some good reasons, as they claim it was for better heat transfer based on what was written in the article. I would prefer it mechanically fastened as miles pictured earlier in the thread.
speedmd
10 MW
h0tr0d said:Miles said:
If those rings are made of metal wont they brake the motor because of the rotor magnets are moving so close?
Also, really awful thermally speaking...
Hi h0tr0d
Not sure I can agree at this stage with so much still to be worked out/presented. Center (end turn area) is totally open to do whatever in terms of air flow as is the outer end turn area. Interesting type motor design, and looking forward to see what miles does with this build.
h0tr0d said:If those rings are made of metal wont they brake the motor because of the rotor magnets are moving so close?
Also, really awful thermally speaking...
They don't have to be metal.
I think that the possibilities for air cooling are pretty good with this topology... Not to mention the possibility of a having a 'wet' stator.
The big advantage is that the completely straight flux path in the core modules allows taking full advantage of grain-oriented steels. With rectangular section laminations, there will be almost no wasted material, justifying the use of high performance transformer steels.
liveforphysics
100 TW
I think it's brilliant Miles.
Does the straight flat tooth enable good flux containment to prevent eddy losses in the ribbon?
Do you have a way to model the grain oriented material for flux path behavior?
Pardon my ignorance, but can't you get enough of a permittivity advantage with a grain oriented cobalt steel that you would actually want the iron/copper ratio's skewed towards having a bit more copper space?
Do you have a way to model the decrease in core losses from going grain oriented?
Does the straight flat tooth enable good flux containment to prevent eddy losses in the ribbon?
Do you have a way to model the grain oriented material for flux path behavior?
Pardon my ignorance, but can't you get enough of a permittivity advantage with a grain oriented cobalt steel that you would actually want the iron/copper ratio's skewed towards having a bit more copper space?
Do you have a way to model the decrease in core losses from going grain oriented?
Thanks Luke,
The present iron to copper ratio was arrived at by:
-Concern over eddy losses, as per your first point.
-Keeping close to the 0.2 slot opening to slot pitch factor for minimal torque ripple.
-Taking advantage of the lower iron losses from the topology and the use of G.O. steel, together with the dual magnets, to milk the magnetic circuit to the max.... For 4t, the conductor cross-sectional area is 7 sq.mm for this design compared to 4.8 sq. mm for my 125mm inrunner...... Copper losses are linear to resistance but the square of Kt (equal no. of turns). If I've gone too far, I'll increase the slot size.....
Modelling is tricky for axial flux. I'm stuck with using linearised circumferential sections in FEMM. I wish I had access to 3D simulation software...
The present iron to copper ratio was arrived at by:
-Concern over eddy losses, as per your first point.
-Keeping close to the 0.2 slot opening to slot pitch factor for minimal torque ripple.
-Taking advantage of the lower iron losses from the topology and the use of G.O. steel, together with the dual magnets, to milk the magnetic circuit to the max.... For 4t, the conductor cross-sectional area is 7 sq.mm for this design compared to 4.8 sq. mm for my 125mm inrunner...... Copper losses are linear to resistance but the square of Kt (equal no. of turns). If I've gone too far, I'll increase the slot size.....
Modelling is tricky for axial flux. I'm stuck with using linearised circumferential sections in FEMM. I wish I had access to 3D simulation software...
h0tr0d
1 kW
- Joined
- Apr 28, 2012
- Messages
- 460
http://www.emworks.com/product/preview/1/EMS?no_redirect=true
See the videos, looks pretty good.
Piratebay it or start selling your radial designs and/or prototypes to make money to buy the software.
May the (electromagnetic) force be with you, Miles!
See the videos, looks pretty good.
Piratebay it or start selling your radial designs and/or prototypes to make money to buy the software.
May the (electromagnetic) force be with you, Miles!
Yes. I'd need Solidworks or Inventor too, though.... I did try getting 3D Systems interested in EMS.....
Motor software is about as expensive as it gets... The cheapest license of JMAG is 19,000 Euros..... and that's just for "2D"......
Motor software is about as expensive as it gets... The cheapest license of JMAG is 19,000 Euros..... and that's just for "2D"......
Kingfish
100 MW
^ +1
The part I found interesting is the discussion about the slotless toroidal stator in relation to the axial pancake layout, and it gave me pause to reflect:
With regards to radial flux layout, unless I’m mistaken – that is a toroidal layout albeit across the breadth of width having the maximum dimension as opposed to length in the case of axial being the maximum, with the implication that width governs strength in the perpendicular to the axis whereas length parallel to the radius governs transmission along the axial plane. It is a good point to raise about physicality and especially about deflection upon thin stators that brings to bear material sciences and mechanical strength against shock.
I haven’t quite bought off that a toroidal-wound axial-flux stator provides more efficiency because of the greater employ of copper, although willing to accept the value of torque is gained. Which is more important? and how can we tell with the free tools at hand? It must be modeled.
Miles, I have come to the point where I must depart from FEMM alone to answer questions as this and have moved onward to Inventor for more advance simulation; I would not have mentioned it at all until you brought it up. In a way I feel we’re solving the same problem though from different ends and applications. It is a good thread. 8)
Lots to consider, much to appreciate. KF
The part I found interesting is the discussion about the slotless toroidal stator in relation to the axial pancake layout, and it gave me pause to reflect:
With regards to radial flux layout, unless I’m mistaken – that is a toroidal layout albeit across the breadth of width having the maximum dimension as opposed to length in the case of axial being the maximum, with the implication that width governs strength in the perpendicular to the axis whereas length parallel to the radius governs transmission along the axial plane. It is a good point to raise about physicality and especially about deflection upon thin stators that brings to bear material sciences and mechanical strength against shock.
I haven’t quite bought off that a toroidal-wound axial-flux stator provides more efficiency because of the greater employ of copper, although willing to accept the value of torque is gained. Which is more important? and how can we tell with the free tools at hand? It must be modeled.
Miles, I have come to the point where I must depart from FEMM alone to answer questions as this and have moved onward to Inventor for more advance simulation; I would not have mentioned it at all until you brought it up. In a way I feel we’re solving the same problem though from different ends and applications. It is a good thread. 8)
Lots to consider, much to appreciate. KF
Kingfish
100 MW
I elected not to purchase the Simulation package: I could not justify the cost for this project. Instead the plan is to use their Cloud Services. That said, Autodesk does not appear to have a robust electrical simulation plan equal to their mechanical/structural/civil package. I’m presently looking into Em by EmWorks.
The guys at work use Solidworks but I could never get a straight answer about the costs for software, and they are after all behind a corporate wall of investment, one that I cannot match. The Cloud Services offer a path forward without throwing a lot or money away.
Between the two sims, we can determine heat, hysteresis, strain, magnetic study, conduction, convection…
Make no mistake, I’ll be giving up a few nights at the pub to afford this new GF.
If it builds me a new electric horse? So be it! 8)
Hi ho Silver, away! KF
The guys at work use Solidworks but I could never get a straight answer about the costs for software, and they are after all behind a corporate wall of investment, one that I cannot match. The Cloud Services offer a path forward without throwing a lot or money away.
Between the two sims, we can determine heat, hysteresis, strain, magnetic study, conduction, convection…
Make no mistake, I’ll be giving up a few nights at the pub to afford this new GF.
If it builds me a new electric horse? So be it! 8)
Hi ho Silver, away! KF
https://www.google.com.au/#newwindow=1&q=low+inertia+axial+flux+servo&safe=active
low inertia axial flux design paper.
Zappy
low inertia axial flux design paper.
Zappy
Fibre filled PEEK seems like a good candidate for the chassis. It's properties are approaching soft metals. Its price is up there with gold plated silver wire though. G.O. steel would seem to have a very large advantage for this design. The stator stack could literally be guillotined up (very accurately) wasting very little precious cobalt.
Zappy
Zappy
dotrick
10 W
Miles,
I would like your opinion about a radial motor design vs axial...
Today, I took a look on the datasheet of EVO motor (http://www.gkn.com/driveline/our-so...ocuments/Datasheets/AF-130-Tech-Sheet-V13.pdf). The W/Kg is very good!!! It's a big motor not for a bicycle but it seems a good idea for bicycle application...
To make this motor it seems easier than radial motor, do you agree with me?
item : Radial motor (outrunner) / Axial motor
airgap (<1mm) : difficult / easier
Cooling : difficult / easier
Torque : Very good / good
end winding : Difficult to cool / no end winding
Density : good / excellent
...
Let me know!
I would like your opinion about a radial motor design vs axial...
Today, I took a look on the datasheet of EVO motor (http://www.gkn.com/driveline/our-so...ocuments/Datasheets/AF-130-Tech-Sheet-V13.pdf). The W/Kg is very good!!! It's a big motor not for a bicycle but it seems a good idea for bicycle application...
To make this motor it seems easier than radial motor, do you agree with me?
item : Radial motor (outrunner) / Axial motor
airgap (<1mm) : difficult / easier
Cooling : difficult / easier
Torque : Very good / good
end winding : Difficult to cool / no end winding
Density : good / excellent
...
Let me know!
I'd say axial flux motor construction is more difficult. With radial, the static forces are integrally balanced and centered. With axial, you have to design a stiff enough structure to ensure a consistent airgap is maintained. If the magnets and core come into contact during assembly or through accident, that's a non-trivial event
major
10 kW
dotrick said:
The EVO has end windings as seen here:
Figure lifted from: http://www.docstoc.com/docs/106122369/Transient-Thermal-Modeling-of-an-Axial-Flux-Permanent-Magnet-_AFPM
dotrick
10 W
Miles said:I'd say axial flux motor construction is more difficult. With radial, the static forces are integrally balanced and centered. With axial, you have to design a stiff enough structure to ensure a consistent airgap is maintained. If the magnets and core come into contact during assembly or through accident, that's a non-trivial event
If you use your topology the axial force are unbalanced yes it's true.
If you use the topology with one rotor and 2 stators, the axial force are balanced.
See this doc : https://dl.dropboxusercontent.com/u/71224521/AFPM - Topology article1380714510_Mahmoudi et al.pdf
Miles,
What is the latest performances do you think you can get with your motor? I'm interested...
major said:
Thank you for this useful information for me!!!Do you know the topology of the EVO motors? 2 stators? 2 rotors?
Lebowski
10 MW
Miles said:I'd say axial flux motor construction is more difficult. With radial, the static forces are integrally balanced and centered. With axial, you have to design a stiff enough structure to ensure a consistent airgap is maintained. If the magnets and core come into contact during assembly or through accident, that's a non-trivial event
And this is why I make mine with an air core....
