Mid Drive - 180x60mm Outrunner

[dotrick]

Hello,

I'm trying to build a mid-drive motor for my new ebike... ...something like this :

Motor features :
- high torque >= 30Nm
- Continuous power >= 5kW
- weight <= 10kg
- Outrunner motor
- Rotor diameter <= 180mm

ebike features :
- torque at rear wheel >= 150Nm
- top speed >= 100kph
- Battery voltage <=100V
- Battery current <=55A

I did several simulations with emetor. The parameters and results are here :
- https://dl.dropboxusercontent.com/u/71224521/simul#40.pdf
- https://dl.dropboxusercontent.com/u/71224521/simul#40.txt

If you have any advice, don't hesitate!

 

[Miles]

Hi Dotrick,

Your stator and rotor yokes are a tad saturated......... Almost 7T!
Because of the rotor construction, you've ended up with very little material just where you need the most - between the poles.

Saturation for silicon steels is arrived at by 2T (for Cobalt steels 2.2T).
Maximum flux density in the teeth should be 2T. The stator yoke should be around 1.5T.

Perhaps choose a layout with a greater pole number? For 2000rpm, it would be reasonable to double the number of poles. This will also help by reducing the thickness of the back-iron required.

 

[dotrick]

Thanx Mike!

Yes I saw this results... Today, I did the simulation with FEMM. I have no problem...

I don't understand! :?

 

[Miles]

That's a bit more reasonable! You need to alter the inputs in the Emetor template to correspond with the geometry you used in FEMM.

Are you planning to use a laminated rotor?

 

[dotrick]

I changed nothing between Emetor and Femm simulation... :? Did you find differences between FEMM and EMETOR simulation? (for your application)

Yes, I think about a laminated rotor. I have a CNC for alu but not for steel. If I use a laminated rotor, I'm able to do an alu rotor. After I will insert laminated rotor in my aluminium rotor. And laminated rotor is better isn't it?

 

[Miles]

I changed nothing between Emetor and Femm simulation... :?

Well, the stator yoke is certainly different! Also, the rotor back-iron is less than the magnet thickness in Emetor and twice the magnet thickness in FEMM!

 

[Miles]

And laminated rotor is better isn't it?

It won't make much difference for your 15s 10p configuration. I was thinking only of the constructional aspect - the internal faceting.

 

[dotrick]

Yes, you are right! It was not the same shape...I adjusted the yoke height of the rotor and the stator. Now, it's better but the flux density is always too high... My rules of thumb is : tooth width/2 = yoke height stator and yoke height rotor ( for the same materials)

@Miles : Do you use only Emetor to design your motor?

 

[Miles]

It's better to have the stator yoke greater than half the tooth width. Your rule of thumb isn't appropriate for the rotor yoke.

I use FEMM, as well as Emetor, but for magnetic simulations only.

 

[Miles]

I quickly played around with your design. The problem is the high level of torque ripple - due to having a relatively low slot/pole LCM of only 30, I think. Narrowing the slot opening might help. As you can see, with the appropriate amount of Iron, the weight has increased..... As I said before, at 2000rpm it might be better to have more poles.

 

[Miles]

18slot 22pole double layer looks quite promising...... Might be worth trying that?

 

[madin88]

Dotric, nice to see you going to built your own motor for middrive use.
Do you also have thought of an inrunner? With an inrunner it would be more easy to cool stator and windings.

do you know the Plettenberg Nova motors?
If you could build your motor like the Nova 30, it would be a burner

take a look:
http://www.plettenberg-motoren.net/index.php/en/motor-serias-nova/motor-serias-nova-30

 

[dotrick]

18slot 22pole double layer looks quite promising...... Might be worth trying that?

18 Slots 22 Poles are you sure? or 18 Slots 12 Poles?
I will try this configuration...

Do you also have thought of an inrunner? With an inrunner it would be more easy to cool stator and windings.

do you know the Plettenberg Nova motors?
If you could build your motor like the Nova 30, it would be a burner

I prefere Outrunner because I want a motor with high torque density. If i choose an inrunner motor,I have to made a high ratio motor/rear wheel.
I didn't know the plettenberg motors. The Nova 30 is powerful!!! 160mm x 80mm : 105Nm for only 120V... I would like to see the torque/Power vs RPM curve...
And I think it's a water cooled system.

 

[Miles]

18slot 22pole double layer looks quite promising...... Might be worth trying that?

18 Slots 22 Poles are you sure? or 18 Slots 12 Poles?
I will try this configuration...

18t 22p never gets a mention but it looks good to me. Winding factor is only 0.902 but that's better than 0.866.... LCM of 198 ....... It doesn't look too bad for rotor losses, either. Maybe I'm missing something?

 

[dotrick]

18 slots 22 Poles: How do you do this with Emetor?

 

[Miles]

I'm not sure I understand the question? You'd set it up in the same way that you set up any other slot/pole combination.....

 

[Miles]

I had a quick try with the 18t 22p configuration. It still needs a lot of attention......

 

[Miles]

Second try. This time with non-linear materials to show I'm not "cheating"

 

[madin88]

I prefere Outrunner because I want a motor with high torque density. If i choose an inrunner motor,I have to made a high ratio motor/rear wheel.
I didn't know the plettenberg motors. The Nova 30 is powerful!!! 160mm x 80mm : 105Nm for only 120V... I would like to see the torque/Power vs RPM curve...
And I think it's a water cooled system.

I think these are more peak values than continuos, but for only 5kg motor mass still very impressive.
As far as i know these motors only will be produced at request, and the Nova 30 is about 4000-5000$
Some curves would be really interesting to see..

You plan <10kg for your motor. Do you think you could make it lighter?
To me it looks like you are building a motor like the minimonster http://endless-sphere.com/forums/viewtopic.php?f=31&t=40859&hilit=minimonster#p597073

 

[Miles]

You plan <10kg for your motor. Do you think you could make it lighter?
To me it looks like you are building a motor like the minimonster

I have to say that the overall concept is not at all clear, to me. Maybe Dotrick could give us a few more clues?

 

[dotrick]

Perhaps choose a layout with a greater pole number? For 2000rpm, it would be reasonable to double the number of poles. This will also help by reducing the thickness of the back-iron required.

Why do you want to increase the pole number? To reduce the cogging torque? I would like to understand because I'm working to reduce cogging torque and increase torque...

I have to say that the overall concept is not at all clear, to me. Maybe Dotrick could give us a few more clues?

No problem! I will do that for you!

 

[Miles]

Why do you want to increase the pole number? To reduce the cogging torque? I would like to understand because I'm working to reduce cogging torque and increase torque...

Increasing the number of poles gives you the chance to do both of those things.

 

[rjoe]

Aren't you worried about putting that much torque/power through a bicycle drivetrain?

 

[dotrick]

Aren't you worried about putting that much torque/power through a bicycle drivetrain?

Good question!
I will design the ebike frame for this motor. I will choose components (chain, gear, rear wheel ...) for this motor.

 

[dotrick]

You plan <10kg for your motor. Do you think you could make it lighter?
To me it looks like you are building a motor like the minimonster http://endless-sphere.com/forums/viewtopic.php?f=31&t=40859&hilit=minimonster#p597073

Yes under 10kg! More lighter? I will do my best What is your weight target for a mid drive motor?

I didn't know the minimonster before your post. Yes it's similar... but not the price 285$, it's the lower price i found for the lamination!

I have to say that the overall concept is not at all clear, to me. Maybe Dotrick could give us a few more clues?

What do you want to know more?
Electrical features? Mechanical features? ebike features?

A preliminary design :