## Idiots guide to water cooled mid motor making

Electric Motors and Controllers
larsb
1 kW
Posts: 304
Joined: Dec 10, 2014 5:12 am
Location: Gothenburg, Sweden

### Re: Idiots guide to water cooled mid motor making

I am trying to wrap my head around the target kV value.

Thermal limits set what you can get out of the motor, so i assume that it's required to keep max losses the same as in the original revolt motor. I think i can take care of the copper losses if the water cooling is successful - the magnets and eddy losses are what concerns me.

Power output is proportional to the rpms so if i can increase max rpms then power is increased the same amount.
I think that the limiting factors for using higher rpms are exponentially increasing eddy losses and the possibility of motor disintegration

Eddy current losses are proportional to rpm squared.
(edit, source:https://endless-sphere.com/forums/viewt ... p?p=240035 )

Eddy-current losses are also approximately proportional to the square of the lamination thickness
(source: www.emetor.org/glossary/lamination-thickness/)

So.. Going from 0.5mm lamination to 0.3mm keeping same losses:

Eloss0.3=k*0.3^2*rpm0.3^2
Eloss0.5= k*0.5^2*rpm0.5^2

Same losses:
Eloss0.3=Eloss0.5

0.3^2*rpm0.3^2=0.5^2*rpm0.5^2

(Rated)Rpm0.5=3120 (guesstimation from where efficiency drops off in pic from Revolt motors below)
image.jpeg (93.26 KiB) Viewed 185 times
rpm0.3=Sqrt(0.5^2/0.3^2*3120^2)

--> rpm0.3=5200
kV=85 is needed to get 5200rpms with my 72V system assuming loaded rpms are 85% of unloaded

The same goes for the magnet segmentation. dividing each pole into 5-7 segments will have a positive effect, should be possible to lower the losses and heating to between 25-50%.

I am also considering to make the motor 10-pole as it lowers the fundamental frequency by 40% (from 7x to 5x rpm). I found that recommended max frequency for 0.35mm lams is 400hz according to an application note for electric steels.

7pole pair is 5200/60*7=607Hz
5pole pair is 5200/60*5=433Hz

Max rpm to get 400Hz on 10poles is 4800. How the hell do the rc outrunners fit this recommendation? They don't..

I have seen some researched motors where 10 pole was slightly higher efficiency than 14 pole.

I wonder what the reason is for the popularity of 12slot14pole RC motors?
I guess it might give a slightly higher torque density than 10-pole but really don't know

I hope some of you can check my logic above, it's been a while since i was in school and it sure wasn't electric machines i studied
Last edited by larsb on Feb 07, 2018 2:25 pm, edited 3 times in total.

larsb
1 kW
Posts: 304
Joined: Dec 10, 2014 5:12 am
Location: Gothenburg, Sweden

### Re: Idiots guide to water cooled mid motor making

I will need to do a test wind to be able to get the correct kV so the rotor is next to start thinking about.

Magnets are 5x5mm
Original rotor airgap dia is 113.65 mm, divided by 14 "poles" gives each group of 1 pole and 1 side magnet a length of 25.5mm so that's perfect with five magnets and a 0.1 mm tolerance gap between each.

The 5 magnets of each pole group will be 2x side oriented magnets and 3x north or south. I found this halbach optimization where an optimal ratio is shown to be 19/26 of the sideways to radial pole magnets so this is the best possible ratio for the 5 magnets of each pole group.
image.jpeg (77.05 KiB) Viewed 148 times
image.jpeg (72.15 KiB) Viewed 148 times

larsb
1 kW
Posts: 304
Joined: Dec 10, 2014 5:12 am
Location: Gothenburg, Sweden

### Re: Idiots guide to water cooled mid motor making

I sketched a jig to position the N&S magnets on the rotor:
image.jpeg (68.43 KiB) Viewed 109 times
I plan to glue these magnets first. The sideways magnets will be placed after as these will be kicking and screaming to avoid their brothers

10-pole halbach rotor with N-S magnets marked:
image.jpeg (134.19 KiB) Viewed 109 times
Best ratio for sideways magnets/pole magnets is 42% (19/45) as shown in optimization graph above.
3/7 sideways magnets: 43% for 10-pole rotor
2/5 sideways magnets : 40% for 14-pole rotor