Idiots guide to water cooled mid motor making

The LRK winding is not necessarily linked to using double controllers but since LRK is only 1-layer winding (every other tooth is wound, the rest are empty) that means if you fill half of the slot with one LRK winding, then the empty teeth can have another set of windings / "another motor" wound on them.

In sensored motors it should mean there must be two sets of halls for the two motors but i really don't know how the influence of the diferent coils will be.

Maybe John in CR can support :?:
 
Maybe like most other instances of dual drive rides? It seems people really fear the sync problems if going with more then 1 drive. But from people who actually done it, they seem to report back no sync issues and no need to do anything. Hook things up and it will work. Or so it seems.
 
I don't agree with some discussed theoretical dual drive limitations (as you say).

In the case of front and rear hub motors they strive to drive a big damping weight forward together, no reason for any issues.
in case of connected motors on same belt/chain it's still a large resistance to both motors. It might be more sensitive to oscillations but i don't see a high risk.

But for magnetic fields in close distance with possible induced currents it might mean trouble and bad efficiency..
Who knows without a scope and a dyno :D
 
Motor design ready :D Fun work and relatively easy. Only thing left is to decide how to wind. It really is a tough choice as i want higher kV than before (70-90) and i seem to have to wind parallel thick wires to get it.
1 stator hub
1 stator
Copper wire
2 bearing hubs
2 sideplates
2 40x52x7 bearings
some magnets..
stator hub final1.JPGstator hub final2.JPGstator hub final3.JPGstator hub final4.JPGstator hub final5.JPG

There seems to be a lot of dead space in the motor, that's to fit the endturns and solder connections with 10mm clearance from coils to the rotating sideplates.

I think one reason for the heating of the RV120 motor i rewound might be that the end turns create eddy losses in the sideplates so i am going for a larger clearance in the new motor.
 
Will the coils you leave "empty" do some work by winding it like this (think about max torque and iron saturaion)?
what would be the advantages / disadantages of the LRK wind vs regular winding every teeth?
 
madin88 said:
Will the coils you leave "empty" do some work by winding it like this (think about max torque and iron saturaion)?
what would be the advantages / disadantages of the LRK wind vs regular winding every teeth?

The primary benefit of LRK seems to be that it provides an "electrical gear reduction", reducing kV whilst increasing torque.

(Note:No expertise here; just my interpretation of what I read a the link above.)
 
Buk___ said:
madin88 said:
Will the coils you leave "empty" do some work by winding it like this (think about max torque and iron saturaion)?
what would be the advantages / disadantages of the LRK wind vs regular winding every teeth?

The primary benefit of LRK seems to be that it provides an "electrical gear reduction", reducing kV whilst increasing torque.

(Note:No expertise here; just my interpretation of what I read a the link above.)
The "electrical gear reduction" is just a feature of both due to the number of pole pairs. dLRK or winding all teeth sees two peaks of magnetic flux from the magnets of the rotating rotor. The Flux curve is more like a camel back. With LRK the winding sees a pure sine wave with a normal sine like peak. Or in math compare the peak of dLRK= sin(x) + 1/3 * sin(3*x) with LRK = sin(x).
lrkvsdlrk.PNG
 
My understanding is that the flux path in LRK(alternating teeth wound) is the same as in dLRK (all the teeth are wound)
The path of the flux will still be through both the wound and empty teeth and contribute to the torque.

@Tecnologic: ...
Edit:See pdf:s below. The Single layer winding is more trapezoidal in these.
 
i do not find it any more, but it was in this thread at rc groups:
https://www.rcgroups.com/forums/showthread.php?2940879-12-teeth-stator-DLRK-improvement-proposal-for-the-motor-designer-masters/page2

or as a ref: https://www.rcgroups.com/forums/showpost.php?p=38837153&postcount=34 last sentence

But this i mainly a geometric thing. Consider one pole pair passing a coil, in time this would give a flux linkage in sine form. The same pole pair passing a U shape coil (same as DLRK on two teeth) u see 2 peaks in the same direction right after each other. The first peak is when the Northpole is right over the first part of the coil. In this situation the south pole is just before the 2. part of the U coil. Next are booth poles in between the coils, thats the slight dip and then the south pole is over the 2. part of the coil wich is wound in the other direction as the 1. part so the flux linkage keeps the sign and has its second peak.

I hope this explanation is kinda understandable :). My graph is also more an estimate. The thing is dLRK is more trapezodial than LRK in its Flux linkage over angle wave form.
 
When searching for LRK winding i was going over a thread in the german pedelec forum.
A guy planned to do such winding scheme because it would be less work and an easier job, but others were saying it would be not as efficient as winding every single tooth, and that it is an old technique (used 15 years ago).
What i later found out is this winding could still be a good choice for motors with very long stators, but for small stators it's the wrong choice.
It also depends on the number of teeth and poles if it may be a good or rather bad choice.

here the thread:
https://www.pedelecforum.de/forum/i...e-100-kaufen-oder-selbst-wickeln.24281/page-2
 
As the end turns grow with number of layers in the winding it's true that LRK is not a good scheme for short stacks. The torque building wire length gets low compared to the end turn length.

In my 60mm stack there is a ratio of effective wire / end turns like this:
60/(Pi*(9+1.25*2)/2)=3.3 for dLRK
60/(Pi*(9+1.25*4)/2)=2.7 for LRK

So the dLRK is better than LRK in this regard. I think it should be easier to pack the slots in LRK wind so i will probably get back some of this difference due to better copper fill. :D
 
I found this presentation:
View attachment LRK vs DLRK.pdf

Eddy losses and harmonic content shown are higher in LRK than in dLRK. I cannot judge how many percent that would mean in efficiency without knowing the power used so is it relevant or not...?

I will possibly go back to dLRK winding plans because of this
 
I got the paper on full winding/dLRK vs alternate tooth winding/LRK:
View attachment Dlrk vs lrk paper.pdf

The LRK type motor in this paper has 17% higher phase resistance for less than 10% increase in torque and also larger eddy losses.

Edit: I calculated the stack height where the LRK is better in torque vs copper losses and realise i made an error before:
Torque is proportional to I and copper losses to I squared. Under the assumption that the LRK wind gives 3.5% higher torque (following the winding factor difference and also the results from the turn calculator sheet) then to get same torque in dLRK the current needs to be increased 3.5% and copper losses will then increase 7.2%-->
The phase resistance can be 7.2% higher in LRK wind (1.035*1.035) and still give same performance/copper loss. In my stator the break point is roughly a stack of something like 85-90mm (see below) when making an average turn length comparison.
(87+Pi*(9+1.25*8)/2)/(87+Pi*(9+1.25*4)/2)=0.072
I compare the halfturn length for LRK and dLRK
9:stator tooth width
87:stack height
1.25: thickness of one layer
8: Average number of layers in single layer wind
4: Average number of layers in double layer wind

dLRK wind is better than LRK in torque per copper loss in my motor as my stack height is only 60mm


Other LRK vs dLRK comparisons:
LRK vs dLRK is verified in testing in this article on small RC motors (not the best type for LRK):
image.jpeg
dLRK is better in this testing also.
 
this is a good write up thanks for sharing. Considering the losses makes sense, for why u see dLRK so often.
 
There are fantastic winders out there. I salute them! Check out this thread by user 'modisc731':
https://www.helifreak.com/#/topics/487902

I think i'll aim for at least 60% fill in the rewind after being inspired by the thread above. That's 28turns of the 2.5x1.25 wire.

My recalculation of the copper losses above makes the double LRK again the strongest candidate for the winding (if the resulting torque is indeed higher.)
 
larsb said:
There are fantastic winders out there. I salute them! Check out this thread by user 'modisc731':
https://www.helifreak.com/#/topics/487902

I think you must be a member in this board as the link doesn't work, but google directed me to some threads.
Man, this guy really knows how to improve the winding and copper fill^^
thanks for sharing
 
I found this paper when searching for ways to lower eddy current losses:
View attachment Minim of eddy losses in 12slot motor (Dajaku - Gerling).pdf
Clever and could be simple to implement with almost no decrease in torque.

The reason i am a bit concerned for the eddy losses is the composite rotor i will be doing, it will not have great thermal mass or conductivity so rotor/magnets will be more sensitive to heat.
 
I did a check to compare the copper fill with different wires.

From left to right slot in picture, absolute max theoretic figures below.
Round wire dia 1.5mm:
58 turns per tooth, pi*1.5^2/4*58*2/267= 77%

Square wire 2.5x1.25:
32turns per tooth, 32*2.81597*2/267= 67%

bundled wire 7x1mm dia:
28 turns total 28*7*pi*1^2*4= 58%
fill study.jpg

Single round wire will be the best for fill in theory. But i doubt it's possible to route six layers well without many crossings and it gives an absurdly low kV.

Using the 7x1mm bundled wire in a single layer/LRK wind is attractive since it's a lot less work and gives a higher kV.
28(turns)*6(slots) is only 168 turns in total vs 30*12=360 turns to wind for the square wire/double layer. The smaller wires should also be better for eddy losses.

The dual LRK / dual controller setup let's me double the controller power available and gives me some redundancy at the cost of a lot of added complexity. And it's fun :D (but my battery will not keep up :wink: )

I am leaning towards trying the double LRK with 30 turns of the square wire, 63% fill. That's 7% higher than in the previous wind so it might not be doable since its 10 wires extra in the slot compared to previous wind. I'll cut the wires for 30 turns and see, it can always be cut shorter later on.
 
I've been reading up on 6-phase motors, it's hard to draw conclusion.
I think to turn the motor into a 6-phase with this stator has small benefits. The winders that have tried 6-phase on stators from 3-phase motors did not get huge improvements. I think the whole stator/ rotor(motor) needs to be redesigned to get the big benefits like higher torque and lower losses.

In my case it's best improvement seems to be the possibility to use two controllers and double the available controller power.

This paper by Gerling on loss reduction on single layer and double layer winded motors shows 2x higher harmonic losses in the single layer wind vs double layer. It focuses on using flux barriers to lower harmonics but contains a lot of interesting data.
View attachment Harmonics Reduction single layer vs dual layer - gerling.pdf It's quite a good paper to see the harmonic content difference.
 
larsb said:
In my case it's best improvement seems to be the possibility to use two controllers and double the available controller power.

My understanding is this is only practical advantage. AFAIK there is no real inherent efficiency or performance benefit, maybe slightly reduced torque ripple. "John in CR" on this forum has posted a bit about the 6 phase hub motors he has used.
 
Punx0r said:
larsb said:
In my case it's best improvement seems to be the possibility to use two controllers and double the available controller power.

My understanding is this is only practical advantage. AFAIK there is no real inherent efficiency or performance benefit, maybe slightly reduced torque ripple. "John in CR" on this forum has posted a bit about the 6 phase hub motors he has used.


Haven't done my homework on 6 phase motors, or rather 2 x 3 phase motors in one casing. But I have been intrigued by them since I first read about the hubmonsters. Pulling from memory in John's hubmonster thread I think there was also benefits from smoother take off. Don't know if the motor comparison spreadsheet is understandable for you - my lack of schooling makes it all greek, but if you can make some sense of it, I think the hubmonster is in that spread sheet. Maybe you can compare the 6 phase to other 3 phase hub motors to get a better understanding if there is any gains in efficiencies, torque or anything else that matters?

Regarding the usage of two controllers rather then one, to me that seems like a huge deal. Rather then running one expensive "Adaptto style" controller to surpass 10 kw, if choosing 6 phase motor then any two decent china controllers will deliver stable combined 10+ kw power. And with hi amp run, like accelerations from standstill, that will put less strain on each controller while riding like a hooligan. In my head dual controller should give some benefits in regards to blown FET's or similar issues that can take place when pushing hard for max amp and acceleration.
 
I think comparison would be difficult as the motors are different in many other ways that will have effects on performance & efficiency. You'd need identical motors but just wound differently (and I can't think of any examples of this).
 
Well I wasen't thinking a direct comparison. More that if a 6 phase dd hub was compared to sort of similar 3 phase dd hub from the motor spreadsheet there could maybe come a conclusion?

Like many dd hubs have efficiencies around high 80's, while the hubmonster had 94 % efficiency.

AFAIK there ain't a whole lot of these 6 phase motors around. And to find two similar motors rc style in 3 phase and 6 phase will be a tall order.
 
There was a lot of discussion a couple of years ago about a proprietary ebike system that used 5 phases. The manufacturer claimed all sorts of supposed benefits, many of which didn't stand up to scrutiny. My recollection is that after a lot discussion and analysis the conclusion was 5 phases over 3 offered no tangible benefit, but was a lot more expensive as the controller was oddball.

Whereas the purpose of running 2x3 phase is the exact opposite: maximising the utility of existing, inexpensive controllers.
 
Back
Top