Minimising rotor losses

larsb

1 MW
Joined
Dec 10, 2014
Messages
2,820
Location
Gothenburg, Sweden
The main problem killing my Revolt motor was eddy losses in the rotor. I guess the magnets aren't electrically isolated from each other and the back iron ring, also it looks like there's metal foil between the magnets.

I see something similar on the pics of your motor. When a rotor is built like that it surely is not for performance..

[Ed. This topic split from here: https://endless-sphere.com/forums/viewtopic.php?f=30&t=94975&start=25#p1395517 Miles ]
 
Maybe i have overread, but i am right that you going to replace the magnets and use JB marine as glue?

After the discussion here about the thing with insulation and segmentation of the magnets, i did check my Neumotor 8057.

I measured 12-30mOhm (yes milli ohm) between them so there is no insulation to the back iron, but at least they do not touch each other (1,5mm gap) and they are bended.
Funnily one magnet is insulated (multimeter shows mega ohms). They probably used more glue or it was already dry when they did install that one.

The point is i think about replacing the magnets by using 3 or 4 pcs (with non conductive coating) instead of a single one to lower the eddy losses.
Now the motor has around 350W of no-load losses at 5500RPM (80mm dia x 57mm long stator) and the temp goes up to 45-50°W in a few minutes without any load.
The good thing is i have a second rotor lying around i can play with :)

I already had contact with a china magnet supplier and i think i will ask for a quotation. Let me know if you ElectricGod or somebody else would have interest too.
 
I had a research paper with some good CAE simulations for the eddy losses in magnets and backiron but cannot find it.

It showed the eddy losses can become large in the magnets and that the backiron material in some cases is best as high conduction material, in some cases low conduction material. If i remember correctly the backiron didn't matter as much as it's further away from the stator fields.
 
I think for the IPM case it's a matter of distance also. The variable magnetic field from the coils are a bit away from the magnets, whereas in a surface mounted magnet motor the closest item to the coil fields is the magnet. BUUT, that's just an educated guess - my prime reasoning comes from the fact that the rotor get's hot quickly in my motor so there's for sure eddy losses in it.
 
I don't think that's correct, that's a typo in the underlined text.

How does segmenting something with the goal of keeping full contact change anything as far as conductivity goes?

How does higher conductivity lower current?

They were good finds, similar study as in the paper i read!

This figure is interesting, it only shows 1.5W heat losses per magnet.
image.jpeg
I think that's not enough loss to be a problem. The other link you showed had max 70000 w/m3 losses and that would be less than 1W for a revolt sized magnet.

I think higher losses than this is present in my rotor so the question is what is the root cause.. Is the slot geometry, coil field, magnet properties, all combined a lot worse than this simulation result?
 
ElectricGod said:
... Of course that means you have a motor with 200 poles to eliminate the eddy currents. Clearly there has to be a middle ground between ridiculous magnet counts and minimizing eddy losses. I think this may be the reason why hubs and outrunners have lots of magnets. This may be the place where eddy current losses are the worst so use lots of magnets. While in an inrunner, it's negligible so much fewer magnets is viable.

https://www.jmag-international.com/catalog/22_IPMMotor_MagnetLoss.html

Just to clarify: Segmented magnets is done for each pole. Doesn't mean more or less poles in the motor. An actual higher pole count equals increased eddy losses as the electric frequency goes up.

The magnet pole area and stator tooth area are normally almost the same, a divide in 4 or even 2 pieces of magnet per pole (and tooth) instead of one magnet gives almost the full improvement, after 4 the improvement is diminishing.
 
I'll leave the testing to you, see no reason to go that route, there's no connection to reality. Why is stator segmented? Because of eddy currents. Is a segmented stator higher or lower conduction in the stator slice plane vs a full lump of metal? It's lower.

It's the same physics behind both types of segmentations.
 
larsb said:
ElectricGod said:
... Of course that means you have a motor with 200 poles to eliminate the eddy currents. Clearly there has to be a middle ground between ridiculous magnet counts and minimizing eddy losses. I think this may be the reason why hubs and outrunners have lots of magnets. This may be the place where eddy current losses are the worst so use lots of magnets. While in an inrunner, it's negligible so much fewer magnets is viable.

https://www.jmag-international.com/catalog/22_IPMMotor_MagnetLoss.html

Just to clarify: Segmented magnets is done for each pole. Doesn't mean more or less poles in the motor. An actual higher pole count equals increased eddy losses as the electric frequency goes up.

The magnet pole area and stator tooth area are normally almost the same, a divide in 4 or even 2 pieces of magnet per pole (and tooth) instead of one magnet gives almost the full improvement, after 4 the improvement is diminishing.
Edit:
As larsb points out a couple posts down, I am wrong here.
[strike]It's actually increased stator hysteresis losses not stator eddy losses from higher electrical frequency.
[/strike]
With regards to eddy current losses in the rotor magnets, I think it is the same as the stator eddy current losses like larsb said. The stator creates a magnetic field that moves synchronously with the rotor, but there is still a magnetic field (from the stator) that is moving relative to a conductor (the magnets) so there is an induced current in the conductor (magnets). These eddy current losses in the magnets seem to range from trivial in lots of motors like the ones EG posted to large like in larsb's Revolt and in Emrax motors.
 
@ ElectricGod

you mentioned non of the motors you posted pics from have electrically insulated magnets.
Couldn't it be that the magents have some kind of transparent coating for electrical insulation?
Have you measured the resistance between two magnets or what gives you this idea?
 
thepronghorn said:
"..It's actually increased stator hysteresis losses not stator eddy losses from higher electrical frequency."

@thepronghorn:As the frequency is one of the factors in below formula i doubt that you are correct. (It's from Wikipedia).
Eddy current losses are proportional to the square of the frequency:


Or from this link:
https://www.quora.com/Why-does-higher-frequency-cause-higher-eddy-current-losses-in-an-iron-core-transformer

or this link from a 1975 book (i particularly like the sentence ""it is well known that..):
https://onlinelibrary.wiley.com/doi/pdf/10.1002/eej.4390950202

It would be more fun to see some motor testing than this. Couldn't "electricgod" move on to the actual motor testing (or are you on vacation without access to motors just like i am? :D ) What do we know about the motor before it's tested? After all, the Revolt motor might be just fine despite their track record.

I must say that even to refer to someone as "electricgod" is peculiar. Remember when Prince suddenly became "the artist formerly known as Prince"? Couldn't Electricgod do something similar and become "Electricsomewhatknowledgeableperson Formerly Known As Electricgod") Short for that would be EFKAE or even FK :D
 
If you reread your own start of this thread you begin to review a motor and say it is better than another motor, but you have no backing this. It's exactly what causes people to buy bad stuff after they read it, see that you have 2000 posts and some builds under your belt and call yourself electricgod. No sane person would do that unless they are at least semi-skilled, right?

You don't know so much on motor design, I guess we're all here to live and learn. Just stick to facts, otherwise your motor review capabilities are doubtful.

I think the tone in your posts sometimes corresponds to the arrogance of your alias and because of this your alias seems like you mean it.

As for me letting others solve my problems: I don't have a problem to solve really - i know my revolt motor is bad. I am riding it until it breaks, then that chapter is closed. My interest in this thread is more to help others to avoid bad products. But when you can prove a root cause to the bad performance then i am really interested since it gives us all knowledge.

I think you got madin88s' question wrong by the way: did you measure resistance on all the motors you showed pics on? If you did i missed it. I believe you have only measured the Revolt motor, the facts on isolation or conductivity is not known for the others.

And if you read this through, it might be enough to set the conducting vs insulating magnets straight:
image.png
 
ElectricGod said:
...I can see they have no deliberate electrical isolation between the magnets.

...I'm pretty sure none isolate the magnets electrically.

...I have no idea what the resistance is between the magnets in any of them.

...It seems that no one I can find in my collection cares about electrical conductivity or not between the magnets."
Yes exactly, you seem to state that there's no isolation. But you don't know. But you're pretty sure.

It's the same with the magnet coverage where Revolt according to you is better. It's not necessarily better to have 100% coverage, but you don't know this. So you state that it's better.

I can say that the winding in your motor looks better than in my original Revolt. But my fill was really low at 22% and i know this. You really don't know how much it is since the coating fills the voids of all crossing wires.

Etc etc. That's my point. Just stick to what you know. I'll backpedal out of this thread until you have some facts like phase resistance, inductance and noload current and some testing done.

And what about high conduction and segmentation between magnets? Is it good or bad to include a copper tape or even to solder a copper wire?

Link to full paper:
https://scholar.sun.ac.za/bitstream/handle/10019.1/45192/wills_reducing_2010.pdf?sequence=2&isAllowed=y

The magnet and backiron segmentation in this paper makes a 4% efficiency increase. If a motor can disperse 2000w heat at 85% efficiency then this would mean you can increase continuous output from 13kW to 18 kW with this improvement. :D

I think you just might kill your motor with the copper wire test if you run it too long at a decent amperage. Or during the soldering.
 
larsb said:
thepronghorn said:
"..It's actually increased stator hysteresis losses not stator eddy losses from higher electrical frequency."

@thepronghorn:As the frequency is one of the factors in below formula i doubt that you are correct. (It's from Wikipedia).
Eddy current losses are proportional to the square of the frequency:
Skärmavbild 2018-07-19 kl. 22.18.56.png

Or from this link:
https://www.quora.com/Why-does-higher-frequency-cause-higher-eddy-current-losses-in-an-iron-core-transformer

or this link from a 1975 book (i particularly like the sentence ""it is well known that..):
https://onlinelibrary.wiley.com/doi/pdf/10.1002/eej.4390950202

It would be more fun to see some motor testing than this. Couldn't "electricgod" move on to the actual motor testing (or are you on vacation without access to motors just like i am? :D ) What do we know about the motor before it's tested? After all, the Revolt motor might be just fine despite their track record.

I must say that even to refer to someone as "electricgod" is peculiar. Remember when Prince suddenly became "the artist formerly known as Prince"? Couldn't Electricgod do something similar and become "Electricsomewhatknowledgeableperson Formerly Known As Electricgod") Short for that would be EFKAE or even FK :D

Sorry yes you are right I was mistaken.
 
Back
Top