The base frequency of a motor isn't enough to show these effects. The pwm frequency may be enough, but the ripple caused by pwm isn't large compared to the main current. So these effects shouldn't change how you wind a motor. It's very important in the design of high frequency transformers though.
Miles' 90mm inrunner build thread
- Thread starter Miles
- Start date
major
10 kW
speedmd said:According to all the formulas I find so far, radius (distance from the iron) of the loop is key.
No, that is not the way I understand it. The distance between the conductor and the iron means little. Your equation is for a conductor in free space (or air). Our magnetic circuit or path is the iron (mostly as there is the air gap from rotor to stator). Your equation is what I think is the Biot-Savart law also known as the differential form of Ampere's law. What we need to consider is the integral form of Ampere's law which says the line integral of B around a single closed path is equal to the permeability of the medium times the current enclosed. Since our closed path is mostly iron at like 4000 times the permeability of air, the position of the conductor doesn't matter as long as it passes thru the magnet circuit (closed path). Most will concede that there may be some slight leakage flux linking the conductor inside the iron circuit but it is generally considered insignificant except with large air gaps (rotor to stator) and/or if saturation occurs in the iron.
speedmd
10 MW
Thanks for setting me straight on this. Sometimes with only a bit of knowledge it will get you off on the wrong track. 
Your correct the formula was applied to a free air electromagnet. Still interested in the real values of current variation throughout the wrap and how much if any effect it would apply here. At 95% efficiency levels we are searching for crumbs from here on. Much to learn.
Found a good link on magnet winding fill factor with various wire shapes. http://kgmagnetics.org/APNOTES-06/An-102.pdf
Your correct the formula was applied to a free air electromagnet. Still interested in the real values of current variation throughout the wrap and how much if any effect it would apply here. At 95% efficiency levels we are searching for crumbs from here on. Much to learn.Found a good link on magnet winding fill factor with various wire shapes. http://kgmagnetics.org/APNOTES-06/An-102.pdf
Just wondering if it would be worth using a thicker wire on the 2 turns section of the coil? It will reduce the total heat. It will reduce heat production on the part of the coil/tooth furthest from the ambient, conductively. I guess I should do the calculation 
Something like this:Miles said:Just wondering if it would be worth using a thicker wire on the 2 turns section of the coil? It will reduce the total heat. It will reduce heat production on the part of the coil/tooth furthest from the ambient, conductively. I guess I should do the calculation
Attachments
major
10 kW
speedmd said:The super high performing stuff I have seen was almost always wound like this which has me thinking.
The primary reasons for edgewound coils like that are ease of termination and consistent heating from turn to turn. There is also a lower voltage between adjacent turns in the coil compared to multilayer construction which benefits insulation concerns. I do not think the technique is used for magnetic reasons.
speedmd
10 MW
Thanks for the post. Lots of added issues making them also. Most of the ones I was looking at were high voltage applications and insulation was also a big concern.
Does the tooth need perfectly parallel walls. You could squeeze out a bit of room if you were not near saturation as on the bottom right image.
Does the tooth need perfectly parallel walls. You could squeeze out a bit of room if you were not near saturation as on the bottom right image.
speedmd
10 MW
Looks like the tapered tooth shaft image is from the Chevy Volt. Older article in car and driver has more images and info on it.
http://blog.caranddriver.com/we-build-the-chevy-spark-ev’s-ac-permanent-magnet-motor/
Press formed copper bars fit in the early version very tightly according to the article and welded on the far side.
http://blog.caranddriver.com/we-build-the-chevy-spark-ev’s-ac-permanent-magnet-motor/
Press formed copper bars fit in the early version very tightly according to the article and welded on the far side.
major
10 kW
speedmd said:
Yep, may well be. It is similar to the Remy HVH (High Voltage Hairpin) design. Ref: http://www.remyinc.com/docs/102509_WHITE%20PAPER_fin.pdf Hairpin is reference to the appearance or shape of the preformed coil usually from square cross section magnet wire. It is end inserted (axially) into the slots and twisted at the opposite end and then welded. Remy uses a 10 pole design which is distributed wound unlike your salient pole coils. They're hyped up pretty good over it and it is a great way to make an armature but is really nothing new. Cranking motor armatures have been made like this for decades. It is just inside-out and double turn per coil (effective). The Lightning race bikes use the Remy HVH250 motor and I got the Remy factory tour a few years ago. Cool stuff. That HVH410 motor is a monster....used in buses.
speedmd
10 MW
Hi Major
Looks like they shaped the core to fit wires for max fill factor and maximum heat transfer. Interesting tradeoff. Will need to simulate it at some point to see how it effects the numbers. In the volt article they were also talking about potting the end wrap with thermal epoxy. That Remy motor looks killer.
Looks like they shaped the core to fit wires for max fill factor and maximum heat transfer. Interesting tradeoff. Will need to simulate it at some point to see how it effects the numbers. In the volt article they were also talking about potting the end wrap with thermal epoxy. That Remy motor looks killer.
liveforphysics
100 TW
I checked the site this image came from, it says it's the capacitive current coupling component on the coil...
AKA, this is NOT what current density in a motor coil looks like, and the capacitive element effects in the freq range used by motors would be something like <0.0001% (likely much much smaller even) of current density effects in a motor application. I would recomend editing your earlier post so noob's don't get confused into thinking that is any form of representation of current density in a motors winding.
AKA, this is NOT what current density in a motor coil looks like, and the capacitive element effects in the freq range used by motors would be something like <0.0001% (likely much much smaller even) of current density effects in a motor application. I would recomend editing your earlier post so noob's don't get confused into thinking that is any form of representation of current density in a motors winding.
speedmd said:Image of coil current flow analysis.
liveforphysics
100 TW
Also, just to clarify for noobs who may have been misled or confused by the discussion in the last couple posts, I will summarize.
With respect to an electric motor operating at electric motor operating freq range, the following is true:
1. Amp-turns in the slot is all that matters for the flux produced.
2. How close the copper happens to be to the tooth is irrelevant to the flux produced as long as the copper is in the slot.
3. No configuration of stacking your coils in the slot makes any magnetic performance difference if you still have the same copper fill%.
With respect to an electric motor operating at electric motor operating freq range, the following is true:
1. Amp-turns in the slot is all that matters for the flux produced.
2. How close the copper happens to be to the tooth is irrelevant to the flux produced as long as the copper is in the slot.
3. No configuration of stacking your coils in the slot makes any magnetic performance difference if you still have the same copper fill%.
speedmd
10 MW
liveforphysics said:
Hi LFP
Thanks for checkin on that image. Very little detail I could find on exactly what they were showing and the degree of effect. Link failed to show popup caption. I edited and took out some of the previous posts. I would appreciate any info you could share on how you came up with your estimate. Not clear on how a segmented core adds to reluctance and permeability. I am not yet 100% clear on that, and its impact on flux. Any details how the fill factor is used in calculations is welcome.
I have since gone partially though my long bookmark list and see this much more clearly. Have to thank you for holding my feet to the fire here. I was assuming much too much from the little I know of inductors and transformers and confounded several key facts about the magnetic loop. I have been able to prove to myself that the concerns I have with lumping windings and distance from the iron are near meaningless unless I make some unrealistic assumptions on permeability /have a extremely low fill factor or many many more wraps and even then it works out to many decimal places. Fill factor is still a mistery. May reduce stray flux? Thanks again, much more reading to catch up on.
A logical extension of this idea would be to vary the widths slightly, as well, to give more equal cross-sectional areas....Miles said:Just wondering if it would be worth using a thicker wire on the 2 turns section of the coil? It will reduce the total heat. It will reduce heat production on the part of the coil/tooth furthest from the ambient, conductively. I guess I should do the calculation
speedmd
10 MW
Hi Miles
I like this one the best so far. Is it time to place a order? Count me in.
I like this one the best so far. Is it time to place a order? Count me in.
Making progress........
Time to compromise a bit...
Having dual aspect ratio winding cross-sections necessitates extra soldering joints at the centre of the coil - decided to leave this off the prototype. Also, I've changed to a 4t winding, for the first version, as the 6t winding required thinner rectangular section wire than is commonly available. The fill factor is down to 0.64 to allow for slot insulation and tolerance on the wire sizes etc. Insulation thickness is circa 0.05mm per face but the maximum size within tolerance for wire plus insulation amounts to 0.1mm per face, or 0.2mm, in total, on top of the nominal (bare) dimensions. Working on the basis that the maximum size within tolerance needs to be, at least, an interference fit.....
4t winding gives better cooling but, obviously, lower inductance (0.04537 mH at peak Eta). Hope that's not going to be problematic... Greater resistance from 'skin effect', too - need to quantify this..... Rm is 0.0305 Ohms in the simulation.
Time to compromise a bit...
Having dual aspect ratio winding cross-sections necessitates extra soldering joints at the centre of the coil - decided to leave this off the prototype. Also, I've changed to a 4t winding, for the first version, as the 6t winding required thinner rectangular section wire than is commonly available. The fill factor is down to 0.64 to allow for slot insulation and tolerance on the wire sizes etc. Insulation thickness is circa 0.05mm per face but the maximum size within tolerance for wire plus insulation amounts to 0.1mm per face, or 0.2mm, in total, on top of the nominal (bare) dimensions. Working on the basis that the maximum size within tolerance needs to be, at least, an interference fit.....
4t winding gives better cooling but, obviously, lower inductance (0.04537 mH at peak Eta). Hope that's not going to be problematic... Greater resistance from 'skin effect', too - need to quantify this..... Rm is 0.0305 Ohms in the simulation.
