Coil shape for axial flux generator

[michal]

Hello everyone,
I'm in the process of making an axial flux generator for my VAWT , but find it difficult to digest all the information available on various websites. I'd like to know what is the best shape of coils, depending on the configuration of magnets.
The information I found on different forums is sometimes confusing and since I'm not an expert I hope someone here will help me to get my head around it.
I understand that for a single disc configuration (like the one below for example) the coil should be big enough, so that the coil legs pass over two magnets and cut thru the flux lines.
 
Plan view

 
Section

My problem is that I want an axial flux generator with two discs of magnets, yet found information on some website that the coil shape should be the same for this case also.
As the direction of magnetic field would be different on two sides of coil, would the the generated currents not cancel each other?

 
 
With the mainly vertical direction of magnetic flux (depending on magnet strength, gap size, etc…)  wouldn't I be better to have the coil shaped similar to the shape of magnet?
Please let me know what you think about the best configuration.
Many thanks in advance…

[electrondady1]

hello michal,
 i build verticals as well as they are best suited to my circumstances .
your showing a 3 phase configuration.
3 coils for every 4 mags
and describing a dual rotor axial flux alternator.
i would suggest that you could make the centre hole of the coil the same size or slightly smaller than the magnet and let the shoulders of the coils touch each other .

[michal]

Thank you for quick reply,

And regarding the shape of the magnets  - what would you recommend?

For single rotor I understand the advantage of wedge/rectangular shape, but for double rotor design, would the circular shape be a good idea ?
(I wonder if the coil would be exposed to more uniform magnetic field  ---> round magnet= round coil???)

Any opinion?

[Flux]

I don't think there is an answer to your question. For axials the most logical starting point is with trapezoidal magnets and coils but if you choose to use other shape magnets then logic goes out the window.

For a high efficiency conventionally loaded alternator your rectangular magnets would probably work best with the trapezoidal coils. but for direct battery charging, the circuit resistance has a lot to do with how much you can get out of the thing and changing to rectangular coils would reduce resistance way below the long turns of that trapezoidal coil.

If you are rectifying for direct battery charging then there is little doubt that coils roughly the shape of the magnets is the way to go, so rectangular coils for rectangular magnets and round coils for disc magnets. Even here resistance is a big factor and you may be better off squashing round coils to ellipses. Any turn smaller than the magnet will not link all the flux and won't contribute full voltage, but it will contribute very little resistance and you usually get an overall gain.

How you load the thing has a big influence on how it behaves and use of material and how much you can get out without frying is what you are really optimising for direct battery charging, efficiency has to suffer to get a match to the blades. If you load conventionally with resistors or with a mppt inverter then efficiency can be a factor and it changes a lot of the requirements.

Flux

[electrondady1]

if the magnet is round, then the centre hole would be round, and so the coil will also be round.
if the mag is rectangular as you have shown in your illustration then the centre should be the same shape .
 it will become more rounded as you add turns.
 your coil jig will dictate the maximum size the coil can be.
in an ideal circumstance all magnets would be wedge shaped and the legs of the coil would be parallel to a radius
in the case of a 9 coil stator the coils would form a segment 360/9=40 degrees

i kind of posted before i realized flux had answered he is the authority

[DanB]

Again I'll back up what flux has to say there...
big holes in the coils don't make lots of sense, there is room for copper in there and I'd use it.
I used to have the notion that it really mattered that things should be such that you would have a N pole over 1 leg and a S pole over the other leg.  But really it boils down to coupling the  flux through the center of the coil, I dont worry about the 'shape' of the coil so much.  Trick is fit a lot of copper in there, and again - as flux has said, the hole in the center of the coil can be smaller than the magnets, and the legs can be quite wide.  The windings in the center of the coil that maybe smaller than the magnets produce less voltage than the windings that are the size of the magnets or larger, however their resistance is very low and they still contribute to the output so I would have them (within reason!).  I am not certain about what 'within reason' is - it could be that I would benefit by even be making the hole in the center of a coil even smaller than I currently do.

[electrondady1]

you didn't' mention whether your vertical was lift based or drag based and what sort of load you were planing on putting on it.
i  found in working with drag mills and their lower  rpm it was expedient to build with fewer phases and more poles.

[michal]

Thank you all for the explanation,

My VAWT is a Lenz2 type (2m diameter, 2 m high wing)  and I want to use it for battery charging.

Now, that I have more idea about the shape of coils, I'll  play with different configurations.

Once again thank you all for your replies !!!!!

Regards

Michal

[Ungrounded Lightning Rod]

Rules of thumb:

Hole the size of the magnet:
 - Smaller hole and the opposite sides cancel out when the magnet is over both, so the extra magnet size / copper in the middle of the coil is a waste.
 - Larger hole and (as DanB said) you could have put in more copper and gotten more power.

For three phase:
    - Width of coil is half the width of the magnet.  (So the sides of two coils are over the magnet at max voltage.)
   - Space between magnets is half the width of the magnet, i.e. width of the coil side.  (The third phase will be between magnets and generating nothing when the first two phases combined are at max.  Flux on opposed-magnets is pretty much straight across so you don't need to bother compensating for spreading.)

If you can't get wedge shaped magnets, or rectangles are cheaper, get rectangles that are right at the outer radius and too wide at the inner.  (The extra material outside the wedge/hole profile is just wasted by cancelation.   All it does is increase the force between the plates during assembly.  But the real cost is money, not magnet material, so optimize for that.

The axial flux design is VERY forgiving, which makes it great for homebrew.  Get something "wrong" (off from some ideal) by 25% and all you do is lower your power output or efficiency by something in that ballpark, not cause everything to quit working or self-destruct.  With "free fuel" from the wind there is no need to be optimized to catch all you theoretically could.  Optimize instead for power/(money and manhours) and don't waste effort fine-tuning the design.  (Why work harder for the last 5% when you can just go 10% wider and get 21% more output from an equivalently "crummy" design?)

[lifer]

I have a similar question regarding the coil shape. I'm gonna use rectangular magnets and I know that the rule of thumb is to have the inner hole of the coils the same shape/size as the magnets. Anyway, I think that a better approach is to have an almost triangular shape (with rounded corners for easy winding) like in the image bellow:



(the original image belongs to the thread initiator)

This way, you can have much more (usable) coil windings swept by the magnets (so more output power). When a single magnet is overlapping a coil (it's right above the center of the coil) it doesn't matter how many windings are (simetrically) under that magnet because of the cancellation effect.

Another advantage of a triangular shape is the use of a smaller amount of (unusable) copper - as the upper and lower legs of the coil don't generate any voltage.

Is that correct?   

[electrondady1]

it's difficult to read the modified drawing .
it looks as though the coil legs would be over two mags for less time.

[lifer]

The outer shape of the coil it's the original one (as in the picture). I've just pointed out the new inner hole.

The two magnets will pass over a great number of windings (almost full magnet/coil leg overlapping).

[electrondady1]

what most people do  is build their rotors and then do some test coils .

[SparWeb]

Michal,
The outline of the coils as shown in your first picture looks right.  The field lines pass through the loop to produce flux.  Your second graphic has no complete flux path, and field lines in the air.  The lines drawn in red over that field will capture very little flux, if I assume the lines represent edge-on coils.  Your second graphic has a complete flux path well shown, but there is no representation of the coil between the magnet poles.  If you want to "capture" the flux travelling from pole to pole, the coil has to be big enough to wrap around all of the lines.
Perhaps reading this will help:

http://www.sparweb.ca/Forum/AXIAL_FLUX_HowItWorks.pdf

Lifer,
The coils can be any size that fits, and as long as the inner width is equal to or greater than the magnet width, then it will "link" all of the flux between magnets.  Making the coils bigger will gain you nothing more.  Reducing the size will close off some area, losing some of the flux from the magnets.

[lifer]

I've made a lot of iterations using this software and there's no benefits of extra-large coils indeed. Still I have one more question: the coils have to be wedge shaped (the two adjacent coil legs have to be side by side) or rectangular shape (I'm using rectangular magnets)?

From the software calculations there seems not to be any difference regarding output power but I've read that it's a right design to let each magnet pass over both (adjacent) legs simultaneously.

[SparWeb]

Lifer,
The wedge shape comes only from the space available to pack them in.  If there is plenty of space between your magnets, then there is plenty of room for rectangular coils.  Of course, if there IS too much room between magnets, then maybe the disk is bigger than it needs to be, and tightening up the space between magnets (and consequently making the coils tighter and then wedge-shaped), you will end up with a more compact, lighter, and cheaper axial-flux generator.  Trade-offs.

The "legs" aren't as important as you think.  The flux is "linked" when the field passes through the loops of each coil.  Once the most flux is passing through the most of the coil area, that is the moment where the flux is a peak.  Then the magnets continue moving along until they pass the next magnet pole.  Since the field is reversed, the peak flux at this point is the same, but in reverse.

Between those two moments, the flux has gone from a peak at one polarity, to a peak at the other polarity.  The difference between the two, divided by the time it took for the magnet to go from one to another, is the definition of the open-circuit voltage the generator will deliver.  If you speed it up, the time between magnet passings will be shorter, making the voltage higher.  For more detail on the formula, if you're interested, you can look up Maxwell's laws in a physics textbook or on a website like Hyperphysics.

There are some counter-intuitive things about the relationship between flux and voltage, so I won't confuse the issue, unless the details are of interest to you.

[lifer]

Very informative, thank you very much!

The coil leg width it's constrained by the inner space between coils but you can go rectangular or wedge shaped to the outer end of the coil. So, should I go rectangle (following magnet shape) or should I go wedge (equally spacing all the coils legs across the stator (radially) instead of leg-by-leg)?

(sorry for being too insistent but I'm planning to start building the rotor this weekend and I want to be sure that I'm doing it in the right way. The space between magnets is one mag width)

[Flux]

Make the inside of the coil at the outer radius the same as the magnet width. On the inside you can reduce it below magnet width perhaps 2/3 magnet width. Then wind the coils until they touch on the outside.

I wouldn't make any attempt to keep the legs radial with rectangular magnets, you will end up with a coil with hole wider than the magnet at the outer and all it will do is add resistance, which you don't want.

You are not optimising open circuit voltage, you are trying to find the best compromise between volts and circuit resistance for direct battery charging and resistance is a killer.

If you are using mppt and can match the voltage at each speed to the optimum load then things may be a little different as the effect of resistance is not so great.

Flux

[lifer]

Yes, I'm going to use a MPPT charger and the PM generator will also run at higher RPMs (~600) to get higher voltages. Thank you very much, Flux, now it's all clear!

PS: A little off-topic question: how thick the rotor iron back plate should be? I'm using (12 poles) 15mm thick N40 neo magnets.

[Flux]

1/2" thick minimum for those magnets.

Flux

[lifer]

A brake disk is a good option (being maybe thinner but double sided) or it has to be "pure" iron? (I guess the brake disk is made of cast iron)

[ChrisOlson]

For three phase:
    - Width of coil is half the width of the magnet.  (So the sides of two coils are over the magnet at max voltage.)

I don't know that I necessarily agree with that.  When you put those magnets on the mild steel rotor it turns the whole rotor into a disc with alternating magnetic poles and the shape or size of the magnet has little to do with how the flux is linked across the air gap.  I built a couple ferrite magnet generators with ferrite ring magnets that I got that are one solid ring with 16 poles in the ring.  No shape at all to each pole where you would have a pole with bar mags, etc..  It worked perfectly fine with the big ring mags.

I've also built very compact 20 pole generators for my 150V turbines with only 331mm rotors and 2 x 1 x 1/2 N42's.  That design goes against all the "recommendations" in homebrew, and those generators produce 3+ kW @ 450 rpm, continuous without overheating.



This is the stator for one of those units:

[joestue]

Chris have you tried pushing those magnets closer together?

so it finally hit me why there's an optimum magnet gap:
Pushing the magnets closer together increases leakage flux.
But the coil is smaller and has a lower resistance.
At some point there is an optimum, and this optimum magnet coverage ratio is probably going to be dependent somewhat on the magnet thickness.
inner diameter to outer diameter ratio is also a minor variable.

most permanent magnet machines operate around 150 degrees magnet coverage, which corresponds to about a 3/16th gap for a 1 inch wide magnet.

The same optimization goes for how thick the coils need to be. less flux but more copper..
but stiffness follows flux squared.

[ChrisOlson]

Basically, what I have found is that you can get away with just about anything.  You can build a uni-polar generator if you want with all north poles in it and it works fine.  I've built them with magnets only on one rotor and just a steel disc on the other side of the stator, using the steel disc on the other side for a "core" and it works fine and gets you more voltage than just spinning a magnet rotor on one side of the stator.  I've built them with solid ring magnets, as I mentioned, and that works fine.

Basically, I've gone to putting as many magnets as will fit on a certain size disc, assemble the whole thing up and spin a test coil in it.  If I don't get the voltage I want at the rpm I want it at I either evaluate if I have to remove or add turns, or put stronger or weaker magnets on.

[joestue]

these minor variations are on the order of 10%

how i suggest testing these alternators is shorting the test coil out and wrapping a rope around the shaft with a weight on it.
measure the rpm.
rinse, repeat, with larger or smaller test coils (compensating for the weight of the coil)
adjust the air gap, etc.

smallest rpm per pound of magnet and copper = most efficient alternator.

but this doesn't account for inductance effects at higher rpms

[ChrisOlson]

how i suggest testing these alternators is shorting the test coil out and wrapping a rope around the shaft with a weight on it.

Well, I never heard of that test before.  I just put my magnets on there and spin it at the rpm's the rotor runs at at peak power at 6 TSR and wind the coils to get 160 open volts for the Classic controller.  The more poles you use the better it is for MPPT turbines to make it easier on the input caps in the controller.  Or you can run less poles with gearing to get the freq up to make it easier on the input caps.  Done that too.  Either way works.

The efficiency is generally above 85% at full power when you finally give up on direct battery charging and go to MPPT and get the voltage up to decent levels.  Getting 90 amps to the battery on a 12V system with a 8 ft diameter turbine is nothing with MPPT.

[ChrisOlson]

The only thing that really requires any serious accuracy in building these axial generators is if you intend to run a delta configured winding.  Then you have to make sure to have the phase angles perfectly matched because you're going to be exciting phases in parallel.  If have the phase angles off (due to improper coil spacing, etc.) you'll get the DCC (Dreaded Circulating Currents) in your winding.

[analog man]

The efficiency is generally above 85% at full power when you finally give up on direct battery charging and go to MPPT and get the voltage up to decent levels.  Getting 90 amps to the battery on a 12V system with a 8 ft diameter turbine is nothing with MPPT.

Chris, I dont get what your saying, and I have read other posts of yours on it...this mppt thing..Admitedly I have never used it but unless I have missed something it dosnt sound great to me, why is 85% better than 100%(or close) because thats what I and I suspect others get with direct batt charging, and "finally give up on it" what are you saying? direct batt charging cant work? ....There are lots of people doing fine with it, I am one with no complaints, Getting 85 amps into my batts on a 12 volt system is nothing with direct charging either with the wind cooperating ofcourse. I dont want to knock the op's thread off, if you could open another thread to respond that might be better. Also what has the life of these controlers been, whats the failure rate used in a wind application, I have not had a failure directly charging batts in the 7 years I have been doing it. I thought I read that you have been through a few controlers yourself. (or maybe it was someone else). I have cash in hand, and am always looking for better ways to do things...make the sale...

[ChrisOlson]

Actually, you are lucky if your generator and transmission line efficiency is 50% on a 12V system without MPPT.  The MPPT greatly enhances the efficiency of the generator and reduces transmission losses in the line and rectifier.  For instance, take a turbine running at 800 watts @ 15V at the rectifier and 1.4V forward drop in the rectifier with a line resistance of 0.2 ohm and stator resistance of .25 ohm.   On direct hooked you are dissipating 700 watts of heat in the stator alone, another 560 watts in transmission losses to the rectifier, and the rectifier itself is dissipating 75 watts in heat.  The turbine is really producing 1,575 watts at the shaft and your overall efficiency is only 50.8%

Put that same turbine on MPPT operating at 100 volts.  Now your stator only dissipates 56 watts, your line transmission losses are only 45 watts, and the rectifier only dissipates 21 watts.  Instead of only 800 watts to the battery that you get with direct hooked with a sizzling hot stator -  you are now getting 1,450 watts to your battery, your overall electrical efficiency is 92.1%, and you have not even close to reached the amp limit of the stator yet.

It is all basic ohm's law.  You will never match the energy production of a MPPT turbine with a direct hooked across all wind speeds.  The MPPT turbine will beat it by 2x in energy production and 3x in peak power production, and never stress the electrical components in the system at all.

The Classic controller for wind has proven dead reliable.  I'm sure there has been isolated problems as there is with any product.  But I will guarantee you beyond a doubt that a Classic on a MPPT turbine is easily 5x more reliable than a sizzling hot stator that's going to burn out on a direct hooked one.

[ChrisOlson]

Put that same turbine on MPPT operating at 100 volts.  Now your stator only dissipates 56 watts, your line transmission losses are only 45 watts, and the rectifier only dissipates 21 watts.  Instead of only 800 watts to the battery that you get with direct hooked with a sizzling hot stator -  you are now getting 1,450 watts to your battery, your overall electrical efficiency is 92.1%, and you have not even close to reached the amp limit of the stator yet.

Frickin' forum software doesn't let you edit a post.

I am adding a correction to the above for conversion losses in the controller, introduced into a MPPT system.

The controller will add 6.5% conversion loss (measured with my equipment) at the amp outputs being discussed on a 12V system.

So instead of the 1,450 watts to the battery I mentioned above, that is input power to the controller.  The controller will dissipate 94 watts in heat, giving a final to the battery of 1,356 watts.  Overall electrical efficiency is 86.1%.

[analog man]

Chris, no I dont transmit DC from the tower nor would I suggest anyone too. Its straight 3 phase from the tower 55-60 feet to the house, and the rectifier, controler is 2 maybe 3 feet from the batts...line losses insignificant, to me. Rectifier losses I would agree as a proportion of the voltage, but if its  a concern that can be brought down to a third of the rectifier losses you state just by using diodes with a low drop, The output diodes on any old junked inverter welders are of that type and can be had for nothing or next to it at scrap metal yard where most these things go to die, .166 to .266 range is what I have found from junked xmt 304's, 456's and such type, or ofcourse can be ordered new, I just prefer cheap or free when possible.... efficiency can be looked at many ways, and I suspect you could beat me over the head with numbers from every direction showing me black is white, then prove white is black....To me if my mill is producing a charging voltage are the amps it can make going into my batts and if the answer is yes thats good efficiency for me, I do not one bit care about the max output of the mill or what its efficiency is there, as first my batts will quickly be charged up anyway regardless of efficiency at the high end, and hard driving wind are less common(steady) here, I am happy getting 5 to 15 amps  steady through a day, sure that Mppt can shine at the high end but I think I would be moving if I had to live where a 8 foot mill was usually puting out 90 amps.
I will lookover your numbers more and try to digest them, but I do math on my fingers so it may take a while. 

[lifer]

You're absolutely right, CrisOlson! That's why I choose to go higher with alternator's output voltage (>300V), running it at higher RPMs (600) through a transmission system.

Actually, I'm in the process of building a MPPT controller by myself  - with separate input for a backup PV panels system.

[analog man]

Looking again at your numbers, yes at the high end it makes mppt look good, a bad sign when you have to use a extrem condition to justify it..... so next week sometime when a 4 hour gust keeps my mill at max output I should be very happy if I had mppt, right now and until then I will try to get by while it bumps along doing its usual 5 to 15 amp dance.