Best coil geometry for round magnets

[dciolek]

I have purchased some 2" diam x 1/2" thick round magnets for a 12" rotor design that will use 9 coils.  What is the best coil geometry to use for this type of magnet/rotor design?  Some sources say that a "hole" in the middle of the coil should be approximately the magnet size itself.  Some say that the distance between the centers of the coil legs should equal the distance between the centers of the magnets.  Some say that the leading edge of a north pole should be striking one side of the coil at the same time the leading edge of a south pole is striking the other side of the same coil.  Then there is the thought that you want the coil to be perpendicular to the travel of the magnet.

All these don't seem to drive towards one ideal design.  And most of the coils you see in same designs (such as in Homebrew Wind Power) don't seem to care about meeting certain of the elements above.

Is there a difference between the best "practical" coil design and what fits all the theoretical best situations?  For example, it "seems" like a coil shaped like a trapezoid with the outer and inner legs completely outside of the magnet path and a coil leg width that just fits between the magnets and a coil leg separation between centers equal to the magnet separation between centers meets many of the above.  But it has an extremely long run of wire (more resistance) that doesn't intersect the magnet path at all (not producing anything -- additive current or negating current). 

Is there a practical rule of thumb that matches the coil dimensions and geometry to the magnet (not just # of turns and size of wire)?

[Flux]

I doubt that there has been enough done on this to form any real conclusions.

There is unlikely to be a general answer anyway. What may make the most efficient design may not make the best wind turbine generator, invariably the alternator that best matches the load will give best results.

For maximum voltage you need to link all the flux and this implies a hole as big as the magnet, in reality with the leakage flux associated with round magnets you are probably going to do better if you squeeze the coil into an ellipse, get shorter turns and lower resistance. resistance mainly is the factor that dictates how much material you need for a given output.

Trapezoidal coils will also work ok but I really don't see the need to worry about this cutting at right angles, if the hole is big enough you link all the flux, at best you only change the waveform and this is little importance.

I think trapezoidal coils will have more resistance than elliptical ones. if you are using a star connection I really wouldn't worry, magnet spacing and getting in sufficient copper to make the outsides just touch is the main thing.

What Dan does seems fine to me, but I would probably increase disc size a bit to get more space between magnets if you want value for material and size is not the main factor.

I am sure there will be others with very different opinions so you will probably be no wiser at the end of this question.

Flux

[Airstream]

Resistance is futile!

I've heard DanB explain coil sides a little more narrow than the magnet has most cancellation occurring below/at/near diode cut-out or at low enough voltages there is little or no subtraction from the peak power period, especially in a battery charging application. Made sense to me. .

[tecker]

The best coil form is long straight legs .The hole doesn't matter that much the thing that much but the hole has to be large enough to allow both magnet to work on the same coil.

[willib]

I have purchased some 2" diam x 1/2" thick round magnets for a 12" rotor design that will use 9 coils.  What is the best coil geometry to use for this type of magnet/rotor design?  Some sources say that a "hole" in the middle of the coil should be approximately the magnet size itself.  Some say that the distance between the centers of the coil legs should equal the distance between the centers of the magnets.  Some say that the leading edge of a north pole should be striking one side of the coil at the same time the leading edge of a south pole is striking the other side of the same coil.  Then there is the thought that you want the coil to be perpendicular to the travel of the magnet.

All these don't seem to drive towards one ideal design.  And most of the coils you see in same designs (such as in Homebrew Wind Power) don't seem to care about meeting certain of the elements above.

Is there a difference between the best "practical" coil design and what fits all the theoretical best situations?  For example, it "seems" like a coil shaped like a trapezoid with the outer and inner legs completely outside of the magnet path and a coil leg width that just fits between the magnets and a coil leg separation between centers equal to the magnet separation between centers meets many of the above.  But it has an extremely long run of wire (more resistance) that doesn't intersect the magnet path at all (not producing anything -- additive current or negating current). 

Is there a practical rule of thumb that matches the coil dimensions and geometry to the magnet (not just # of turns and size of wire)?

First off , a 9 coil 12 pole machine with 2" diameter magnets should have a 14" diameter magnet rotor. Not 12.

And on a 14 inch diameter magnet rotor with 2" diameter mags the distance between magnets determines your coil hole diameter , a little over an inch .

Your coil outer diameter is twice the magnet diameter.

[dciolek]

So rule of thumb number 1 = hole size must be big enough to allow N and S pole magnets to work on the same coil at same time (and in the same current direction we should add -- so we don't get negating current).

I can't create rule of thumb number 2 from the other post without a bit more math -- but it seems to suggest that you size your rotor to have enough travel distance for a coil size that is twice the magnet diameter.  I'm guessing (to get a general rule of thumb for when you say 9 coil, 12 pole = 14") you suggest 14" rotor because if I did this on a 12" rotor -- the 4" outer diameter coils overlap two like poled magnets and create some canceling current.

Now -- let's try the "rule of thumb" #3 for how to get to the little over an inch inner hole for the coil from there?  Is it that you want the inner portion of the coil on one coil leg to LEAVE the magnet space on the N pole magnet just as the outer portion of that same leg reaches the outer edge of the S pole magnet?  I can see that making sense.  But if I design the coil size using these parameters -- it has two different legs of the same coil overlapping a single 2" diameter magnet at the same time -- creating negating current.  Maybe that overlap is small enough that you make up for it with the power produced by the thicker coil leg during the high output portion of the sine wave?  This is where it got fuzzy for me...

But I'll continue on and ask if we can get rule of thumb number 4 to choose the wire diameter size such that the number of turns you want winds the coil to that exact leg and hole width (regardless of geometry)?  Now that's some interesting math...

[wooferhound]

This may help you a bit . . .
http://fieldlines.com/board/index.php/topic,143575.0.html

[dciolek]

Thanks!  After reading them all I come up with rule of thumb #5:

Quit asking questions about theory and wind some coils to test!  Adjust from there...

This makes sense -- because I kind of have some fixed variables to work with.

1) My (24) 2" diam x 1/2" thick round magnets are already purchased (not cheap)
2) My 12" rotors are free off my car (cheap and available - maybe not optimal diameter for these sized magnets based on what I am hearing, but better than making new ones at this point)
3) My 11 lb. spool of #16 AWG wire is already purchased (not cheap).

So -- I have plenty of wire to wind one coil as a circle, and one as a trapezoid and check out the difference.  Sounds fun anyway...

[willib]

2) My 12" rotors are free off my car (cheap and available - maybe not optimal diameter for these sized magnets based on what I am hearing, but better than making new ones at this point)

Not really..
Because NONE of what i described to you will work, without the correct size magnet rotors.

That is the starting point.

[dzen1]

best coil geometry


editors note: watch the photo size, we'd like to see 640x480 & less than 150k uploaded, I'm leaving them since its your 1st post...

[fabricator]

OK dzen, that requires elaboration, it is a nice piece of work and it looks like you are mass producing them, they look like maybe 220V AC stators?

[tecker]

The last part of the equation is over all thickness .The stator as whole should be only slightly thicker than the magnets.

[dciolek]

I built templates for a quick and dirty coil winder for three different coils.  Circular (slightly smaller than magnet diameter), trapezoidal (circumscribing the magnet sized hole) and then a modified trapezoid that is 90 degrees to the magnet travel path for about 2/3 the diameter of the magnet and then "cuts the corner" a bit to shorten the coil length and reduce resistance.

My magnets and wire were backordered, so I decided to go ahead and have the rotors machined and paint them (attached).  Then after 3 coats of primer and 3 coats of paint -- they look good -- but I started getting worried.  Should I have attached the magnets to the rotor first?  Adhesive to paint doesn't seem like as strong a bond and the paint could pull loose...

Although my intention was to pin them anyway -- the magnets have 3/16" diameter holes in them so I could screw them down to the rotor.  What happens if I don't use stainless steel screws?  Will it muck up the flux too bad?  I also thought about just using all thread with a slot cut in the head so there was no profile above the magnet but keep the magnet in place while in rotation.

[Flux]

Stainless screws are preferred but I don't think you will see any measurable loss in flux if you use mild steel ones, such big neo magnets will saturate an tiny screw with no obvious ill effect.

keep us posted on your coil experiments, I really am not convinced that there are any special rules that you have to follow, I don't believe there is a perfect geometry.

Flux

[DanB]

There are a lot of compromises and with this sort of alternator, I don't believe there is any 'perfect' or best.  For the 10' turbines I build these days, I am inclined to start using larger rotors just so that I can fit somewhat heavier wire in there.

It is all really very forgiving I find.  If the magnets are so close to one another that there's lots of leakage from pole to pole then you won't get the best bang for the buck from magnets... although, at the end of the day the only concern is that the alternator have a power curve which is a reasonable match to the blades.

I think there are a couple things going on in these alternators.  Lots of folks are concerned that the legs of the coils be just exactly perpendicular to the path of the magnets, and that the geometry be such that if there is a N pole over one leg there should be a S pole over the other leg.  I believe those things are an issue but a small one.

I like slightly wedge shaped coils because it allows me to fit a bit more copper in, that's all.

For the most part, it's all about flux coupled through the coils - for all practical purposes you can view this like a transformer.  In that case, the other stuff doesnt mattter.  Trick is to get enough copper in there of heavy enough gauge so get the power curve you need.  All these 'rules' people are throwing out are questionable -  they may serve as good guidelines but with this sort of alternator there is no 'perfect'.

[ChrisOlson]

I think there are a couple things going on in these alternators.  Lots of folks are concerned that the legs of the coils be just exactly perpendicular to the path of the magnets, and that the geometry be such that if there is a N pole over one leg there should be a S pole over the other leg.  I believe those things are an issue but a small one.

I've built a lot of generators and found you can get away with just about anything and they still work.  But if you want one that runs smooth as silk without a bunch of parasitic harmonics you need to pay attention to every last detail.  And that includes wiping the respective coil legs at precisely the same time, paying attention to perfect coil orientation, stacking the wires in the coils so every one is wound identical, and choosing a magnet/coil shape that optimizes the space available.  A few inches of wire either way with different coil shapes in not even a factor, even in 48 volt stators.

This configuration dciolek is talking about with 12" rotors and 2" round mags will work and make juice.  How efficient it will be is another matter.  If you don't mind burning up some rotor torque in heat and parasitic losses right in the generator you can do anything and still get juice down the tower and a proper match to the blades.  If you want to convert more of that rotor torque to juice then you need a different configuration and you'll still get a proper match to the blades.

To pack the highest copper density in the stator I prefer wedge mags - the reason being is because you're laying things out in segments of a circle and that's what wedge mags are.  When I feel like compromising I use bar mags.  Round mags are my last choice and in this case require generator rotors with at least 14" diameter to get it even close to right.  The only reason round mags work is because they got sheer brute area allowing you to use less turns in the coils.  But get them too close together on 12" rotors and you'll be exciting enough coils at the same time - in different phases - to have no problem melting ice off the stator in the winter time, assuming the rotor can make enough torque to get it turning.

I guess it all depends on what you want in the end.
--
Chris

[dciolek]

Considering I've never dabbled in putting something like this together before -- what I want in the end is to understand how these machines get put together -- most efficiently and effectively.  I'm not necessarily shooting for perfection out of the gate, but more like a very solid understanding of what it takes to assemble them, to actually generate some power and get a feel for what you get out of it based on what you put into it (sweat and $$$).  If after this exercise I understand the different types of choices you can make (magnet size and shape, rotor sizing, blade profiles, blade matching, coil geometry, wire sizing, how to feed your batteries and how to extract what you put in there most efficiently, etc) and how they impact your available power in the end -- then I'll feel like I hit the ball out of the park.  Now that's going to be a lot of work!

I must say -- I am immensely grateful to all of you posting on this board.  You are an incredible resource and so helpful beyond just reading a text straight through and doing what it says.  That's because when I come across something that gets my brain cells fired up, I want to know why and how it works (or works better or works worse than some other way).  A quick post here on that topic brings back a wealth of knowledge and experience I couldn't find in my own little part of the world.  Pretty cool.

The way I see it -- I'll burn through several hundred bucks buying too big of magnets for too small of rotors along with a large coil of what will probably end up being the wrong wire size.  But I chose to do that instead of building straight from the recipie (even though I bought DanB's and Dan Fink's book and read it cover to cover to gain knowledge).  I'll learn some things the hard way I expect, and I'll tweak and learn along the way so that when I can actually fly one of these things -- I'll build it better.  And the kind of straight talk offered by ChrisOlson about the difference between making juice and making the best juice you can make from the resources you have available is VERY helpful in that process.  But honestly, so is the assurance that if you aren't expecting perfection, you can still make some electrons fly.  Honestly, for right now -- this is just tinkerin' for me and a GREAT excuse to run the power tools in the garage for a while...

[ChrisOlson]

But honestly, so is the assurance that if you aren't expecting perfection, you can still make some electrons fly.  Honestly, for right now -- this is just tinkerin' for me and a GREAT excuse to run the power tools in the garage for a while...

What I would do, if I were you, is screw those mags to the rotors, and don't bother glueing them, so you can take them off easily later.  I like to experiment as much as anybody and have done many things like you're about to do that other people have told me don't work.

For a 12 volt turbine with small coils and small wire you might be able to come up with a configuration that will work reasonably well with those small rotors.  But leave yourself an "out" and don't waste your mags if you decide to tear it apart later and rebuild it.  Those mags can make a pretty powerful generator in larger diameter units, and they're not cheap.  So that's what I'd do.
--
Chris

[willib]

What I would do, if I were you, is screw those mags to the rotors, and don't bother glueing them, so you can take them off easily later.  I like to experiment as much as anybody and have done many things like you're about to do that other people have told me don't work.

For a 12 volt turbine with small coils and small wire you might be able to come up with a configuration that will work reasonably well with those small rotors.  But leave yourself an "out" and don't waste your mags if you decide to tear it apart later and rebuild it.  Those mags can make a pretty powerful generator in larger diameter units, and they're not cheap.  So that's what I'd do.
--
Chris

I agree with chris , i wouldn't glue them down either.
The threaded rod idea will work, also if possible a bolt from the other side would work better, But the casting will probably get in the way .
So the threaded rod with a slot in it is probably best.

[DanB]

This configuration dciolek is talking about with 12" rotors and 2" round mags will work and make juice.  How efficient it will be is another matter.  If you don't mind burning up some rotor torque in heat and parasitic losses right in the generator you can do anything and still get juice down the tower and a proper match to the blades.  If you want to convert more of that rotor torque to juice then you need a different configuration and you'll still get a proper match to the blades.

I think you have a misconception here Chris.  With that arrangement, 12" rotors, 12 2" round magnets, the efficiency is pretty much related to the resistance of the copper and nothing else.  There are absolutely no parasitic losses.  I find that I can use the same stator exactly that I would use with 1" x 2" blocks, with the larger disk magnets and the power output (or open voltage) at any rpm is pretty much just related to the increased surface area of the magnet.  Slightly less, for two reasons I think - 1, there is slightly leakage from pole to pole because the edges of the disk magnets are a bit close to one another, and 2 - the inner windings of the coils are too small to couple all of the flux from magnets this size.  But it is surprisingly close.  That said, I totally agree that larger rotors with those magnets will get you more power - at the cost of more steel, and more copper.  (if you don't use more copper though t he increase will be very small) 

My main point here though is you are telling folks that parasitic currents exist in this situation, it is not the case.

To pack the highest copper density in the stator I prefer wedge mags - the reason being is because you're laying things out in segments of a circle and that's what wedge mags are.  When I feel like compromising I use bar mags.  Round mags are my last choice and in this case require generator rotors with at least 14" diameter to get it even close to right. 

I obviously have to agree, wedge shaped magnets are the best choice if your concern is about squeezing the most power from a machine of given diameter. 

The only reason round mags work is because they got sheer brute area allowing you to use less turns in the coils.  But get them too close together on 12" rotors and you'll be exciting enough coils at the same time - in different phases - to have no problem melting ice off the stator in the winter time, assuming the rotor can make enough torque to get it turning.

It does make me wonder how many different configurations of alternator you actually have built - you seem to think that there is some drag / parasitic losses in this situation, and again, there isn't 

If you actually pack your magnets too close together, such that there are times when you actually have two poles over the same leg of a coil at the same time - it will make a slightly funny wave form, however there will be no drag and no parasitic currents within the coil.  (I've taken this to extremes in some cases) - it will not make the machine hard to start, it will not cause it to run hot.  In fact, if your concern is to get absolutely the most power and /or the highest efficiency from a machine of given diameter (at the cost of spending a bit more on magnets) then I should say pack the biggest magnets you can as close together as possible.   At some point it does get silly though.  When I pack 12 2" disk magnets onto a 12" diameter rotor - I am having some very slight flux leakage from pole to pole which is not ideal - but it is small and the resulting alternator is still too powerful for the 10' blades we usually fit to it.  The sine wave produced from that alternator is about as perfect as my eye can see and so long as they are wired in star... the only 'drag' on the machine is that from the bearings.

Someday... if you like, come by and turn one - and  you'll see!  And watch my listeroid run at 2kW sustained output.  Look at the sine wave and feel how cool the stator runs. 

I guess it all depends on what you want in the end.

It really is all a compromise between keeping things compact (and making the best use of steel/resin), making the best use of magnetic material, making the best use of your copper, keeping it fairly easy to manufacture and matching the alternator to the blades.  Alternator efficiency is also important - again though, in this case... the efficiency is still related to little more than the resistance of the copper, and parasitic currents do not exist.   

[ChrisOlson]

I think you have a misconception here Chris.  With that arrangement, 12" rotors, 12 2" round magnets, the efficiency is pretty much related to the resistance of the copper and nothing else.  There are absolutely no parasitic losses

Have you ever calculated the efficiency of these generators that you build that are "too powerful" for your 10 foot blades?  You might find, like I did, that it's not so much the generator being "too powerful" as it is the fact that the efficiency curve looks like a truck driving off the edge of a cliff.

So I'll leave it at this being I obviously don't understand axial flux generators - Ed Lenz has probably done more than anybody I know in figuring out how to build more efficient axial flux generators.  Ed has provided me with countless tips and ideas on how to go about this since I took an interest in it some time back.  YMMV but I really think that a lot of the homebrew designs out there work OK, but they're not as efficient as they could be.  And most of the time that fact is lost in the shining light of success when the design works and puts power down the tower - even though it could be putting more power down the tower.

I've spent the better part of the last year attempting to build new generators that run better than 50% efficient at rated speed and power and I've come a long ways from the first ones I built that simply "followed the published plans".  Sometime if you're up this way I'd be happy to show you my generator collection I got hanging on the wall here - just a rough guess probably 30 or so of them where I've tried different things- some burned up, some perfectly good but didn't work the way I wanted - but generators that only run 35% efficient at rated power don't impress me anymore.  I figure there's no sense to wasting perfectly good rotor torque and contributing to Global Warming by converting wind directly to heat   
--
Chris

[Flux]

I really don't think this is an issue of efficiency at all, its just back to load matching.

When you are experimenting with star delta switching then you are making a reasonable attempt at load matching. Efficiencies upwards of 70% are possible in the high wind region even with direct matching if you set the loading to suit the steep part of the prop curve.

If you design for cut in in the 7 to 8 mph region and have no matching system then all a high efficiency alternator will do is bog the prop down into stall. Unless you devise some method of letting the alternator rpm rise with wind speed then it is inevitable although prop characteristics can help a bit.

The difference in efficiency between the two schemes is nothing more than the normal case of the output of an alternator for a given efficiency being related to its speed.

There are basically only resistive and eddy losses and bearing loss in an axial alternator, eddy loss is tiny unless the wire is too thick.

If you double the cut in speed by using star delta or any other method the high wind capability goes up very drastically. Doubling the cut in speed lets you use half the turns of twice csa wire and that means 1/4 the resistance. This is where the improvement comes from, for a given output current your stator heating is suddenly reduced to 1/4 of the value of a machine designed for low cut in and bogged down.

With mppt loading I can get an overall electrical efficiency of over 70% from cut in to full load, it could be higher with a vastly oversized and costly alternator but it is not cost effective. You just can't achieve anything like this with direct connection from low wind cut in.

In low winds the alternator efficiency is in fact very high but not much above cut in you start loosing prop efficiency and although you don't hit hard stall the performance is in fact quite a bit down on a mppt scheme in the low wind region as soon as you are about 2 mph above cut in wind speed. In the high wind region there is no comparison.

My 6ft machine produces significantly over 1kw in decent winds but the input to the converter is over 80v on a 24v system If you try to bog the alternator down to 24v then there is no hope of any performance, in fact without the converter the alternator is too powerful and it stalls the prop at about 120W. It would in fact work far better without the converter if the alternator efficiency was made deliberately lower.

The alternator is a dual rotor using 8 magnets per rotor rectangular 46 x 30 x 10mm . The stator has 6 coils and is star wound, it is potted in polyester and I suspect the thing doesn't exceed 40 deg C at full load.

Load matching is the only thing that is ever going to change things, there is nothing wrong with alternators as long as you load them correctly.

Star delta change is indeed a step forward, I started there a long time ago, in some wind areas it may in fact be quite good but in other areas it spends most of its time in the wrong mode. I covered all of this a long time ago in the hope that it would cause people to think but most seem happy to live in stall mode and accept the loss of performance, I can see why with battery charging machines where you mostly spend time dumping on high wind days and it is ok if you make it furl within the capabilities if the alternator.  I really don't go along with this constant worry about burn out and blaming the axial machine, the thing doesn't have an inherent design fault, it just has limitations like everything else and you have to stay within the limits.

I admire the work that Chris has done and appreciate that he shares it with us but the improvements are largely in better matching rather than any fundamental things with the alternator. I am not familiar with the blades that he uses and most likely they need different alternator characteristics for the best match and this may be another factor.

I have to admit I have never taken much interest in blades, there is more to be gained from matching than clever blade technology for normal loading. Once you have got the loading right then there could be more to be had from better blade profiles.

I seem to manage an overall cp of 0.3 from cut in to near full load with the buck converter machine with basic hand carved wooden blades, a cp of possibly up to .4 may be possible with decent blades but it is only a guess, I just don't have the data.

Flux

[DanB]

Have you ever calculated the efficiency of these generators that you build that are "too powerful" for your 10 foot blades?  You might find, like I did, that it's not so much the generator being "too powerful" as it is the fact that the efficiency curve looks like a truck driving off the edge of a cliff.

I have as best I can and it tends to pretty much follow what I would predict based upon the resistance of the wire in the stator (slightly worse but close)

So I'll leave it at this being I obviously don't understand axial flux generators

I did not actually say that, but your comment stating that there would be drag, parasitic losses, and heat in the stator in the stator that specifically has 12 2" diameter magnets crowded rather closely together with 9 coils as we build them, is dead wrong.  Since you stated that as a fact, I thought I needed to say a bit aboutl that.

- Ed Lenz has probably done more than anybody I know in figuring out how to build more efficient axial flux generators.  Ed has provided me with countless tips and ideas on how to go about this since I took an interest in it some time back.  YMMV

Ed does great work and he's well ahead of most of us... he's still on a learning curve like the rest of us though.  What does YMMV mean?

but I really think that a lot of the homebrew designs out there work OK, but they're not as efficient as they could be.

That's for sure.  Again, we all have to find a compromise.  At the end of the day though, unless you have some sort of MPPT controller or whatever...  there *has* to be some inefficiency to keep the blades out of stall.  Some will be in the alternator, some will be downstream of that.

 

And most of the time that fact is lost in the shining light of success when the design works and puts power down the tower - even though it could be putting more power down the tower.

Yes I can't argue that.

I've spent the better part of the last year attempting to build new generators that run better than 50% efficient at rated speed and power and I've come a long ways from the first ones I built that simply "followed the published plans".  Sometime if you're up this way I'd be happy to show you my generator collection I got hanging on the wall here - just a rough guess probably 30 or so of them where I've tried different things- some burned up, some perfectly good but didn't work the way I wanted - but generators that only run 35% efficient at rated power don't impress me anymore. 

Honestly - generators that run much less that 60% efficient at 'rated power' (which means sustained output in my mind) burn up so...

I figure there's no sense to wasting perfectly good rotor torque and contributing to Global Warming by converting wind directly to heat   

well... at the end of the day all that wind must heat up one thing or another so I doubt we're contributing much to global warming!  No -= Chris - I only came in to contest some specific statements about parasitic currents and alternator geometry.  Both you, and WillyB seem to think there are some very specific 'rules' and that if they are not followed it's 'wrong' or 'inefficient'.  That is simply not the case.  Any alternator of reasonable geometry may be fine...  they are all inefficient above certain output levels.  Some are smaller and waste magnetic material a bit. Some are larger... have very little magnetic material and lots of steel and copper.  But within reason, there are lots of ways to skin the cat.  Specifically though, the comments about drag, parasitic currents, and heat are wrong.  I waste a bit (not much) of magnetic material in the way I do things with round magnets on 12" rotors I agree...  I do get a slightly more compact machine though and make use of stuff I already have on the shelf (those are my current compromises).  I just have a small problem when people start promoting their 'theories' or 'opinions' as facts.  It may be so but if I disagree I'll usually let that be known too.  

That said - for 10' turbines in the future, I'm inclined to start using smaller magnets again and larger rotors and more copper.  The 'rules' in this thread though that others have put fourth about alternator geometry are 'not necessarily so'... I'm fairly sure of that.  

[DanG]

YMMV "your mileage may vary..."

They used to say "see the big picture" at the railroad I worked at, its a joy to read this kind of thread...

Don't comparisons skew fast when we insert the variable air gap and site specific drop lead resistances, that calculated efficiency becomes just a subjective bench test without a set of standard test conditions?

[DanB]

Hi Dan - the efficiency of the wind turbine overall is affected by the airgap - and the line (drop lead as you put it).  The airgap has no real affect on alternator efficiency though, it only affects the speed.  The drop line yes...
but really it's just alternator geometry we're 'arguing' about here...

[ChrisOlson]

but really it's just alternator geometry we're 'arguing' about here...

This isn't an argument - it's a discussion.

You can build pseudo efficiency into an air gap axial flux by overwinding it, using bigger wire or more parallel strands - and this is the most common method used.  But it's not the most efficient.  You can select a target ouput and say, "I want >60% efficiency at 'x'" kW" - use the above method and get it.  But this does not make up for inherent inefficiencies in the design of the generator.  All you're doing is using a 10 lb maul to drive a 10 penny nail.

If manufacturers of commercial motors and generators did this they'd be broke because the other manufacturers that have R&D figure out how to build the same thing that performs just as good with less cost in materials.

When I have tested these generators I push them to the full rating of the wire in the stator and measure power in and power out.  The most efficient star-wired home built axial generator I have ever tested came in at 55.7% efficient @ 57.5 volts on a battery charging system.  The most efficient delta-wired generator I've built to date came in at 67.7%.  That difference from 100% efficiency is caused by parasitic losses in the generator, whether those parasites be wire resistance, phase timing off, or what have you.  You're converting shaft input power directly to heat - almost half of it in the best wye-configured axial flux generator I've ever tested, running it at the full rating of its wire.

Only when you know this can you start making changes to coil shapes, magnets and overall layout to arrive at a design that's better.  Throwing the 10 lb maul at it doesn't gain one thing other than getting the desired power output without burning it up - at the cost of extra materials, weight and size.

Getting back to these 12" gens with 2" rounds on them, it might be a good way to use up stuff you have laying around the shop.  And they DO work.  But no matter how you try to justify it and bend the rules of good air core axial flux design, they're not even close to being efficient compared to what you can do with those same mags.  And that's the bottom line here.  And that's why I recommended that they not be glued to those car brake rotors and cast in resin so they're easy to get off later and improve it. dciolek is experimenting and learning here - and my hat's off when I heard that because I LOVE to do that stuff too.  But please don't start him off by trying to convince him that what he's doing is "fine".  Even drawing it out on paper gives me the shivers looking at how coils are going to be excited in that thing with only ~.600" between magnets and the massive overlap on coil legs from a single generator pole.  Put wedge mags on those rotors and then you'd actually have something.

Again, and here's that term, YMMV.  But I've always believed that if there's a mouse trap it can be made smaller, faster, lighter and better - and still pin that mouse under its bail just as good the big, slow, crude, inefficient one.  And use less cheese to lure him in to-boot.
--
Chris

[dciolek]

Excellent feedback!  And I am definately going to pin these mags as suggested and not use adhesive, I'm glad I bought them with the holes already there.

Let me step down the path on "drawing it out" (which I have done extensively in CAD as I experiment).  The thing I find the most vulnerable on in my understanding is that I have yet to have the opportunity to actually wind a coil, so I have no clue how many turns I can fit in that .600 magnet spacing to avoid the "massive overlap" on coil legs from a single generator pole.  I understand that overlap is a bad thing, so the design on paper tries to avoid it.  But I realize paper ain't reality...

One of the "paper designs" I laid out from the little knowledge I have gathered has a hole in the center of the coil just slightly smaller than the magnet (whatever a 2" hole saw leaves in wood when you are done, about 1 7/8" I expect).  And in the CAD version of the design -- I am showing the coil leg thickness of each of the geometries (circle, trapezoidal and hybrid) to be such that as the leading edge of the coil hits a magnet pole -- the trailing edge is leaving that same pole from the magnet two behind it.

The little piece of information I don't have at this point is, "Can I actually wind a coil with enough turns to produce the voltage I want with a wire thick enough to deliver the amps I want at a resistance that is reasonably efficient?"  I'm experimenting to get that answer -- although others here may already know that answer.  I have been pursuing this without that answer so far because I figured I'd find out when I wound that first test coil and adapt from there.  And I realize that I have two directions to go in increasing the number of turns in my coil -- either overlap like poles by increasing the width outward or excite the same coil with current going in opposite directions -- both of them probably not deal killers in small amounts, but increasingly bad I expect as you do more of this.  Just not sure what the "acceptable" limits are.

If there is a good mathmatical approach to determining how many turns you can pack into a given coil space for a given wire size at a given coil/stator thickness -- and that method is reasonably accurate given "homebrew" coil construction conditions (I am not a carefully controlled manufacturing facility -- so my coil density may not be perfect), then I'd love to work on that math too.  Until then, I'm waiting for the wire to show up so I can give it a spin!

[DanB]

Yes this all sounds good and wise and I think you'll be doing well to pin your magnets down and experiment a bit.  I can't argue with Chris's last statement really.
Could you get more power from your magnets at any given rpm by going with larger diameter rotors - and the expense of a bit more copper and steel?  Yes, definitely. 
Will it cause parasitic drag if you're coil legs are a bit wider than the space between your magnets? - no.  You'll get more power and higher efficiency at any given rpm if you take that route considering the rotors and the magnets you currently have to work with.

Just one bit of data that might be of interest.  If I use 12" diameter rotors, and load those up with 12 magnets, 1" x 2" x 1/2" thick (N42 grade) - and wind my stator so that the holes in the coils are just slightly smaller than those magnets and such that the coils are wide enough so they just touch each other on the outsides (can't fit anymore copper in) - I get 'X' voltage at a certain rpm.

If I use the very same stator with 2" diameter round magnets (which have about exactly 1.5 x the surface area/total mass of the smaller rectangular magnets), then I get about 1.4 * X voltage from the same stator, at the same rpm.  And no drag below cut in - and in doing so I am violating all sorts of rules here.  The fact that I do at times have opposite poles over the same leg of a coil does not seem to affect things much, again - the main reason I expect I'm only getting about 1.4 instead of 1.5 - is simply that the inner windings of the coils are too small to couple all the flux with the larger magnets.

If I used 14" disks, and a fair bit more copper in the coils... with the same gauge of wire and more turns to fit hte coils tightly we could probably get 2*X voltage at the same rpm. 

As it is now though, if I shoot for a specific cut in speed, but using basically the same stator (same size, same shape of coil) and the 2" disk magnets vs the 1" x 2" blocks I can achieve the same cutin speed but cut alternator resistance in about half by using thicker wire and fewer turns. 

All this said, surely as Chris has pointed out, there is always room to optimize the design and make better use of resources.  We do that best by sharing one anothers experiences and opinions and I expect we'll have it 'perfect' in a year or two!  (just kidding)

[ChrisOlson]

Excellent feedback!  And I am definately going to pin these mags as suggested and not use adhesive, I'm glad I bought them with the holes already there.

That's a good move getting the mags with holes.

I made a quick basic sketch in AutoCAD of the layout of this thing to show the problems I have with this design from an efficiency standpoint.  It should be attached to this post.



The maximum width the coil legs can be to prevent cancellation of the coil by leg overlap is .725" and still get it to fit under 12" rotors.  But you can't do this because of the wide mags - you're going to have two poles striking one coil leg on part of the rotation.  So the compromise is to scrunch the hole in the coil.  Now you're going to have leg overlap from one pole.  So the magnet is wiping one leg and that coil (the whole phase) is at full power.  You have electrons flowing one direction in the coil and the magnet strikes the other leg, you have current trying to flow both directions in the coil and it cancels before the magnet is done with the leg it was working on.  The phase doesn't switch polarity and make any power the other direction until that magnet clears the leg it wasn't done with yet.

On a 12 volt you can "fudge" it and get away with it.  On a 24 volt it's worse.  A 48 volt is bad.  You can increase the thickness of the stator past 1/2" and run more air gap to make the coils thicker and the legs narrower.  But to get the desired cut-in you have to put more turns in the coils if you do that, which adds fuel to the fire.  Don't even consider building a delta wound generator with this layout.

Despite any of this, if you can't figure out why your stator bolts keep vibrating loose even though you use locktite, nylock nuts and peen the threads ....... well........ you're looking at the problem.  You can't just simply slam-bang cancel phases like this and have a smooth running or efficient generator.  These things are what I call "buzzers".  The other problem I found when I started testing these after I built my hydraulic motor setup is that they exhibit a characteristic I call "piling up" at the full rated capacity of the wire.  Some people call this "too powerful for the blades".  What it does is put the rotor in stall and not let it make all the power the blades can make.  If you run one of these on my test setup and push it all the way to what it SHOULD put out without overheat, based on the rated capacity of the wire - they burn right up.

So the common method used to keep these things from setting themselves on fire, to get the 1.15 Service Factor, is use bigger wire, or more turns in hand, so you only run the wire at ~75% rated capacity at the output that you need to match your blades.  Once you know what this is you can select your wire size.  When you know your wire size, you can do all the calculations you want on coil size and they won't turn out.  The best method I've found is to set up your winder and wind a test coil.  This will give you an idea of what the size of the coil is going to be, and you can run that coil in between the rotors and figure out how many turns you actually need --- and adjust your pins in your coil winder accordingly.

So this is what I'd call the ultimate compromise configuration.  It'll work and put power down the tower and you'll have fun with it.  And in your case building your first generator, this is probably a very good one to build to learn about how NOT to do it.  But it'll also give you a good idea of what you CAN do and still have it work   

I don't know what size turbine you're building but be aware that this design generator gets VERY hard to turn at higher outputs and rpm's.  The common method used to compensate is to open the air gap.  So it's another compromise.  If you get the cut-in you want but it stalls before it should, you open the gap.  Then it's going to cut in later and make more top end power at the expense of low wind performance.  There's people who will tell you there's no power in low winds anyway, so it's no big deal.  But this is a little generic I think because I've found there IS power in low winds.  My 13 footer made 18.1 amp-hours yesterday at only 5 mph average wind speed for the whole day.  When you live off-grid 18 amp hours is better than zero amp-hours from a turbine spinning on the tower all day below cut-in.
--
Chris

[DanB]

Chris - someday we must get together over a cup of coffee, we would have no end of stuff to talk about 

I see you're point, and I think I've understood that bit from the start.  I used to think it was very important to get the coil legs as you describe them.

At one point Ed Lenze was arguing the point that the legs of the coils should be exactly perpendicular to the path of the magnets.

My current thinking is that it's not really a problem if two poles slightly for a moment of time are over the same leg of a coil at the same time.

My current thinking is that almost everything that's going on here has very little to do with the magnets moving over the legs of the coils.  It's mostly all about flux changing through the center of the coil, like a transformer.  If that is the case, then possibly round magnets are the best choice, because a smaller length of wire can surround the most surface area.  Of course, in an ideal situation (to make the most of the magnetic material) it would be nice to use larger rotors and fit more copper, but that is beside the point right now.  My guess though is we are splitting hairs one way or the other.

I'm the alternator built from 12 of these disk magnets is too powerful for 10' blades because of this:



(I've posted this before so sorry to be redundant)

Assuming 10' blades, running at a TSR of 7, then the blue line is 'betz' limit (cP of about .6).  A more realistic place to put that blue line (which is power at the shaft) should be at about half the power (cP of about .3 for most blades)
The red line is output we get from an alternator that's built with 12" rotors and 1" x 2" x 1/2" blocks (which is basically in line with the geometry rules put fourth here by Chris and WilliB).   The yellow line is output from the alternator I'd build with 2" diameter disks on the same 12" rotors with about half the resistance.  And it's not just internal losses that make it hard for 10' blades to run, it's the fact that at about any rpm (especially those in higher wind speeds) - it produces a lot more power than the blades could possibly have to offer.  Internal losses only add to that problem and that's not shown on the graph above.

Point is though - the red line (1" x 2" block magnets) and the blue line (2" disk magnets) - the differences in power output are pretty much related to the difference in surface area of the magnets themselves, and the fact that I'm crowding magnets a bit doesn't really seem to be showing up as a problem.

I am quite sure though that I don't have an issue with 'cancelling currents'. 

[ChrisOlson]

At one point Ed Lenze was arguing the point that the legs of the coils should be exactly perpendicular to the path of the magnets.

I think what me and Ed (has) pointed out is that there's some efficiency advantages to packing the most power in the smallest package by using your materials better.  Take a Leece-Neville 200 amp brushless truck alternator for instance - that is an amazing piece of equipment being able to put out almost 3 kW at better than 70% efficiency in the small package it comes in.  The engineering in that thing is light years ahead of what we're dealing with in these homebrew turbine generators.  And it does that with ~1/10 the copper in it that a typical 3 kW homebrew turbine generator has in it.  And, unlike normal car alternators, those things are built to put out that kind of power continuous - on a 12 volt system.

There are some of these things that are done in commercial generators that can be applied to the lowly direct drive axial flux.  And those are the things I like to play with   
--
Chris

[DanB]

Yes  - well that sort of alternator has some real advantages when it comes to saving on copper (steel cores and very high rpm).

But I agree totally Chris, there is always room for some improvement.  I expect we get one anothers points here.

[ChrisOlson]

Point is though - the red line (1" x 2" block magnets) and the blue line (2" disk magnets) - the differences in power output are pretty much related to the difference in surface area of the magnets themselves, and the fact that I'm crowding magnets a bit doesn't really seem to be showing up as a problem.

I think there's something wrong with your data.  Unless I'm reading that graph wrong, it appears that Alternator 2 is so efficient it puts out more electrical power than what's available for shaft power to drive it up to ~1,400 watts.

I guess I've never seen one that efficient.

I re-opened this to edit more comments:
Unless this power curve was measured with a power-in, power-out measurement device like I use - in which case Alternator 2 is not even a good match for 10 foot blades.  I've never actually taken the time to plot a power curve for any of my generators at each incremental speed increase.  I should do that sometime because I think my newest delta ones would show a considerably steeper power curve than my old ones that I built that were hard-wired in wye.  I've been pretty impressed by their performance in higher wind speeds because the delta configuration takes more rpm's to generate the voltage required to push amps so it lets the turbine rotor "spin up" and operate at considerably higher rpm's when the wind picks up.
--
Chris