Alternator Output Waveform

[ghurd]

Beating a dead horse...

"If someone does not care to know which configurations will produce what waveform, they need only skip over this thread.  It is appropriately named, after all."

The waveform, while charging batteries with a polyphase machine, will be +1 with some choppy ripple.

The waveform of a single phase machine charging batteries will be... 0, +1, 0, +1... uh... Pulsating DC.
Pulsating.  To expand and contract rhythmically; beat.  Expand and contract with strong regular movements.
DC.  Unidirectional flow of electric charge.

If some shade-tree turbine builder wants to design a nonworking turbine radically different than the one Steven Fahey is explaining, by choosing a single paragraph to work from, nothing anyone can do about it.
He is the same guy who reads the tuneup manual for his V8 Mustang, but can't find 8 spark plugs on his 2 cylinder Onan diesel.
And he is the same guy whose best chance of understanding it will possibly with the term "Pulsating DC".

[SparWeb]

Forgive me, everyone, but I don't think we're helping much.
Forgive me, too, kamikaze, because I don't think you're going to learn what you want to know through the internet.  Technical knowledge is easy to discuss on a forum, but MUCH harder to teach.  I can see you are taking a lot of it in, but it's not fitting together in order.  That's bound to happen when various people with varying ideas come together to try to answer a question that reaches so far and wide.

The internet works really well for some things, but it is absolutely USELESS for other things, and understanding science is one of them.  You just gotta do the work with a book and a pencil plus a bunch of wire, magnets, batteries and whatnot on your desk to relate to it.  I did it in school myself, and even so I came out with a lot of mis-conceptions until I came back to the subject in this hobby years later.  I could even rattle off lots of memorized things about Ohm's law but couldn't use it properly in the real world until I made a lot of mistakes.
What I can see through your questions is that your goal is to understand electromagnetic theory in depth.  I don't think a forum will get you to that goal.

http://www.amazon.com/University-Physics-Harris-Benson/dp/0471006890

McGraw-Hill and Wiley publish a lot of good textbooks on all physics subjects so you shouldn't have any trouble finding one on Amazon (or try the used book stores like Alibris.com because they can be a lot cheaper).

These Shaum's outlines are great references though the explanation is a bit thin.    There are examples to work through.
http://www.mhprofessional.com/product.php?cat=145&isbn=0071632352&cat=145

And this is on the web, but it is only a reference, not a textbook or guide:
http://hyperphysics.phy-astr.gsu.edu/hbase/hframe.html

[oztules]

Not so Sparweb.

I think it was reading and interpreting books in isolation that got him in this in the first place. The net has been able to help undo the damage.

Different people have different aptitude with books. Some folks can construct their IKEA stuff when they get home by native savvy, others try reading the instructions and get it right.... others read the instructions and never get it built..... were not the same. You love books and figures.... I love books and am pretty cool on figures.

In our country, we all read the same news... and all draw different views.... so with  the same information disseminated to the population, we get a polarization of the population... so books dont say the same things to different people, and so are not to be relied on for all learning for everyone.

BUT 10 minutes with a scope and a magnet and a mess of wire will unveil it all, without a single mistake. The laws of the universe (ok at these current temperatures) do not change, and will always reflect this in experimental results.

Short of that, talking to those who have messed with the wire and magnets and scope, has got to be the next best thing.

I know You can get all the formula and set about finding all the variables and stitching them into  sensible mathematical model..... I'm scratching to count my toes.

Were all different.

This thread would have stayed very short, if the those  on the sidelines had pointed out straight away that the base assumptions were in error, or the frame of reference was wrong for the formula to be useful.

Maybe that says that more folks have learn't from this than just the thread starter..... now thats better than a book..... if so the forum has worked well as usual.



.................oztules

[kamikaze762x39]

Oh, I am definitely taking in what all of you are saying, and Oztules is right.  The book got me in trouble  

Not to say that I won't check out the links.  

I have to draw the concept visually in my mind.  It is the only way I understand things.  This is proving to be difficult with these new insights.  I completely understand the result now, but I cannot quite seem to understand why.

I know a transformer needs an alternating field to produce output.  I always thought the fluctuating field was exactly equivalent to relative motion--minus the third condition of Faraday's disk, which still blows my mind...

I figured one could visualize the field as a bunch of strings pulling the electrons along.  If the strings are sitting still (a static field) nothing happens... but if if it moves, the free electrons get dragged, and presumably dropped off a few atoms down...

...but then, I find out that the electrons recede as the field decreases in intensity... I am not sure how to visualize that.  It is almost as if they pop back where they were...

It is almost as if the electrons take on a wave quality... Maybe that is where I am going wrong... by only visualizing them as particles... If they had the quality of more of a conjoined medium, like the surface of a pond, than any disturbance would cause a ripple both above and below the original surface plane.  

Could it perhaps be something like this?  I am pretty sure I have just lost everyone    Might as well keep going.

I know there are many vectors to consider besides a simple 2-dimensional wave pattern here.  It is only my attempt at transposing the concept to something visually workable... much like a space-time light cone gets reduced to a simple 2D image.

Maybe a unidirectional current produced by a constantly growing field is akin to an infinitely long pole dropped into a infinitely deep pond.  I suppose the waves could only radiate outward without collapsing back in as we would see with a rock that drops below the surface...  Of course, then we are looking at deflection along the radial plane, and not just the axial plane...  which of course could be akin to frequency and amplitude...

[rossw]

I know a transformer needs an alternating field to produce output.  I always thought the fluctuating field was exactly equivalent to relative motion--minus the third condition of Faraday's disk, which still blows my mind...

The reason I mentioned the transformer is specifically because there is no "relative motion", only a change in field-strength.

I figured one could visualize the field as a bunch of strings pulling the electrons along.  If the strings are sitting still (a static field) nothing happens... but if if it moves, the free electrons get dragged, and presumably dropped off a few atoms down...
...but then, I find out that the electrons recede as the field decreases in intensity... I am not sure how to visualize that.  It is almost as if they pop back where they were...

I don't think thinking in this way actually helps your cause.  Sticking with your "elastic, dragging the electrons along" for a moment - if you drag a magnet *along* the length of a wire, you get no output. Its when you cut *across* the wire you get output.

It is almost as if the electrons take on a wave quality... Maybe that is where I am going wrong... by only visualizing them as particles...

This isn't a good analagy, and if it confuses things even more then disregard it!
Perhaps you can think of the "magnetic field" as being like a sponge full of water.
When you squeeze it, water comes out. When you UNsqueeze it, it sucks water in. If you just wiggle it about, or leave it sit there, the water just sits in it and doesn't flow anywhere. Squeezing (or unsqueezing) changes the sponge density - just like a changing magnetic field density changes.

The main thing I think you HAVE to come to basic grips with is that the entire premise on which ALL this is built, is it's got nothing to do with "north" and "south" poles specifically, and everything to do with "changes in magnetic field over time"

[oztules]

I think we need to separate out the two devices.. ie transformers and alternators and relative motion concepts.

The bosons (force carriers) for the fields are photons. They have no mass and so travel at the speed of light..... so no matter what you do with them, they will approach any object at the same speed... lightspeed. So we can't change the relative speed of the field to the transformer coils as you have previously supposed. So relative motion plays no part in the discussion of transformers. It relies only on the rate of change to the field, and any change we make to it's intensity will be transmitted at the same speed (light speed)... so it is only the change in field at play.

In an alternator with perm magnets, the field intensity around the magnet is fixed (ignore back MMF from the coils for simplicity in this case).... so guess how we get a changing field seen by the coils.

Now relative motion of the magnets as referenced to  the coils give a change in relative distance in time and so a changing  field comes into play.

Stop the motion, and we stop the changing distance.  We have plenty of field, but no emf.... the rotor is stopped.


..............oztules

[Ungrounded Lightning Rod]

I am pretty sure it is a misconception that a dual-rotor configuration consisting of all North poles on one and all South poles on another will produce a waveform above the zero reference (pulsed DC).

A dual-rotor with all north poles on one rotor and all south poles on the other will have "consequent poles".

Think about the magnetic field that went from one pole on one rotor to an opposing pole on the other rotor.  That field line must be a closed loop.  It will come back somehow.

With alternating poles:  It will mostly go out the back of the magnets into the rotors, split sideways, and essentially half of it will come back through each of the neighboring pole pairs, forming about half of THEIR field.  It will go through two pole-to-pole air gaps and four layers of magnetic material.

With all the poles pointed the same way:  Some of it will go through the disk to the hub and back through the shaft.  But there are a lot of strong magnets so that will saturate with lots left over.  Some of the rest will jump back from disk to disk between the pole pieces or inside or outside of their inner and outer radii.  Some more will come out the sides of the magnets - looping from near the back toward near the back of the opposite pole (again either between the pole inner and outer radii or further in/out), from near the back to near the back of the opposite pole, or from near the back to near or on the front of the same pole.  Some more will make it to the rotor and out, but loop back to the same magnet near or at the front rather than crossing to the other side.

The field that makes back to the opposite rotor's magnets forms "consequent poles" of the other direction - north poles on the all-south-poles rotor and vice-versa.  But they are spread out, with only part of their field at a radius where it will cut the coil-sides and induce a useful voltage.  The crossing field looks like the all-N rotor is really NsNsNs where the Ns are the magnet poles and the ses are the effectively weaker consequent poles.  Meanwhile most of the field that loops from front to back of the same magnet is field that would have made it across in a NSNS arrangement, weakening even the main pole-to-pole field.  (This corresponds to the field being weaker because it goes through two pole-to-pole air gaps, two pole-thickness air gaps, and two magnets, rather than two pole-to-pole air gaps and four magnets.

In the NSNS arrangement the strong pole-to-pole field from an N to an S will be cutting a coil side where the wires are going one way, about the same time the strong S to N field from an another pole pair is cutting the opposite side of the coil where the wires are going the other way.  The induced voltage humps on the two sides of the coils are strong and add.

With the NNNN arrangement you have a couple possibilities:  One is that the coil spacing is the right one for an NSNS rotor pair with the same number of magnets.  In that case the weaker but still moderately strong field from an N to S will be cutting one side of the coil while another N to S is cutting the opposite side of the coil.  Because the wires in the two sides of the coil are going in opposite directions the two (substantially weaker) voltage pulses will cancel out and you get essentially nothing (except for junk wiggles from asymmetry).  Ditto when the still weaker s to n consequent pole fields cut the two sides of the coil.

The other is that the coils have a different size / half-coil spacing.  Then the N to S field from a pole pair might be cutting one side while the s to n part-of-a-consequent-pole field is cutting the other.  The weak pulse and the puny pulse add up to something appreciably weaker than if you'd had an N to S and an S to N fied of an NSNSNS arrangement doing the work.  Later the N to S pair make it to the other side of the coil while another s to n pair comes up on the first side.  The same thing repeats in reverse and you get a weak plus puny voltage hump going the other way.  Similar stuff happens if the voltage humps aren't lined up nicely:  You may get asymmetric waveforms.  But the area under the voltage waveform ABOVE the zero line is always equal to the area under the voltage waveform BELOW the zero line.

Another way to look at it is to think of the flux lines transitioning between the inside and outside of the coil.  Any line that comes in must go out and vice-versa.  So over a whole cycle the total number of lines coming in (and inducing voltage one way) equals the total number going out (and inducing voltage the other way).  Then only way to beat this equivalence involves one sliding contact per turn of the coil, which isn't very practical.

By the way:  A better way to do the consequent pole trick is to take a pair of NSNSNS rotor and replace every other magnet on each rotor with a block of iron, so half your poles are magnets and half are explicit iron pole pieces acting as the consequent poles.  The field will then be mostly jumping from pole to pole again.  It will be a tad less than half as strong as if the poles were all permanent magnets.  But it will be a lot better than letting your consequent poles spread out and increasing the gaps.

[Ungrounded Lightning Rod]

I figured one could visualize the field as a bunch of strings pulling the electrons along.  If the strings are sitting still (a static field) nothing happens... but if if it moves, the free electrons get dragged, and presumably dropped off a few atoms down...

Think of the field lines as little springy wedges (or better yet as carrying a crosswise electric field with them, with strength proportional to their speed relative to the electrons they affect).  As a field line running in one dimension moves along a second dimension it applies a force to the electrons that gives them a push, not along its motion, but crosswise to it along the third dimension.  (The "right hand rule" lets you figure out the direction of the push from the direction the magnetic field and the direction of its motion.)

In a transformer the lines are moving because, when the primary current is rising, new circular lines are created in the primary coil wires and jump outward (through other primary and secondary coil wires) into the magnetic core material.  Similarly, when the primary current is collapsing the magnetic field loops shrink and vanish into the wires, cutting through other wires as they go.

A generator differs from a transformer in that the field is dragged through the secondary by moving magnets around rather than expanding a primary coil's field from nothing and then collapsing it again.

[SparWeb]

Oztules, you have a point there.  I can't give up on the old world-wide-interweb thingy just yet.

Back to the OP,

Have you read this FAQ?     http://fieldlines.com/board/index.php/topic,143565.0.html

It was turned "read-only" before I could go back and edit it for clarity and add more diagrams to it, but it's a start.

Carrying on from there, may I ask if you're familiar with the terms "flux linkage" and "electromotive force".

The member Flux has referred to "linkage" but I haven't seen you engage the subject yet so maybe it escaped you.  The subject of the FAQ I was writing introduces it.  What we're concerned with is the quantity of magnetic field that passes through an area defined by a coil loop.  I've made a few statements about the "mathematical" perspective but I guess I'd better stay in the real world.  Usually where we are concerned the magnet is the source of the field, the wire coil passes by it and as it goes by more and more of the field goes through the inside of the coil, until a point where the coil is as close to the magnet as possible.  At this point the coil has the most field passing through it.  The equation for this comes from Faraday's Law (check it out on the hyperphysics pages):

F = B * A * N
  where F = Flux
            B = average magnetic field
            A = area of the coil
            N = number of coils

It's important to note that the coil probably isn't the same size as the magnet, isn't always lined up on the center of it, may be close to it or far away, may not even be perfectly face on to it, for that matter.  These practical considerations can get in the way of relating the theory to doing an actual test.  In fact it's right here that many PMA builders get confused when they start trying the math.  It would seem, from the equation, that increasing the size of the coil would increase the flux.  Not always the case.  The magnet is only so big, so there are only so many field lines coming from its face.  A 2-inch diameter coil is well suited to a 2-inch diameter magnet.  If you need lots of turns then the inside ones have to be a bit smaller and the outside ones have to be a bit bigger.  But making a 4-inch diameter coil is pointless because the magnet's field will fit inside the loop and the rest of the loop won't have any field passing through it.

Something else to consider from that equation.  It does not specify any manner of creating the magnetic field.  It also doesn't tell you where the area came from.  You can use it for laminations in a transformer or a compass needle if you wanted.

A worked example:

Across the face of a cylindrical magnet I measure a uniform field strength of 0.5 Tesla.  The field lines diverge beyond the end of the magnet but on the tip they're pretty straight.  The magnet is 2 cm in diameter, so the area of its end is 3.1 square cm.
I then take a piece of wire, bend it into a square with sides that measure 1 cm across, so the area enclosed inside it is 1 square centimeter.
When I place this one loop in front of the end of the magnet, the LINKED FLUX is:

F = (0.5T)*(0.0001m^2)*(1)  =  0.00005 Weber  =  0.05mWb

The standard unit for flux is the Weber.  In our machines it's usually pretty small, so a "milliWeber" is easier to use.

Let's play with the example for a bit.  What if I made another square of wire, but it measured 4 cm on a side this time?  Its area would be 16 square cm now, much bigger much more flux linkage, right?  Not quite as much as you think.  The size of the magnet becomes the limiting factor.  I showed before that the face of the magnet only offers 3.14 square cm of area.  This is the only part of the area inside the loop where the field passes.  That means that only 3.14 cm^2 of flux linkage is possible...

F = (0.5T)*(0.000314m^2)*(1)  =  0.016 milliWeber

Better, but not 16x better.

The situation can become much more complicated inside a PMA because there are many other magnets and coils in the area.  The field lines diverge in places, so they don't go perpendicularly through the area everywhere.

The best we can do is approximations and simplifications.


The electromotive force is the next step, and it actually IS Faraday's Law, not just the piece of it I bit off first.  Without EMF, you can't tell why anyone would care about the quantity of flux in the first place.  I've written enough for now and I need some sleep.  If I haven't put everyone else to sleep yet, I'll try to continue tomorrow.

PS
Ghurd, I finished putting together your kit.  Testing will begin (in earnest) soon.

[kamikaze762x39]

Ungrounded Lightning Rod,

I have thought about the vectors you speak of concerning the flux paths.  The problem I saw with an alternating pole arrangement is that with an NSN on top of a SNS, you would have quite a few flux lines attempting to travel parallel to the vector of motion (tangential axis on the rotor)...

That close NS arrangement in the gap should be nice and strong, but if the poles have opposite poles adjacent to them, the lines will be drawn to them, and not only directly across the gap...

So if we think in terms of NsNs or SnSn, could not much of the "wild" flux find its way around the radial edge of the rotor to meet with its counterpart?  It really is a question... just thinking out loud.  Perhaps that is why we end up with a n pole and not a N pole. 

Now... if we continue along that course... We may possibly even discover that NsNsNs has much bigger N's than does NSNSNS, because NSNSNS has a lot of flux drawn along the tangential axis of the rotor.  Widen the gap, and this adverse effect (if, in fact, it is adverse) becomes even more observable I am assuming. 

So, if my premises are correct, the tangential flux is essentially wasted in the NSNSNS arrangement, just as the "wild" flux is wasted in the NNNNNN arrangement.  It is my hypothesis that the consequent pole arrangement will have a more sharply focused field in the gap, thus more (delta)B, hence more power.  The wasted flux in one configuration or other may offset this balance.  I don't know...  But for some reason, I intuitively went with the consequent pole arrangement on a wooden rotor.  Let the stones fly. 

In a transformer the lines are moving because, when the primary current is rising, new circular lines are created in the primary coil wires and jump outward (through other primary and secondary coil wires) into the magnetic core material.  Similarly, when the primary current is collapsing the magnetic field loops shrink and vanish into the wires, cutting through other wires as they go.

See... that is exactly the kind of thing I am trying to look at.  The field is cutting in different directions as it shrinks and grows.  I wish I knew what those directions were.  I thought I did using the right-hand rule, but it seems that other factors are at play.

Sparweb,

I will check out the FAQ, and you are correct in assuming that flux linkage has escaped me for the time-being.  Just from looking at your explanation of it, it is the concept involving a sort of saturation of usable flux in the field to coil relationship.  I have not hammered out the figures yet, but I will try to.  The formula looks simple until an integral gets thrown in there, with which I have had no experience.  Yet I see that you have posted what seems to be a simplified version of the formula wiki has.  F=BAN looks simple enough, so I will check into it.  Thanks.

I understand EMF, but flux linkage is a new concept to me.  I wish we hadn't jumped straight into digital in school, because it seems I missed a lot of AC concepts that should have been explained better. 

Rossw,

I don't think thinking in this way actually helps your cause.  Sticking with your "elastic, dragging the electrons along" for a moment - if you drag a magnet *along* the length of a wire, you get no output. Its when you cut *across* the wire you get output.

Yes, I did have the wrong vector there with the analogy, but the entire analogy was a bit flawed anyway. 

I see now that it has nothing to do with north and south, which seems to be contradictory to Fleming's rule.  I think the problem, as I have been saying, is that I am not seeing all the vectors of force and motion and how they relate to a coil's observation.

So... what I am essentially trying to grasp is--does the changing field produced by the relative motion actually change Fleming's north to south vector?

Oztules,

The boson, light speed explanation helps a bit.  I have studied relativity, and thus now understand that the electrons don't "see" speed changes.  Thus, even in an alternator, relative motion doesn't mean jack to a strand of wire... So, if motion is completely eliminated as a factor, that only leaves us with field intensity, its vector, and its change over time. 

So are we aiming more at change to determine the direction current flow, or more at the field vector created by the change?

[Flux]

The flux linkage concept is very simple.

Any coil is a loop, it may be one turn or many that doesn't matter. Any changing flux linking the loop will produce an emf. The flux needs to be changing for this to happen. It is easiest seen in the transformer but there is no difference if you use magnets.

Any flux from your magnet that doesn't link the coil doesn't contribute to the emf. Looking at it that way you don't need to worry about how the magnets are arranged as long as you can work out which flux lines link your loop.

Consequent poles work fine for decent low reluctance iron circuits but there tends to be a bit more leakage flux that won't link the loop. For magnetic circuits without iron they really don't work well, the flux path is not defined.

Flux

[SparWeb]

I will check out the FAQ, and you are correct in assuming that flux linkage has escaped me for the time-being.  Just from looking at your explanation of it, it is the concept involving a sort of saturation of usable flux in the field to coil relationship.  I have not hammered out the figures yet, but I will try to.  The formula looks simple until an integral gets thrown in there, with which I have had no experience.  Yet I see that you have posted what seems to be a simplified version of the formula wiki has.  F=BAN looks simple enough, so I will check into it.  Thanks.

I understand EMF, but flux linkage is a new concept to me.  I wish we hadn't jumped straight into digital in school, because it seems I missed a lot of AC concepts that should have been explained better. 

Hi Kamikaze.  Now I feel like we're getting somewhere.  You don't really need the calculus so don't worry.  I did put it in your face so that I could gauge whether you did or not.  Couple of things to observe right away:
You used the word "saturation" but not in the proper context.  Later on we could go into that, but for now there is no reason to use the word when discussing the flux linkage between magnet and coil.  Second, if you are just getting used to "flux", then you don't understand "EMF".  Not as much as you think.  (I'm being blunt but I don't have much free time today).

You're not alone in being set adrift by incomplete schooling now that computers do everything for people.

I'll see about giving you a part 2 later but very busy with 1/2 staff at work today.

[kamikaze762x39]

Flux,

I was afraid that might be the case after doing some more digging on the subject.  But, I do have good results so far.

Right now, I have the rotors touching the stator, the stator is unsecured as of yet, so there is a bit of rubbing.  But I do have the rotor locked down pretty good to the butt end of a bicycle frame where the wheel connects.  I sawed it off the bicycle, so the alternator is standing pretty well upright.  I hooked up a multimeter to one phase (6 wires protruding), held the stator with my palms and spun the rotor with my fingers.  I got a 6 volt AC reading just by this slight little push (just an instantaneous peak reading).  

The bearings aren't even lubed (plain bearings).  I plan to drill and tap the end of the shaft so I can hook it to my drill after securing the stator and adjusting the gap.  I have 36 small gauge coils packed pretty tightly on an ~18 inch rotor.  I expected voltage  

I mixed iron powder with polycrylic and used a brush to paint the mix onto the wires of the coils.  I figured I would try increasing permeability directly on the wires rather than putting a core in the center of each coil.  There is very little tug by the magnets, so cogging should not be a problem.  Maybe it was a mistake to coat the wires in iron... I don't know, but, again, 6 volts with only a very slight push... far less than the device's full potential, I'm sure.

 I tried to better "define" the flux path in this way, but I am considering running some iron laminates from the back side of each rotor, through the jacking bolts and to the back side of the other rotor.  I'm not sure yet.  Maybe I'll post pictures eventually.

SparWeb,

Hey, not a problem  There are people paying a couple hundred bucks an hour for this kind of thing.  I'm just glad so many people like to see themselves type as much as I like reading it.  I look forward to part 2.

[Ungrounded Lightning Rod]

I have thought about the vectors you speak of concerning the flux paths.  The problem I saw with an alternating pole arrangement is that with an NSN on top of a SNS, you would have quite a few flux lines attempting to travel parallel to the vector of motion (tangential axis on the rotor)...

That close NS arrangement in the gap should be nice and strong, but if the poles have opposite poles adjacent to them, the lines will be drawn to them, and not only directly across the gap...

Consider:

 - Pull the stator out of the middle and bring the rotors together until the poles touch:  Zero gap.  No sideways leakage at the middle because it cancels by symmetry - all the flux at the pole piece faces goes to the opposite pole.

 - Open some gap a bit.  Some of the flux takes the harder path sideways.  But most of it still goes across to the face-on opposing pole.  The flux that goes sideways corresponds to the weakening of the field because it must fight its way through the added magnetic reluctance (analogous to electrical resistance) of the added gap, and some of it choses instead to go through instead through the reluctance of the air toward the adjacent pole sideways.

 -  But the gap gives you room to insert some coils to generate power.  And the weakening of the field is an unavoidable consequence of making the gap.  If there wasn't a pole nearby some of the field would still chose to fold back beside the magnet and you'd still lose strength in the gap.  And the backside of the magnet is ALWAYS nearby with these strong, thin, magnets.

 - Widen the gap and the field weakens more but you get to add more coils.  What's optimum?

 - When open the gap the first few wire thicknesses the field is hardly weakening at all.  You get to add more layers of coil and it's all gravy.  When the gap is far wider than the magnet spacing the field (a dipole field) is falling off by a drastic inverse-cube law (inverse square from loss of the field from each magnetic particle times another inverse-linear due to the smaller angle to the adjacent canceling pole).  The weakening field loses you more than the added room for more turns gains you.  Somewhere between that is a sweet spot.

 - Assuming the next-door neighbor pole is "far enough away", the sweet spot is about where the gap is the same thickness as the sum of the thickness of the magnetic material in the opposing pole pieces.

 - Because the change-of-power vs. width-of-gap-full-of-coils function is continuous, this happens where it is horizontal, i.e. not changing appreciably with gap adjustments.  So it's "get it in the ballpark" not "critical for good performance".  Also:  Magnets come in discrete sizes convenient to vendors.  Getting the absolute most out of the ones you have may not be an issue - compared to, for instance, matching the genny to the prime mover or having enough copper and cooling to avoid a meltdown.

 - Nevertheless you want to be moderately efficient, if only to avoid paying extra for the next size thicker magnets.  So how far apart is far enough?  Simple:  If the sideways gap is larger than the thickness of the magnets and larger than the gap between the rotors, the contribution from adjacent-pole cancelation will be smaller than the contribution from front-to-back flux leakage.

 - Normally the side-to-side gap between the magnets will be something like half the width of the pole, and much wider than the thickness of the magets.  So its contribution is minor.  (Lots of magnets close together, as when trying for high frequency, may get you into the geometries where you have to worry about it.)

(Meanwhile, with NSNSNS, you WANT the field to connect between the back sides of the opposing poles.)

So if we think in terms of NsNs or SnSn, could not much of the "wild" flux find its way around the radial edge of the rotor to meet with its counterpart?

Yep.  That's where the consequent poles come from in NsNsNs:  The backside field bending forward around the magnets.

... We may possibly even discover that NsNsNs has much bigger N's than does NSNSNS, because NSNSNS has a lot of flux drawn along the tangential axis of the rotor.  Widen the gap, and this adverse effect (if, in fact, it is adverse) becomes even more observable I am assuming.

Except that with NSNSNS the backside fields mostly want to connect to the adjacent magnets' backsides, rather than come forward and partially cancel the front poles.

Also:  Remember that your NSNSNS is really N_S_N_S_N_S  while your NNNN is really NsNsNs.  So your NNNN case has MUCH closer spacing between its N and consequent S poles than the NSNSNS design, giving it much greater adjacent-pole cancellation.

So, if my premises are correct, the tangential flux is essentially wasted in the NSNSNS arrangement, just as the "wild" flux is wasted in the NNNNNN arrangement.  It is my hypothesis that the consequent pole arrangement will have a more sharply focused field in the gap, thus more (delta)B, hence more power.

Nope.  The field is actually more focussed in the NSNS arrangement, due to the diffuse nature of the consequent S poles and of the field as it crosses the gap from one N to any S pole on the other side, reducing the variation with position that gives you generation.   Dig around on the site and you'll see some nice computed diagrams of this sort of flux.

But for some reason, I intuitively went with the consequent pole arrangement on a wooden rotor.

Sigh.  New data:  The rotors are wood.  Now in addition to jumping the gap between the rotors the field also has to jump the big gap between the backs of the magnets.  LOTS more reluctance.  Lots more backside field coming forward and canceling frontside field.  Hint:  NSNS still wins big.

In a transformer the lines are moving because, when the primary current is rising, new circular lines are created in the primary coil wires and jump outward (through other primary and secondary coil wires) into the magnetic core material.  Similarly, when the primary current is collapsing the magnetic field loops shrink and vanish into the wires, cutting through other wires as they go.

See... that is exactly the kind of thing I am trying to look at.  The field is cutting in different directions as it shrinks and grows.  I wish I knew what those directions were.  I thought I did using the right-hand rule, but it seems that other factors are at play.

While you could figure it out with the right-hand rule, it's easier to use a rule of thumb:  The expanding and contracting field, as it cuts the metal of the wire carrying the current generating it, always induces a voltage that opposes the voltage that is trying to CHANGE the current.  When current is increasing and the field is expanding the induced voltage opposes the voltage pushing the current.  When the current is decreasing and the field shrinking the induced voltage helps make up for the loss of the applied voltage that had been pushing the failing current.

[kamikaze762x39]

Ungrounded Lightning Rod,

I see what you are saying now.  I should have done some remediation on magnetic circuits.  Too late to change it, though.  The best I can hope for at this point is using some iron to create a path.  Anyone know if magnetic circuits work well with a common ground like electric circuits?  I may tie all the "loose ends" to the same chassis, such as a bolt or the center shaft.

The magnets would be easy enough to remove from the rotors and swap... but this would be a very bad thing to do since I have 12 magnets and 36 poles.  I would not get any net current due to the opposing arrangement.  The coils are already cast, and I'll be a monkey's uncle if I'm going to wind and cast a new stator.  I think I can make it work more effectively with a few modifications.

[finnsawyer]

Oh, I am definitely taking in what all of you are saying, and Oztules is right.  The book got me in trouble  

Not to say that I won't check out the links.  

I have to draw the concept visually in my mind.  It is the only way I understand things.  This is proving to be difficult with these new insights.  I completely understand the result now, but I cannot quite seem to understand why.

I know a transformer needs an alternating field to produce output.  I always thought the fluctuating field was exactly equivalent to relative motion--minus the third condition of Faraday's disk, which still blows my mind...

I figured one could visualize the field as a bunch of strings pulling the electrons along.  If the strings are sitting still (a static field) nothing happens... but if if it moves, the free electrons get dragged, and presumably dropped off a few atoms down...

...but then, I find out that the electrons recede as the field decreases in intensity... I am not sure how to visualize that.  It is almost as if they pop back where they were...

It is almost as if the electrons take on a wave quality... Maybe that is where I am going wrong... by only visualizing them as particles... If they had the quality of more of a conjoined medium, like the surface of a pond, than any disturbance would cause a ripple both above and below the original surface plane.  

Could it perhaps be something like this?  I am pretty sure I have just lost everyone    Might as well keep going.

I know there are many vectors to consider besides a simple 2-dimensional wave pattern here.  It is only my attempt at transposing the concept to something visually workable... much like a space-time light cone gets reduced to a simple 2D image.

Maybe a unidirectional current produced by a constantly growing field is akin to an infinitely long pole dropped into a infinitely deep pond.  I suppose the waves could only radiate outward without collapsing back in as we would see with a rock that drops below the surface...  Of course, then we are looking at deflection along the radial plane, and not just the axial plane...  which of course could be akin to frequency and amplitude...

[finnsawyer]

Oh, I am definitely taking in what all of you are saying, and Oztules is right.  The book got me in trouble  

Not to say that I won't check out the links.  

I have to draw the concept visually in my mind.  It is the only way I understand things.  This is proving to be difficult with these new insights.  I completely understand the result now, but I cannot quite seem to understand why.

I know a transformer needs an alternating field to produce output.  I always thought the fluctuating field was exactly equivalent to relative motion--minus the third condition of Faraday's disk, which still blows my mind...

I figured one could visualize the field as a bunch of strings pulling the electrons along.  If the strings are sitting still (a static field) nothing happens... but if if it moves, the free electrons get dragged, and presumably dropped off a few atoms down...

...but then, I find out that the electrons recede as the field decreases in intensity... I am not sure how to visualize that.  It is almost as if they pop back where they were...

It is almost as if the electrons take on a wave quality... Maybe that is where I am going wrong... by only visualizing them as particles... If they had the quality of more of a conjoined medium, like the surface of a pond, than any disturbance would cause a ripple both above and below the original surface plane.  

Could it perhaps be something like this?  I am pretty sure I have just lost everyone    Might as well keep going.

I know there are many vectors to consider besides a simple 2-dimensional wave pattern here.  It is only my attempt at transposing the concept to something visually workable... much like a space-time light cone gets reduced to a simple 2D image.

Maybe a unidirectional current produced by a constantly growing field is akin to an infinitely long pole dropped into a infinitely deep pond.  I suppose the waves could only radiate outward without collapsing back in as we would see with a rock that drops below the surface...  Of course, then we are looking at deflection along the radial plane, and not just the axial plane...  which of course could be akin to frequency and amplitude...

[finnsawyer]

No, trying to ram the flux from several magnetic paths through a single section of iron will drive the iron into magnetic saturation.  Go with ULR's flux splitting arrangement.  That's what usually happens anyway.

The problem that I see with this thread is ignoring the role of time in creating the induced voltage from the loop or coil.  Faraday's Law states that the voltage induced around a loop is proportional to the area of the loop times the time rate of change of the magnetic flux through the loop.

             That is:  V=Axd(B)/dt.

Now, A is assumed fixed and unchanging.  B changes with time and has direction.  That is, B can be negative or positive and can increase or decrease in magnitude.  As an example, you can have a positive B decreasing with time causing a negative voltage and the opposite B (negative) becoming more negative with time causiing a positive voltage in the same loop or coil (at different times).  Or positive B increasing giving a positive voltage or negative B becoming less negative giving a negative voltage.  For more fun check my alternator design in my diary.


I see what you are saying now.  I should have done some remediation on magnetic circuits.  Too late to change it, though.  The best I can hope for at this point is using some iron to create a path.  Anyone know if magnetic circuits work well with a common ground like electric circuits?  I may tie all the "loose ends" to the same chassis, such as a bolt or the center shaft.

The magnets would be easy enough to remove from the rotors and swap... but this would be a very bad thing to do since I have 12 magnets and 36 poles.  I would not get any net current due to the opposing arrangement.  The coils are already cast, and I'll be a monkey's uncle if I'm going to wind and cast a new stator.  I think I can make it work more effectively with a few modifications.