Best coil geometry for round magnets

[DanB]

Hi Chris - that graph confuses people and I never seem to describe it right.  It's not that it produces more power than there is available at the shaft.

the 'betz' curve there is just calculated and stuck on there for so folks can get a rough idea what the maximum power at the shaft could be from a 10' diameter blade.  The power from both alternators is measured (off a tractor PTO).  The point here is, it (alternator 2) puts out more power than 10' blades could ever hope to provide at the shaft... which is why a good bit of resistance is required to be added to the line to make these run.

It also shows the increase in power I get when using basically the same stator but just changing from 1" x 2" rect. magnets to 2" diameter round magnets.

The graph obviously does not show power in to the alternator - I've measured that but I don't have it here now.  It does pretty nearly line up with what you'd expect if you take power out and add I^2 * R losses in the stator just knowing the resistance of the stator - and my measurements of torque are not that accurate so... my actual tests are probably not as scientific as what you are doing. 

Actually, Alternator 1 is a pretty good match for 10' blades running at a reasonable TSR and most of those machines I and other folks have built worked out really well, but the resistance within the stator is such that they tend to burn up if sustained output over 600 - 700 Watts happens for long.  So my 'quick/dirty' solution, which seems to upset the alternator geometry puritans    was to put in larger round magnets, cut the resistance in the alternator by about half while keeping the same cutin speed... and that's 'alternator 2'.

[ChrisOlson]

Hi Chris - that graph confuses people and I never seem to describe it right.  It's not that it produces more power than there is available at the shaft.

Yes, I just made an edit to my previous post and figured that out.

What really surprises me is that you can only get 600-700 watts from a gen with bars in it without burning it up.  I have pushed those to 1.5 kW using what I'll call an "Ed Lenz design" but using bars instead of wedges, and 10" rotors instead of 12" - with absolutely no heat problems.  If you check one of my posts in another recent thread here, I put some pictures of one of these generators in that post because Ungrounded Lightening Rod wondered about it.  I used wedges in that one but the only change I have to make for it to use bars is wind 5 more turns for the same air gap.  The resistance is not all that different between the two - but the one with wedges makes 11% more power at the same rpm than the one with bars - and it runs 2-4 degrees C hotter than the one with bars, measuring the temperature with my infrared pyro that I use to check exhaust port temperature on diesel engines when adjusting injectors.  That reading is taken at the full amp rating of the wire that they're wound with, and these are delta generators.

It seems my gogle thing got messed up where I had those pictures so I'm putting the picture of the rotors I used on that one here.  I'll have to dig to see if I can find the picture of the stator with the coil layout - I had a picture of it with all the coils in it in my brand new steel stator mold that I built.
--
Chris



I found that photo and put it on here as an attachment.  This generator is the most efficient one I've built to-date.  It came in at 67.7% power efficient @ 1,560 watts on a 24 volt system with the voltage running at 29.0 volts.  After 10 minutes the stator temperature was 49.7 degrees C (about 120 degrees F).  You could burn your hand on it - but that's at the full amp rating of the wire - and well below the rated temperature of the wire.  Very efficient generator and I'm extremely happy with it.  I could make it more efficient yet if I'd reduce the turns in the coils slightly, cut it in in wye and switch it to delta with an automatic controller of some sort.  This one fit under 10" gen rotors and I could get it down to 9.5" and maintain proper leg spacing to prevent heat buildup and use a ventilated stator with exposed coils (not encased in resin) to provide better cooling for a better Service Factor in power spikes.  I'm thinking I could push this one to 2 kW with those changes for a period of time that I'd have to arrive at thru testing and burning one up on my test setup.

Just wanted dciolek to see what you CAN do with a little engineering and burning up a few stators to figure it out (plus the help of Ed Lenz and several other people who gave me pointers)   

[DanB]

What really surprises me is that you can only get 600-700 watts from a gen with bars in it without burning it up.  I have pushed those to 1.5 kW using what I'll call an "Ed Lenz design" but using bars instead of wedges, and 10" rotors instead of 12" - with absolutely no heat problems.

There are two possibilities there... either alternator design and geometry, where you may have me beat...  or rpm, or both.  At what rpm do you get 1.5kW?  It's a lot to do with what's driving it and how we load it.  Of course I could, from the same alternator get 3kW or whatever, depending upon how fast I run it and the load.  One particular version of this alternator, for charging 48V batteries needs to cut in at 140 rpm and the limit is about 15 amps which should happen somewhere between 300-400 rpm, depending on battery voltage and line loss  - it's wound with AWG 16 wire for that application and we can't push that wire much harder for long.

The similar alternator on my diesel engine is wound with 3 strands of AWG 15 wire (if I recall correctly) and it's good for 40 amps (2kW) sustained output and efficiency I believe is around 90% - it doesn't warm up at all at 2kW, but I run that at 600 rpm and it wouldn't start charging at all until about 550.

[ChrisOlson]

There are two possibilities there... either alternator design and geometry, where you may have me beat...  or rpm, or both.  At what rpm do you get 1.5kW?

Hi Dan,

I edited that above post when I located that picture of the coil layout.  This generator actually has wedges in it and it makes 1,560 watts @ 29 volts @ 330 rpm.  I built it for 13 foot PowerMax blades but I haven't put those blades on it yet to see how good the match is.  Those blades put 2.3 kW to the shaft @ 24 mph and with the generator at 67.7% that should be dead on what this one put out when I tested it - right around 1.5 kW.  It should be letting the rotor spin at ~6.5 TSR @ 24 mph.

I'm still driving it with a set of cheap 9.4 foot WindMax blades because the 13's I ordered for it haven't come yet.  I had to open the air gap so wide you can just about slip a 2 x 4 between the rotors and stator to run it with those little blades.  This gen has AWG 14 in it and it's delta wired (basically the same as two-in-hand AWG 14 in star).

Edited note for Dan:
If you decide to try building one of these to test for yourself, I had to make a compromise between having the coil legs perfectly parallel with the magnet path vs how much leg overlap I was going to get from a generator pole.  Ed's theory is that having the coil leg parallel is more important so I ended up, right at the very inside edge of the magnets, with .050" overlap.  Those wedge magnets are somewhat of an ideal shape so they're narrower at the bottom than even 1" bars, which helps quite a bit in designing coils to fit.  Ed's little 8" generators that'll push 80 amps on a 12 volt system are more ideally laid out, but this one didn't turn out too bad for a 24 volt.
--
Chris

[ChrisOlson]

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.

DanB,

I read this yesterday and found this statement pretty interesting.  The way you describe above sounds kind of more like an iron or steel core unit.  The way I envision these air core generators is very simply highly concentrated flux lines between the two rotors being cut by wire.  Wire that is laid out parallel with the direction of travel of the flux makes power, wire that is perpendicular to it makes no power.  Wire that is 45 degrees to it only makes half the power it could.  So design the coil so the part that's going to make power is positioned to cut the maximum lines of flux - and don't waste any of it by overlapping coils with one pole, having coils too short so the magnet wipes wires that aren't laid out parallel to the magnet travel, etc.

In other words, the way I look at it, it has EVERYTHING to do with magnets moving over coil legs.  And the more magnet area you have, in the proper balance of length and width, the more precisely you can wipe the coil leg with the flux, turning it on and turning it off again at the right time to get a perfect sine wave.  And the more perfect you get that, the less harmonics and vibration you have - and you get the corresponding increase in performance.

It's kind of like building engines - I'm sure you've heard of the term "balanced and blueprinted".  This procedure is usually good for 7 hp on the dyno with your run of the mill 355 cubic inch circle track engine.  All it involves is balancing the crank, flywheel and harmonic balancer on a dynamic spin balancer and either drilling or welding counterweights to balance that assembly.  Then weight match all the reciprocating assemblies, including dynamic balance of the rod big ends.  Upon assembly you clearance everything precisely - piston to bore, mains and rods, crank endplay.  The result delivers that 7  hp on the dyno every time as compared to an engine that's built using production line tolerances - merely by tuning up the bottom end and getting rid of vibration and excess frictional losses.

So that's my current theory- pay attention to details when you build them.
--
Chris

[DanB]

Hi Chris - now this is getting interesting.  This topic really is something I think about often and after your last post which I read last night, just before bed... I found myself dreaming about 'alternator geometry' all night.  

I read this yesterday and found this statement pretty interesting.  The way you describe above sounds kind of more like an iron or steel core unit.  The way I envision these air core generators is very simply highly concentrated flux lines between the two rotors being cut by wire.  Wire that is laid out parallel with the direction of travel of the flux makes power, wire that is perpendicular to it makes no power.

I expect you meant to say things the other way 'round.... wire that is perpendicular to the path of the moving poles makes power, wire that is parallel makes none.  At any rate I understand what you mean.

Wire that is 45 degrees to it only makes half the power it could.  So design the coil so the part that's going to make power is positioned to cut the maximum lines of flux - and don't waste any of it by overlapping coils with one pole, having coils too short so the magnet wipes wires that aren't laid out parallel to the magnet travel, etc.

In other words, the way I look at it, it has EVERYTHING to do with magnets moving over coil legs.  And the more magnet area you have, in the proper balance of length and width, the more precisely you can wipe the coil leg with the flux, turning it on and turning it off again at the right time to get a perfect sine wave.  And the more perfect you get that, the less harmonics and vibration you have - and you get the corresponding increase in performance.

Right.  I really did used to think exactly the same thing, and if you look at the machine we build with 1" x 2" blocks, we're pretty close to getting that bit just right.  It's part of the reason we made wedge shaped coils - and for a while, I was using wedge shaped magnets, which ~ I still have to agree, must be ideal.  

But again, if you look at the difference in power out on my graph, between the alternator that's using the blocks, vs the one that's using the disks - I don't think it supports that theory.



The image above shows the wave form from a 12' turbine we built a couple months ago - that machine actually crowds the magnets slightly worse than our 10' machines.  Often times in that machine, there are the edges of two poles over the same leg of a coil at the same time and it really doesn't show up as the problem I would've expected.

I used to think that conductors that were 'parallel' (not the legs of the coils  but the tops and bottoms) to the changing magnetic field did very little for us... except add resistance.  So there in my thinking was... should it be a triangle like you have, with legs perfectly perpendicular to the path of the magnets, and a very short bottom and a fairly long 'top' (and the top in that case would do nothing) - or... a sort of compromised oval shaped but still slightly wedge shaped coil, which might be about the same when all things are considered.

It was really somewhat disturbing when I went to one of Hugh Piggotts workshops to help out and he had a new 6' turbine that he'd designed, with 1" x 2" magnets that were in there sideways.  (in other words - the magnet rotors were small, 8" in diameter I think and the magnets and coils were in there sideways)  In that machine, with 1" x 2" magnets, the hole in the coil was 1" tall and 2" wide (according to your current thinking and mine at the time if I understand you correctly, only 1" of the coil was making power and 2" were doing very little).

So I did some fairly unscientific tests when I got home.  One involved a single strand of wire instead of a coil and I measured that on my antique scope... if the magnets were setup on the rotors tall, instead of wide (like we normally do)  - I got close to twice the voltage induced like you would expect.  If it was a coil though (a single wire) - there was very little difference (there was some difference).

Honestly - I really do think you can look at these like a transformer for the most part, it's mostly about flux coupled through the coil.  The tops and bottoms of the coils (those bits that are parallel to the path of the magnets) have almost as much EMF induced in them per given length as those bits that are perpendicular (not quite as much... but almost as much).  I really think there is a bit of both things going on, but the major part is simply flux changing through the coil and the concern about having certain poles over the legs at the same time is a more minor issue.  And I believe that the wave form shown above, and my tests  between the fairly correct 'alternator 1' and the fairly obnoxious 'alternator 2' with the crowded 2" magnets (the graph on the previous page) add evidence to that.

At any rate - it's all very interesting  - these air core axial flux machines are not quite as simple to fully understand as transformers, or more conventional alternators.  There's more stuff going on to think about I think.  

The performance of your alternator is definitely impressive... it's similar to what I'm getting, definitely a bit smaller in size though.  Some of the difference could be to do with air gap and battery voltage - in my tests I tried to keep my battery down to about 50 Volts and I run a pretty wide airgap, especially in 'alternator 2' with the larger magnets in it.  Another possibility is the shape of your coils, I'm still not convinced of that but have not ruled it out.  My bet though, is you could get more power from the machine if you shrank the hole in the middle, widened them so that they actually touch each other, so you fit more copper in there, even though doing so would be slightly violating some of your guidelines.  You're definitely are getting impressive results though - and again, I think there is some factor with getting the right poles over the legs at the same time, but I believe it's a smaller factor than  you think it is.  (again - if it was a big problem, it should show up in the wave form)

So... shall I take my old volvo, spend gobs of time and money to  blueprint the engine, put in a nice cam, port / polish the head, and install a nice pair of weber DCOE 40's... and maybe get 160 hp, or just go get a chevy with a sloppy old big block 400 and call it good at 300 hp!  (just kidding here... )

[ChrisOlson]

So... shall I take my old volvo, spend gobs of time and money to  blueprint the engine, put in a nice cam, port / polish the head, and install a nice pair of weber DCOE 40's... and maybe get 160 hp, or just go get a chevy with a sloppy old big block 400 and call it good at 300 hp!  (just kidding here... )

No, the bore/stroke ratio on those small block 400's is bogus - not a good choice.  Something a little more square like a Enderle injected 572 KB Hemi with a 12-71 Littlefield blower would be more impressive.  Pretty much the ultimate in light weight, can't break it, and output per unit of displacement   

We're going to have to get together sometime with coffee, lots of napkins to draw on and permanent markers to figure all this out.  Some of the best designs on earth came from coffee shops with drawings on napkins - just ask Burt Rutan sometime how Voyager was conceived   
--
Chris

[Flux]

I originally thought this was about electrical efficiency but now I see it is about the efficient use of materials and this makes more sense.

All factors being equal then the only sensible choice for an axial machine is trapezoidal magnets. This means trapezoidal coils and you can get the most wire in the appropriate space, I don't think there is any doubt about that.

When we consider costs then unless you have a special supply of magnets the common rectangular and circular ones will be cheaper because of volume supply and use and to me it seems that a few shapes are very much more cost effective than the others. this is why I have always used rectangular magnets.

If you go for the cheapest materials rather than the ideal ones it somewhat changes the argument. Given an equal choice i could see no reason to use rectangular magnets in something with areas based on sectors, but if you do go for rectangular ones there is no reason whatsoever to have more losses or lower electrical efficiency, you will end up with a different mechanical layout and you may use a bit more magnet or copper but there is no question of parasitic losses unless you do something silly.

I don't find round magnets cost effective but if your supply is such that they are cheaper then go for round magnets.

Once you depart from trapezoidal magnets then the ideal coil shape and dimensions will change so there really are no exact rules unless you specify all the parameters.

Once you depart from the ideal magnet shape then there is little point in worrying too much about coil shape, there is no law that requires straight sides. the requirement is that for maximum use of materials you link all the flux of each magnet without linking the neighbouring circuit.

Ideally the hole in the coil needs to be magnet size and shape or bigger and when the magnet is directly over the magnet you don't have parts of the coil linking adjacent poles. this sets some limits on coil dimensions relative to spacing of magnets on the disc and there will be an optimum magnet spacing.

So far this only affects emf and in reality the winding resistance is a very big factor , so much that optimising the emf without regard to resistance is not a great idea. This is why in practice you can benefit from coil holes below magnet size, the extra turns still add emf but don't link all the flux. if the added resistance is tiny this extra emf still helps.

Similarly the coil shape sensibly should be the same as the magnet to keep the minimum wire length and minimum resistance. In real life again you can squeeze the inner part of the coil to get more copper in the main area without big losses in emf and if by doing so you get thicker wire in then it works better. Circular magnets will have least resistance with round coils but if the magnets are crowded then squashing them elliptical or triangular may well be better.

If you optimally link all the flux in a single turn with any magnet shape the mean emf will be the same and the waveform will be virtually rectangular. As Dan pointed out it doesn't present a problem if you turn rectangular magnets sideways as long as the flux linkage is the same. It is the area of the magnet that matters, Turns are closed loops there are not active bits and dead bits the best result will come when the resistance of the loop is a minimum and that goes in favour of round magnets and coils but there other factors involving space and stacking that in reality favour the trapezoidal.

If you want delta with no circulating current then you need a sinusoidal phase voltage and that means a specific distribution of turns to cancel the in phase odd harmonics with the out of phase ones. It doesn't have to be exact by any means but you do need to make some effort.

Just as Sparweb has found my tests always show the highest efficiency in star, with IRP next and delta worse but these are fairly tiny factors and if you feel there is some particular virtue in delta then use it but you will need to be more careful with the winding. I don't think it is any issue at all with star /delta changeover but I still prefer star at cut in. Irp clears any circulating harmonics below cut in but it returns if the rectifier conducts. Above cut in the whole waveform is such a mess that harmonic loss is inevitable with all connections, the only gain is with star in the first few watts above cut in, beyond that use whatever you fancy.

I am sure we can optimise shapes for all magnet geometries but cost per unit output depends so much on the source and cost of the component parts.

I can go along with all this but I can't see that that more power is available from a given wind input with one winding or another if the overall electrical efficiency and the loading is identical.

I have little doubt that Chris has found more or less the ideal spacing and coil dimensions for trapezoidal magnets and as such will be able to get more out for a given size of alternator, it may not be the cheapest method or the most convenient method unless others can get the magnets at a sensible price. Similarly you can get the same performance from smaller magnets and more copper or vast magnets and relatively small amounts of copper but the cost effective machine will be determined yby the price you have to pay for copper and magnet and that changes with time and source of supply.

If you want cheaper designs with almost identical performance look at air gap radials without slots, I find it far more cost effective and the cooling is much better, there are many ways to skin the cat.

Flux

[DanB]

All this stuff said, just one mostly unrelated thought for dciolek:

If your rotors are 12" in diameter and you use those round magnets, and look at my graph on page two of this discussion, you could expect output something like 'alternator 2' assuming you're charging batteries.  And as others have pointed out - you could get more power at the same rpm (or more efficiency... or both) with better geometry, which in my mind will basically mean, 14" rotors, and larger coils wound with thicker wire and perhaps a few less turns, but such a machine will really miserably stall 10' blades so you might be looking at larger blades in that case.

My one concern that I just now noticed... if your brake rotors are 12" in diameter, I wonder what the bolt pattern is?  For 12" rotors I normally use a 4" diameter bolt pattern and there is very little room left on the inside (like 1/4") after I wind the coils.  If your bolt pattern is larger in diameter, you may not have room for the copper where it really should be (that copper at the bottoms of the coils, in between the magnets, and the studs which hold the whole thing together).

Just something to ponder... i'd be a shame to wind a stator and find out the hole in the middle isn't big enough!

[ChrisOlson]

which in my mind will basically mean, 14" rotors, and larger coils wound with thicker wire and perhaps a few less turns, but such a machine will really miserably stall 10' blades so you might be looking at larger blades in that case.

I think what you're saying here, Dan, is that those 2" disc magnets are simply too big for this size generator and blades.  There's a LOT of magnet area there - 3.14 square inches, compared to 2 for bars and 2.35 with wedges.  What if dciolek would build a different configuration?  Like a 8 pole 6 coil?  I see no reason why those big mags wouldn't work better for 10 foot blades with that sort of setup - a better "match" to the available power to drive it.

Just something to ponder... i'd be a shame to wind a stator and find out the hole in the middle isn't big enough!

Yes.  This is a big one!  If you noticed in the photos of my generator, I had to make the coil interconnects on the outside diameter of the coils (I'd rather do it on the inside).  There just wasn't room on the inside and still keep the hole big enough.
--
Chris

[ChrisOlson]

The performance of your alternator is definitely impressive... it's similar to what I'm getting, definitely a bit smaller in size though.  Some of the difference could be to do with air gap and battery voltage - in my tests I tried to keep my battery down to about 50 Volts

Dan,

We might even be using different testing methods.  When I did this I hooked up the rectifier and two batteries and put a Doc Wattson inline between the rectifier negatives and the battery negative.  Then I hooked up my Sun AVR to apply a load to the batteries to keep the voltage at 29.  I took the readings off the Doc Wattson, even though the actual load on the AVR was a bit less than what Doc Wattson said was going into the batteries.

I think the Doc Wattson amps were between 53 and 54 and the battery amps going to the AVR carbon pile was around 50 most of the time.  And I had to constantly adjust the knob a bit on the carbon pile to keep the voltage at my 29 VDC target.

If I drop the voltage down in the low range (24 volts) I don't know what it would put out because I didn't test that.  All I know for sure is that 53-54 amps with AWG 14 in delta didn't appear to be too much for it because it only got to 49.7 degrees C on my tester after 10 minutes (the carbon pile got hot - otherwise I would've tested it longer).
--
Chris

[ChrisOlson]

Often times in that machine, there are the edges of two poles over the same leg of a coil at the same time and it really doesn't show up as the problem I would've expected.

I think (but don't know for sure) that this can't be.  I try to think about electrical flow in terms of hydraulics - volts (pressure), amps (flow), ohms (resistance).

If you think about this, you have one generator pole over one side of the coil leg and another pole over the other side of the same leg.  The coil has to be excited two directions at once.  The windings on the inside of one leg are excited and current flows (tries) one direction.  The windings on the outside of that same leg are excited by an opposite pole and current flows (tries) the other direction.  That current flow has to meet somewhere in the middle of the coil and cancel.  Nothing moves but it piles up electrons under pressure that would move if the pressure was different on each end.

Now think about this in terms of hydraulics.  It's a little different because you can't "excite" a hydraulic tube in the middle and get oil to flow out the end.  But if you take a hydrauilc tube and apply pressure to both ends of it, nothing will flow if the pressure is equal.  But it robs pump power to create that pressure in the tube.  This is why closed center hydraulics, once thought to be the Hot Tip, have been replaced by PFC in more modern systems.

And the other thing that makes me think it has to be is that if you bring a generator up to rated speed with no load on it and let it run that way for an extended period of time - the windings in it eventually warm up a bit - usually enough to feel it by touch.  Even though there's no current coming out of it, things are being excited in there and it creates heat.  My Winco tractor drive generator - single phase, 12 kW, self-exciting, brushless - gets warm enough after only 10 minutes at rated speed to melt snow off it - with no cord hooked up to it.

I have to believe this same thing happens in a coil.  Any time you pass the fliux contained in the rotors/magnets over wires that cut it and be excited, it has to rob a bit of power because the electrical pressure in that coil increases even though nothing flows and it cancels.  You simply cannot cause an action without having an equal and opposite reaction someplace - usually against whatever caused the action.  If people are going to insist that delta circulating currents rob power, then I'm going to insist that circulating currents in a single coil, caused by uneven pressure as the two poles pass over one leg, rob power.  And also that circulating current in that coil caused by ONE pole passing over BOTH legs of the coil rob power.  In the dead middle where the pressure is equal nothing much probably happens.  But as it passes over and the distribution of flux changes and causes current to flow back and forth against changing electrical pressure in that coil it's taking power to move those electrons on the surface of that wire.

My gut feeling, based on what I've learned and tested myself, is that if you built two generators - one with "perfect" geometry and the other with stuff overlapping like this - and tested them side by side with whatever device you want to use to accurately measure shaft input power, and whatever you wanted to use to measure what you get out of the generator, the one with "perfect" geometry would win.  It might be very small, I don't know.  But my gut feeling is that it's larger than you suspect.
--
Chris

[dciolek]

My one concern that I just now noticed... if your brake rotors are 12" in diameter, I wonder what the bolt pattern is?  For 12" rotors I normally use a 4" diameter bolt pattern and there is very little room left on the inside (like 1/4") after I wind the coils.  If your bolt pattern is larger in diameter, you may not have room for the copper where it really should be (that copper at the bottoms of the coils, in between the magnets, and the studs which hold the whole thing together).

I just ran into that question yesterday when I did the CAD for the Stator (did the rotor first). Because I am using a car rotor -- I have a raised lip from the geometry of the rotor rather than a flat steel plate -- so it actually isn't the bolt  pattern that limits me in coil leg width.  It is the raised section that would typically show the studs to the wheel and have the lug nuts cranked down against it.  As I measured it out -- the coil leg width that keeps me from passing two poles at the same time is pretty close to the same width that would allow me to just barely miss the raised area of the rotor.  I guess what this means is that if I need more turns -- I have no place else to go but towards the center rather than towards the outside, regardless if I am OK with the "crossing two poles" question.  That means I am crossing two sides of the same coil at the same time -- so if there is a negative to that, it will show up if I can't fit the turns in that I need in that fixed width.  My design radius for the center of the 2" round magnets is 4 7/8" until I actually get the magnets and see what extra space that gives them on the outside.  I might adjust when I place them day 1.  You see, these are old rotors off a car, not steel plate that I control the diameter.  So there is rust off those rotors has taken it below 12" and I will probably want to mount them slightly inside that outside edge (right?).  But I have room on the inner leg of the coil based on my last calculation if I place the magnets at least 4 7/8" radius (which makes the outside edge at 5 7/8" radius).  Maybe I have room to mount them further out, or even overlap the outside edge -- but I didn't think that was a good idea based on having some steel to keep the flux lines happy.

What would happen if I mounted the center of the magnets at 5 7/8" radius and adjusted my stator appropriately?  That gives me more coil room, but leave half the magnet without any steel behind it.  Is that worse than crowding things?

I think what you're saying here, Dan, is that those 2" disc magnets are simply too big for this size generator and blades.  There's a LOT of magnet area there - 3.14 square inches, compared to 2 for bars and 2.35 with wedges.  What if dciolek would build a different configuration?  Like a 8 pole 6 coil?

Cool, I had just been thinking of that when the 14" rotor deal came up saying I was crowding the magnets otherwise.  Maybe I am better with the fewer poles and coils.  I don't think I'm going there though.  First of all, I already ordered the 24 magnets, so I'd hate to have 8 of them laying around doing nothing.  So I think the plan is to ride out this design with 12/9 and see what happens.  If I don't like it -- I unpin the magnets and start with a fresh rotor machined from the custom steel plate that everyone uses instead of the ones off my Chevy Impala SS (which is pretty peppy, but not even close to the HP you guys were talking about earlier).

[ChrisOlson]

What would happen if I mounted the center of the magnets at 5 7/8" radius and adjusted my stator appropriately?  That gives me more coil room, but leave half the magnet without any steel behind it.  Is that worse than crowding things?

That would be not recommended.  I think it would work but be even a bigger example of what you CAN do and still have it work   

I tell you what - your space there is WAY too tight.  IIRC, the Chevy bolt pattern is either 5 x 5 or 5 x 4.75 (depending on the year) and it's just not going to fit.  You're going to need two 14" x 1/4" steel discs to get the hole in the stator big enough to clear that big bolt pattern.  What I would recommend is that if you could find a welding shop there locally, have them plaz you out a couple discs.
--
Chris

[dciolek]

IIRC, the Chevy bolt pattern is either 5 x 5 or 5 x 4.75 (depending on the year) and it's just not going to fit.

Its probably the 5 x 4.75, because I measured it off at 4 5/8" (sounds like my measurement error to centerline).  Along these lines though, the bigger problem is not the bolt pattern -- it is the fact that I was mounting these rotors back to back (not the flat face, but the raised face) and bolting them together with the lug nuts to the hub.  This back to back created the spacing between rotors and it seemed to work out just right for the airgap to work with the 1/2" high disc magnets (raised flag - no adjustable airgap if I need it tighter, only further apart is possible).  That being said, I actually have more to clear with the stator than just the bolt pattern.  When I first laid it out in CAD, it seemed that the coil leg width that I had available to keep from overlapping poles was about what I had available on the inside clearance of the stator.  I'll have to double check that I guess.  On the other hand, it looks like a 14" rotor would have slightly more than an inch of coil leg width (vs. .600) without overlapping opposite poles or the magnet space itself on the inside of the coil.  I didn't think about room for connecting leads and wires and stuff -- those don't show up on CAD


Or more likely -- I will be using this go around to experiment only and end up buying the larger rotors before I start thinking about using the magnets with blades in some permanent setup.  I was going to test this monster out attached to a spinning bike (heavy fly wheel stationary) first and worry about blades once I cut my teeth.  I realized that would take a different stator (lower RPM required for cut-in on bike) and produce less power (lower diameter wire necessary), so I figured the coil space available might work for something like that.  I always knew I'd remake the stator down the line -- but didn't figure I'd also redo the rotors.  Either way -- its an experiment only and not a production model.  If I had started the coil winding and figured out I didn't have room -- I would have come to the same conclusion that I need to move to a large diameter rotor.  You've saved me an experimental step here -- because I had nothing reliably mathematical to go on to estimate how many turns will produce the voltage I need in the coil leg space I have available at the RPMs expected from the blades this would have eventually been matched with.  I had some designs mapped out for a 12.5' diameter set of blades that got carved from a 2 x 10 -- but that's another adventure I thought I'd leave for the future.

[ChrisOlson]

Well, if this is a 12 volt 8 foot so you only have to wind like 30-33 turns with one strand of wire it still might fit if you make the coil interconnects on the outside of the coils instead of inside.  But if the plan is to use 12 foot blades this is getting more hairy than I thought.
--
Chris

[DanB]

if I was going to use those rotors, I'd not run them the way you have them pictured, there's no room for copper left.  The tops of the 'hats' should be facing out - and then it might work but it would be tight.

When you consider all the mucking around you'll have to do/compromises to make those work, I'd be inclined to get some rotors cut from steel and may as well make them a bit larger.

[dciolek]

Well, I'm convinced!  My original idea to try to come up with a "find it in the junkyard" project using the rotors/lug nuts/hub/half-CV shaft/knuckle (maybe)/strut (maybe) of a car to make this work doesn't seem to have much promise in the real world of producing an "effective" turbine.  Project abandoned in favor of something that will work better with the magnets and wire I obviously purchased too soon before I knew the whole plan.  I personally find that sometimes that kind of leap is necessary to get me out of design forever/implement never stage.

if I was going to use those rotors, I'd not run them the way you have them pictured, there's no room for copper left.  The tops of the 'hats' should be facing out - and then it might work but it would be tight.

The only reason I ran them back to back was so I could use the studs and lug nuts that were already on the wheel hub.  Flipping them around, the studs aren't long enough to engage both rotors, and I would have to replace with something I'm not going to find on that hypothetical "car in the junkyard".  The axle haft CV shaft and axle nut was going to couple what was going to be the structural base for the blades to this monstrosity.  That's where I ran into my first "must-weld" situation.  I couldn't find an easy way to attach blades to the half shaft without welding a six inch or so steel plate to the shaft.  



You should also know that I was also trying to be as self sufficient as possible in case there was a situation where I didn't have a local plasma cutter, welder, lathe or stuff like that available to me (for whatever reason).  With all of the feedback I have received, I seems too challenging to continue pursuing with those limitations and restrictions.  It was fun while it lasted though and brought out a heck of a discussion and lots of great information.  Off to find a local shop to work with...

Thanks!

[DanB]

You can use junkyard bits, I made lots of turbines that way a few years ago, they worked fine and most of them are still flying.  Here's a page about one:
http://www.otherpower.com/timsturbine.html

But my thinking lately for most folks is... you're going to invest a lot in time, tower, batteries, power electronics etc etc so don't make *too many* compromises on whatever sits at that tower top!  And you don't need to goto the machine shop necessarily.  If you have a cutting torch and a drill press you can probably make all these bits yourself.  I've seen folks cut out nice steel rotors out of thick steel with a saber saw and a pile of metal blades (ear protection and patience required)

[dciolek]

Hmmm...

A cutting torch sounds fun.  I almost decided to break down and learn how to weld too.  My wife and kids already think I went off the deep end playing with this stuff.  I can only imagine what they would say if they pulled into the garage to see me in a welder's mask with sparks flying.

[DanB]

A cutting torch is a really nice thing... lots more fun than a new big LCD TV and just about the same price.  You can have loads of fun with it and actually do a couple of useful things!  (if you learn to use it)

[ChrisOlson]

A cutting torch is a really nice thing... lots more fun than a new big LCD TV and just about the same price.  You can have loads of fun with it and actually do a couple of useful things!  (if you learn to use it)

I agree.  The two most important tools in the shop are the fire wrench and perma-bolter.  You can have CNC mills and lathes and all sorts of other fancy equipment, but if you got no torch and welder you're pretty much screwed.
--
Chris

[scoraigwind]

Often times in that machine, there are the edges of two poles over the same leg of a coil at the same time and it really doesn't show up as the problem I would've expected.

I think (but don't know for sure) that this can't be.  I try to think about electrical flow in terms of hydraulics - volts (pressure), amps (flow), ohms (resistance).

If you think about this, you have one generator pole over one side of the coil leg and another pole over the other side of the same leg.  The coil has to be excited two directions at once.  The windings on the inside of one leg are excited and current flows (tries) one direction.  The windings on the outside of that same leg are excited by an opposite pole and current flows (tries) the other direction.  That current flow has to meet somewhere in the middle of the coil and cancel.  Nothing moves but it piles up electrons under pressure that would move if the pressure was different on each end.
--
Chris

I can't let this go unchallenged.  I have just read all the way though this thread and it has been fascinating stuff.  But there does seem to be a persistent misapprehension on Chris's part about the role of flux linkage in the process of inducing currents in wires.

The way it works can be analysed in various ways including looking at the various angles of various wires and their positions as they move, but the simplest way (which must ultimately give exactly the same answer) is to look at the changes in flux linkage - the rate of change of flux linkage - in the coil.  This is what Dan calls thinking of it like a transformer.  And that's where the end windings that appeared to be doing nothing actually show their value, because they bring a larger area of flux into the coil and increase the flux linkage.  And that's where my smaller stators with the coils 'the wrong way around' actually work very well to stay clear of the moving parts in the centre and still get the best flux linkage.

But I digress.  The main point I have to make here is that the voltage produced by a coil depends on the net rate of flux linkage (added from all of the turns in series).  If the coil is open circuit there is no current (try as it might) and there are no electrons piling up anywhere.  When the switch closes or the voltage gets high enough to make the diodes conduct then a current will flow around the entire circuit.  There will not be more current here than there because there is no significant capacitance, so the charge can't 'pile up'.  This will not cause parasitic losses nor extra vibration.  Sorry, I just don't buy that stuff.

Personally I try to design alternators so that the full flux of the magnet passes through even the innermost turns, and I do this by making the rotor big enough where possible.  But I accept that you can cut some corners with those inner turns and still contribute a bit of extra voltage, and the resistance of those extra inner turns is small. I have done this myself when using very long magnet blocks.   In such cases where there is 'cancellation' of the flux you will get a small loss in the overall voltage, but it is not the case that currents will start flowing hither and thither along the same wire.  That's not going to happen.

I just had to set the record straight on that score.

You can achieve radically higher power handling with a Leece-Neville 200 amp brushless truck alternator by using a laminated core (that is a pig to start in low winds and saps all the power at that level) and running it super fast, but that type of design is not going to be suitable for a wind turbine.  Holding up engineering excellence in design in one area and arguing that the axial flux alternator is therefore rubbish in another area is illogical and misleading.  I am more inclined to say that we are talking about small changes in performance here that don't honestly make a lot of difference, whereas what is important is the relative cost of one magnet compared to another and to the cost of wire, and finally above all, what matches the blades to the load as Flux pointed out right at the start.

[ChrisOlson]

I can't let this go unchallenged.  I have just read all the way though this thread and it has been fascinating stuff.  But there does seem to be a persistent misapprehension on Chris's part about the role of flux linkage in the process of inducing currents in wires.

I may misapprehend a lot of things but I don't misapprehend what I've measured.  Just yesterday I was running various stators with my hydraulic motor and comparing them to various iron core rings to measure eddy losses in a new iron core generator I'm building.  I was using 10" rotors with bar mags and in arriving at some benchmark figures for this new generator I ran two different stators - one with considerable leg overlap on the insides of the coils (the coils come to a "vee" with a single pin used in the bottom of the coil winder) and shorter coil legs so the magnets "wipe" the end windings of the coil, and the one in the photo.  They are otherwise wound identical with the same number of turns, both wired delta.  Referring to my notes, it took 19 more lb-inches of input torque (~74 watts) to drive the one with lots of leg overlap @ 330 rpm with 24 watts less output than the one in the photo.  Absolutely nothing different in those two stators except the coil shape.

On the turbine you'd get 98 more watts out of the one I showed in the photo.

If you can show me some concrete measurements on this indicating different, then I'll go for it.  Otherwise I'm going with what I've learned thru playing with these things and will continue with my misapprehensions.
--
Chris

[DanG]

Referring to my notes, it took 19 more lb-inches of input torque (~74 watts) to drive the one...

Those power figures seem skewed.

1 horsepower is defined as 550 ft-lbf/s (33,000 ft-lbf/min) and 1 Horsepower is defined as 745.7 watts

74 Watts to inch-pounds: 74w equals 0.09924 horsepower equals 54.582 ft/lb or 654.984 in/lb...

So given 19 Inch-pounds, to watts:  19 in-lb equals 1.58 ft-lb equals 0.0029 horsepower equals 2.14 watts

[ChrisOlson]

So given 19 Inch-pounds, to watts:  19 in-lb equals 1.58 ft-lb equals 0.0029 horsepower equals 2.14 watts

Mechanical (brake) hp is given by:
torque x rpm /5252

1.58 lb-ft x 330 rpm / 5252 = 0.0993 hp, multiplied by 746 watts per hp = 74 watts for a quick dirty figure.

HP is work over time and you're not considering the shaft speed in your figure.
--
Chris

[scoraigwind]

I can't let this go unchallenged.  I have just read all the way though this thread and it has been fascinating stuff.  But there does seem to be a persistent misapprehension on Chris's part about the role of flux linkage in the process of inducing currents in wires.

I may misapprehend a lot of things but I don't misapprehend what I've measured.  Just yesterday I was running various stators with my hydraulic motor and comparing them to various iron core rings to measure eddy losses in a new iron core generator I'm building.  I was using 10" rotors with bar mags and in arriving at some benchmark figures for this new generator I ran two different stators - one with considerable leg overlap on the insides of the coils (the coils come to a "vee" with a single pin used in the bottom of the coil winder) and shorter coil legs so the magnets "wipe" the end windings of the coil, and the one in the photo.  They are otherwise wound identical with the same number of turns, both wired delta.  Referring to my notes, it took 19 more lb-inches of input torque (~74 watts) to drive the one with lots of leg overlap @ 330 rpm with 24 watts less output than the one in the photo.  Absolutely nothing different in those two stators except the coil shape.

On the turbine you'd get 98 more watts out of the one I showed in the photo.

If you can show me some concrete measurements on this indicating different, then I'll go for it.  Otherwise I'm going with what I've learned thru playing with these things and will continue with my misapprehensions.
--
Chris

My brain is only small, and I have by now lost track of all of the different coil shapes you refer to.  and when you start on horse-power I really do start to get perplexed because here in the UK we discovered some years back that mechanical power can also be measured in Watts.  But I shall retreat into the theory and let you do the bench test stuff.

What I was trying to say was that there is no chance of parasitic currents in a disconnected winding, arising from sub-optimal layout of coils and magnets.  But now I find that you are connecting the coils in delta...  well there can be a lot of parasitic currents there.  I always use series-star (wye) connection so as to avoid any parasitic currents in the winding sapping the power at low speeds.  Delta windings can be handy as an option for more power at higher speeds (only one third the resistance, so much lower losses on high power) but in my opinion its a mistake to use delta windings for low winds because the parasitic losses are always higher than a star connection.

Whatever the details, I would still staunchly contend that you will never get two currents in opposite directions in the same coil due to 'cancellation' of magnetic poles.  The currents in the coils are driven by the overall net voltage induced in the coils, although often in delta connections if the waveform is not sinusoidal and so you can get all sorts of current that you neither need nor want wasting power before you even start to produce any output.  But if you are messing around with cores in the coils I don't suppose that parasitic losses worry you anyway.

Clearly a coil that does not link all of the flux will produce a lower voltage than one which does link all of the flux, and the reason for creating the first type of coil is that it can have lower resistance which may turn out to pay off later (or not) so it is not a simple matter to say which of the two is going to be overall 'better' but I still contend most adamantly that the coil shape will not produce parasitic eddy currents in a star connected winding.

[ChrisOlson]

but I still contend most adamantly that the coil shape will not produce parasitic eddy currents in a star connected winding.

Well, I have also run those generators and measured input torque, unloaded, in both wye and delta configuration to measure these "parasitic losses" in delta and that was a wash.  Absolutely not one bit of difference unloaded.  In delta, hooked up and charging batteries, they are more efficient (measuring power in and power out) from cut-in all the way to rated power, and the gap widens the faster they turn.

The reason I have been testing them in delta recently is because the new iron core generator I am building will be a delta unit.  I wanted some benchmark figures to compare because the new radial flux iron core unit I'm putting together will have the same continuous power rating, based on wire size.  When I get it built I'm going test them for efficiency and operating temperature to see how they compare, the axial vs the radial, the axial using coils cast in fiberglass resin, the radial with open coils wound on a powdered iron core.

Stay tuned.  You haven't seen all of what I'm up to yet   
--
Chris

[DanB]

Chris - you may have run these tests and I think tests are good...
but I've contended as have others that there could not possibly such currents in star, no matter how lousy the coils might seem.

I think your work is quite good.  My point, and Hugh's at this point is that... there cannot be parasitic losses in the coils when wired in series / star - which is really the statement you made that got me heavily into this discussion to begin with.

WilliB putting up 'rules' about how it 'must be' and you're suggesting that if we get it wrong it will cause drag and heat (even before cut in).  When I see something posted here *as fact* that I really feel is *dead wrong* then I must contest it (even if I find out later that I'm wrong!.. in which case... I'll confess)

But when folks are posting questions who are new, and perhaps don't understand... I'm quite keen to see suggestions about 'better design' or how to get a bit more for the buck or whatever.  But when false statements are made and suggested to be fact (and I've done my share of that too so...)... and I don't believe they are fact, then...

My honest opinion here...
You're slightly off in your understanding of how these things work and I've posted plenty about that already in this thread.  (Im probably a bit off too)

You're doing fine work and I will be staying tuned and very much looking forward to the program.  I expect you're building fine alternators... probably much finer than I am because for you.. attention to detail is built in and for me it's always a challenge.

You maybe taking steps backwards with iron cores, springs in the furling system, fiberglass blades made in china and wiring stators in delta.  But time will tell... 

[ChrisOlson]

You maybe taking steps backwards with iron cores, springs in the furling system, fiberglass blades made in china and wiring stators in delta.  But time will tell...  

I may be, according to popular consensus.  But then I've found in the past that popular consensus is not synonymous with wisdom and is most times arrived at thru compromise.  So I throw popular consensus out the window and look at how one of the most successful small wind turbines of all recorded time does it - the Jacobs.  The newest ones, evolved over almost a century from the 31-20 down to the little 23-10, have a delta-wired Winco generator with iron core, springs in the furling system, and fiberglass blades.  If they can do it, so can I.  I'm guessing that they know some things that compromise missed along the way.

Oh - and those Jakes also rotate left and normally furl left (although they can bend to the right too) ---- which is also wrong according to popular consensus.  LOL!
--
Chris

[DanB]

I must be behind the times.
I didn't know that there were any Jakes out there using anything other than wooden blades and a variable pitch system (they don't furl at least not the ones I've seen)

The tails can only be folded up to shut down the machine with a winch at the bottom of the tower.

They are nice things though...  very neat machines.

But you're wrong about one other thing.  Those are not the 'very best' small wind turbines ever...  mine are. ;-)
(just kidding again...  )

[ChrisOlson]

I must be behind the times.
I didn't know that there were any Jakes out there using anything other than wooden blades and a variable pitch system (they don't furl at least not the ones I've seen)

Nope - all the new ones have fiberglass blades, a Winco three-phase generator (they all use the same gen, just de-rated with smaller blades on the smaller models), and they furl left due to torque on the gearbox.  the prop governor starts feathering the blades at 25-30 mph with the gen loaded, and about 15-20 mph if it's unloaded.  At about 40-45 mph the machine folds up and furls left out of the wind.

Every once in awhile when I'm feeling deprived I go over to the factory just across the border in Prior Lake, MN and drool over them.  Steve (the GM) and all the guys there at Wind Turbine Industries Corp (the company who bought the manufacturing rights to the Jake) know me pretty good.  I've tried to convince the guys there to sort of get me the pieces like Johnny Cash and his "One Piece At A Time" Cadillac.  But they claim sneaking a 25 kVA Winco generator out in a lunch box is a bit of a challenge.  I'm putting all my pennies in the piggy bank and when I save up enough to drop 118 Grand on a new 31-20 I'm going own one.
--
Chris

Edit:
The Jake actually does furl left before the 40-45 mph point due to torque on the hypoid gearbox.  But once the loading on the blades exceeds the tail's ability to counteract torque, they fold up and shut down.  The Jake's governor is centrifugal so they don't fully feather the blades and just stop.  If the rotor slows down the blades flatten out and the turbine goes back to work.  Someplace, if I dig thru all my videos and stuff I got a movie that I took of a 31-20 folding itself up in very high winds with tree branches and pieces of buildings flying by the blades.

[scoraigwind]

You maybe taking steps backwards with iron cores, springs in the furling system, fiberglass blades made in china and wiring stators in delta.  But time will tell...  

I may be, according to popular consensus.  But then I've found in the past that popular consensus is not synonymous with wisdom and is most times arrived at thru compromise.  So I throw popular consensus out the window and look at how one of the most successful small wind turbines of all recorded time does it - the Jacobs.  The newest ones, evolved over almost a century from the 31-20 down to the little 23-10, have a delta-wired Winco generator with iron core, springs in the furling system, and fiberglass blades.  If they can do it, so can I.  I'm guessing that they know some things that compromise missed along the way.

Oh - and those Jakes also rotate left and normally furl left (although they can bend to the right too) ---- which is also wrong according to popular consensus.  LOL!
--
Chris

It's great to have people like you trying stuff out and reporting to the world Chris, and preventing us getting bored and complacent.  I hope you do not mind when we 'consensus' folk drone on about the best way to do stuff.  My only concern is that homebrew people might get mislead by some of the stuff you are stating such as axial flux alternators never being good for more than 35-40% duty cycle and the weird idea that you seemed to have of parasitic currents in star/wye connected stators.  A good debate helps to bring out and clarify the reasons for our respective views.  Even if it ends up that we have to agree to disagree.

I would never claim that wooden blades are a more suitable material for a manufacturer in the USA or in China compared to wood, but I do suggest that wood is a much more suitable material for homebrew for numerous reasons - it's nicer to work with, it has better fatigue resistance, you don't need to make moulds etc etc...

As for springs, I have seen several people move from using springs to using gravity but never the other direction.  In most cases where a manufacturer uses a spring in a turbine it has been a troublespot - for example the Proven turbines still have problems with the blade springs after nearly twenty years, and the old windseekers, now called Lakotas? suffering spring failures on the tilt-up system.  Anyway it's 'just a materials problem' as the VAWT people would say.  But gravity is free of charge and very reliable.

Star and delta - what can I say?  You can always achieve the same voltage output for the same internal resistance using star rather than delta. (By adjusting the wire size and number of turns to do that.)  In star there is no possibility of parasitic currents arising within the windings (no internal circuits for them).  I and others have measured the mechanical power drain from the parasitic currents in open circuit delta alternators.  We therefore choose to avoid using delta windings.  Your experience has obviously been different.  The world is a strange place.

Iron cores work fine if you don't mind the losses in low winds and you are ready to fabricate them.  I will watch with interest to see how you get on with your iron filings.  Cores can save on magnet material, but magnets have become pretty cheap so I would still recommend to a home-brewer who wants to find the easy road to a machine that works in light winds that he/she uses an axial flux stator without cores.  I used to work with laminated cores in the old AWP turbine that I designed (and there are still ten or more around here working) and it was a breath of fresh air to escape from those cores:  The drag at low speeds, the limited maximum current that prevented effective braking in high winds, the hassle of making them...  Let's say there are pros and cons.

As for which way you furl in relation to the blade rotation - that is a bit of a refinement.  I did it wrong for thirty years without noticing any problem.  If the Jacobs yaws out of the wind slowly (as I suspect) due to excess torque rather than rapidly in response to a gust hitting the blades frontally then it makes no difference, and in any case its best to make sure you have plenty of clearance at the blade tips so it should never be an issue in the first place.

There's more than one way to build a good turbine to be sure, but some of the easier ways are already quite clearly documented and the 'problems' that you find with the 'consensus' ways to do it are not always apparent to me.  Hence my occasional impatience with the solutions.