Ebay hoverboard motor 30$ 10Turn, 5 in hand to 50 Turn alt 5W at 12V 40 rpm

[oneirondreamer]

Hi Guys,

It looks like the scourge of scooters and hover boards are going to supply us with wind turbine alternators for years!   I don't advocate theft, but if I lived in a city having problems with them, I'd be sorely tempted to take some home in a faraday bag, as I'm pretty sure they mostly have similar wheelmotors...      I'll add more to the documentation eventually, but I want to get my hydro rotor tested ASAP, so it will be later.   People in the know that want to try it probably don't need much more instruction that what I've got in the title.    Some things I would do differently, and that I'd like advice on are.   The original configuration had all 3 tails to the star point quite different lengths.   The difference would have been as much as 20mm, almost 1/2 a turn in the winding.   I reconnected in series, and was grateful at the time to have as much length in the wires as I had, but now I realize I should have kept them all the same I think.   I think I've gone from a 3 phase alternator, accidentally, to a 9 phase....   You can see the readouts from my data logging system in the pdf's on the instructables site.



In short what I've done is taken the five in hand (parallel) windings from each phase and run them in series.    I think for what I want the ideal winding for this for me would be a 2 in hand 20-25T with a bit thicker wire, and maybe delta.    I will reconnect another, this time trying to keep connection distance the same as phase to phase difference, and set it up for star and delta, from the outside if possible, could put relay in stator rather than leading out.


https://www.instructables.com/id/Micro-Kinetic-Hydro-Power-System/

[Adriaan Kragten]

I have watched the nice video which you have made about this hover board motor. In the video it is shown that the stator has 27 coils and that the armature which turns around the stator has 30 poles. This is an unusual combination but it must have been done so to minimize the fluctuation of the sticking torque. I have described several PM-generators which also have a small difference in the number of armature poles and the number of stator poles. Public report KD 632 describes a 26-pole generator for which a stator is used with 24 slots. Twelve coils are laid in these 24 slots, so four coils per phase. Figure 1 of KD 632 gives the orientation of the coils and you can see that if coils U1 and U2 are about opposite to the south poles S1 and S13, the coils U3 and U4 are about opposite to the north poles N7 and N8. So the winding direction of the set of coils U1 + U2 must be opposite to the winding direction of the set of coils U3 + U4.

I think that a similar orientation of the coils is used for the hover board motor. A difference is that for my generator, the number of coils is half the number of stator slots but that for the hover board motor, the number of coils is the same as the number of stator slots. I have thought how the winding of the hover board motor may look like.

I think that the winding is a 3-phase winding and so there are nine coils of phase U, nine coils of phase V and nine coils of phase W. I think that the nine coils of one phase are divided into three bundles of three coils and that the sequence of the coils is: U1, U2, U3, V1, V2, V3, W1, W2, W3, U4, U5, U6, V4, V5, V6, W4, W5, W6, U7, U8, U9, V7, V8, V9, W7, W8, W9.

The armature has 30 poles so 15 north poles and 15 south poles. Assume that a north pole is opposite to coil U2. A south pole will then be about opposite to coils U1 and U2. So the winding direction of coils U1 and U3 must be opposite to the winding direction of coil U2. There will be an angle of 120° in between coils U2, U5 and U8. So if a north pole is opposite to coil U2, there will also be a north pole opposite to coil U5 and to coil U8. So the voltage generated in coils U2, U5 and U8 will be in phase. It must be possible to prove that the phase angle in between the currents generated in coils U2, V2 and W2 is 120° and that so a 3-phase current will be generated.

Assume that the north poles are labelled N1 - N15 and that the south poles are labelled S1 - S15. Assume that the armature has the position such that N1 is just opposite to coil U2. The armature pole angle is 360 / 30 = 10°. The stator pole angle is 360 / 27 = 13.333°. So the difference is 3.333°. This means that if N1 is just opposite U2, S15 will be about opposite to U1 and S2 will be about opposite to U3 but there will be a shift of 3.333°. You get the same magnet situation if the armature has rotated 30°. So an armature angle of 30° corresponds to a phase angle of 360°. So an armature angle of 3.333° corresponds to a phase angle of 40°. This means that there is a phase angle of 40° in between the voltages generated in coils U1, U2 and U3. Adding of sinusoidal voltages which are out of phase but which have the same frequency, results in a total voltage which is also sinusoidal. If it is assumed that the maximum voltage of one coil is 1, it can be proven that the total voltage of the three coils U1, U2 and U3 is given by: Utot = sin 50° + sin 90° + sin 130° = 2.532. If the voltage generated in the three coils would be in phase, the total voltage would be 3. So the voltage is reduced by a factor 2.532 / 3 = 0.844 because the voltages are not in phase to each other. This voltage reduction is acceptable for the given coil configuration. As there is a phase shift in the voltage generated in U1, U2 and U3, these coils must be connected in series. Parallel connection of these coils would result in internal circulating currents.

So I think that the three coils of one coil bundle are connected in series. The three coil bundles of one phase can also be connected in series and this gives the highest voltage at a certain rotational speed if the winding is rectified in star. As there is no phase shift in between the voltage generated in each of the three coil bundles of one phase, the three coil bundles can also be connected in parallel. If the three coil bundles of one phase are connected in parallel, the voltage becomes a factor 3 lower but the current becomes a factor 3 higher.

I am curious if my estimation of the coil orientation is correct. As these motors are rather cheap and as the peak on the sticking torque will be rather low, it might be possible to use this motor as a windmill generator if the shaft and the bearings are strong enough. Have you measured the peak on the sticking torque at stand still position?

[oneirondreamer]

Hello Adriaan,

Thank you for your interest, I had hoped you would find these interesting.     The silly scooter craze is going to mean a massive surplus of these motors (and even the controllers, some of which have regenerative braking....).  The scooter motors are similar in geometry. 

These are wound in a distributed pattern, of AaABbBCcCAaABbBCcCAaABbBCcC, with the repeating sets of 3x3 coils joined in series (capital for clockwise) for each phase.     The winding is 5 strands in hand of what seems to be about 20G wire, with 10 turns around each pole.    What I've done is connected each of the 5 in hand turns (from each phase) in series, so that instead of 10 turns of 5 sets x3, I've got 50 turns of 1 wire x 3 sets. 

To try to get an understanding of the cogging load, I left the alternator open circuited, and attached the lightest load which would continually turn the alternator, but was not enough mass to start it turning.   This mass was 0.25kg/m.    RPM was around 15 sec/rev by my recollection.  The cogging loads are present, but not strong. 

Here someone has done some interesting FEA on these rotors
https://www.electric-skateboard.builders/t/rewinding-of-hoverboard-motor-looking-for-advices/76373/35

Something curious about them is that there seems to be no steel can, but also no magnetic field on the outside of the hub.   I'm wondering if these magnets are “pulsed” in such a way as to create a halback effect, like they do with the flexible magnetic strip?    The beauty of these things is that there is a a small amount of magnets doing a lot of work.   

[Adriaan Kragten]

Assume that the north poles are labelled N1 - N15 and that the south poles are labelled S1 - S15. Assume that the armature has the position such that N1 is just opposite to coil U2. The armature pole angle is 360 / 30 = 12°. The stator pole angle is 360 / 27 = 13.333°. So the difference is 1.333°. This means that if N1 is just opposite U2, S15 will be about opposite to U1 and S2 will be about opposite to U3 but there will be a shift of 1.333°. You get the same magnet situation if the armature has rotated 30°. So an armature angle of 30° corresponds to a phase angle of 360°. So an armature angle of 1.333° corresponds to a phase angle of 16°. This means that there is a phase angle of 16° in between the voltages generated in coils U1, U2 and U3. Adding of sinusoidal voltages which are out of phase but which have the same frequency, results in a total voltage which is also sinusoidal. If it is assumed that the maximum voltage of one coil is 1, it can be proven that the total voltage of the three coils U1, U2 and U3 is given by: Utot = sin 74° + sin 90° + sin 106° = 2.9225. If the voltage generated in the three coils would be in phase, the total voltage would be 3. So the voltage is reduced by a factor 2.9225 / 3 = 0.974 because the voltages are not in phase to each other. This voltage reduction is very low for the given coil configuration. As there is a phase shift in the voltage generated in U1, U2 and U3, these coils must be connected in series. Parallel connection of these coils would result in internal circulating currents.

I see that I have made a mistake in the armature pole angle in my previous calculation. I have corrected it now but I could not do it in the original post.

[Adriaan Kragten]

Hello Adriaan,

These are wound in a distributed pattern, of AaABbBCcCAaABbBCcCAaABbBCcC, with the repeating sets of 3x3 coils joined in series (capital for clockwise) for each phase.     The winding is 5 strands in hand of what seems to be about 20G wire, with 10 turns around each pole.    What I've done is connected each of the 5 in hand turns (from each phase) in series, so that instead of 10 turns of 5 sets x3, I've got 50 turns of 1 wire x 3 sets. 

To try to get an understanding of the cogging load, I left the alternator open circuited, and attached the lightest load which would continually turn the alternator, but was not enough mass to start it turning.   This mass was 0.25kg/m.    RPM was around 15 sec/rev by my recollection.  The cogging loads are present, but not strong. 

From the pattern you describe I see that I am right how the winding is laid. However, I don't understand your modification. For each phase you have three bundles of three coils. Within a coil bundle you should not change something because there is a phase shift in between the voltage generated in each of the three coils. So the only thing you can do if you want to use the original winding, is to connect the three coil bundles in parallel if they were originally connected in series or to connect them in series if they were originally connected in parallel.

For the starting behaviour of a wind turbine it is important to know the peak on the sticking torque from stand still position so not the torque which is required to make your bucket going down slowly. So you must add more and more weights until the bucket starts moving. This weight in N times the radius of the rope in m gives you the sticking torque in Nm. The required torque to make the bucket going down slowly will be much lower because of the inertia effect of the wheel and the bucket.

Although the generator housing is made of aluminium, there must be an internal iron ring to make that the magnetic flux is guided at the outside of the magnets from a north to a south pole. Without an iron ring, a very high magnetic resistance will be created and this will reduce the maximum torque level dramatically. May be you can't see this iron ring if it is cast in the aluminium housing.

[Adriaan Kragten]

I have read your last post several times and now I think that I understand what you mean. Is it right that the original coil is wound with five wires in parallel and that you have ten turns per coil? So is it right that you have disconnected the soldered ends of the five wires and put all five wires in series? In this case the voltage becomes a factor five higher as now you have 50 turns per coil.

I assume that for the original winding, all nine coils of one phase are connected in series but that the generated voltage for your rather slow running wind turbine is too low with the original winding if you want to charge a 12 V battery. So therefore you want to increase the voltage. But this kind of modification is rather complex and the factor five with which the voltage increases may be too high. It would be nice if the motor could be used as generator without modification of the winding. May be you can use a 6 V battery in stead of a 12 V battery if the voltage is too low. The power which you can generate at a certain rotational speed doesn't increase if you modify the winding. Only the power loss in the rectifier and in the cables of a certain diameter from the generator to the battery increases.

I have asked the Dutch supplier of these motors some questions and one of them was if they have information about the use as generator. May be I buy one myself to test if it can be used for a small horizontal axis wind turbine for 12 V battery charging without modification of the winding.

[oneirondreamer]

I have read your last post several times and now I think that I understand what you mean. Is it right that the original coil is wound with five wires in parallel and that you have ten turns per coil? So is it right that you have disconnected the soldered ends of the five wires and put all five wires in series? In this case the voltage becomes a factor five higher as now you have 50 turns per coil.

Yes, this is correct.   Yes, this is to increase the voltage output, and does not increase total power available, actually probably reduces it quite a bit because of the increased resistance, and very low current handling capability. 

It is now a bit higher voltage than what I would like, and I think that if I do a more careful job of the reconnecting (making sure to keep connections all the same length, and as close to the same length as the connection that jumps from AaABbBCcC group to group) than it will be even higher voltage.   What I hoped for was to have 6 in hand, so I could have had either all six in series or two groups of 3 or 3 of 2, to be able to tune Volts/rpm out.  I'd like to rewind one 6 in hand.

What I think I'll do for now is on my next one, the one to go on my hydro turbine, I'll do the conection more carefully, and I'll connect in Delta, to lower Volts, and give a higher current capacity.   

For this hydro turbine, it will need to be very capable of taking high torque at well under 100 rpm to battery charging voltage, but it will never fear the windy day where velocity is 1000X an average day.

[Adriaan Kragten]

If modification of the winding by connecting the five wires of one coil is series gives you a too high voltage, you can reduce the voltage by a factor 3 if you connect the three coil bundles of one phase in parallel in stead of in series. So compared to the original winding, the voltage is now increased by a factor 5/3. But now the whole modification becomes even more complex. It is a pity that one has used five wires in parallel for the original winding. I one should have used four wires in parallel, you could have doubled the voltage by connecting two wires in parallel and two bundles of two wires in series.

In my previous calculations I have shown that there is a difference in armature pole angle and stator pole angle of 1.333°. This means that the number of preference positions per revolution is 360 / 1.333 = 270 if it is assumed that a preference position is created when a magnet is exactly opposite to a stator pole. So the preference positions should lay very close to each other. Can you feel this if you turn the armature? This calculation of the number of preference positions per revolution is only true if the magnet poles are spread around the armature at a very accurate pole angle of 12° and if the stator poles are also made very accurately. I have made a 22-pole generator in combination with an Indian stator with 24 slots and theoretically it should have 264 preference positions per revolution but in practice I felt 22 preference positions per revolution. This might also be the result of four outside grooves in the used stator stamping.

When the motor is used in the hover board, it uses a certain nominal battery voltage. Do you know this voltage? This DC voltage is transferred into a 3-phase AC voltage with varying frequency by a special inverter. This is a rather complex device and the required electronics will make a substantial part of the price of a complete hover board. But I am still surprised by the rather low costs of a single motor.

If you want to use this hover board motor as wind turbine generator, it must be possible to connect the rotor to the generator armature. The armature is normally provided with a rubber tire. If you remove the tire, is there any flat plane with threaded holes to which the rotor hub can be connected?

[oneirondreamer]

If modification of the winding by connecting the five wires of one coil is series gives you a too high voltage, you can reduce the voltage by a factor 3 if you connect the three coil bundles of one phase in parallel in stead of in series. So compared to the original winding, the voltage is now increased by a factor 5/3. But now the whole modification becomes even more complex. It is a pity that one has used five wires in parallel for the original winding. I one should have used four wires in parallel, you could have doubled the voltage by connecting two wires in parallel and two bundles of two wires in series.

Thank you!, that was exactly my next question.    Yes, I will try this on Stator 2, as it will still be much faster than rewinding, and should be able to leave the Hall sensors in place, in case I want to use a sensored controller.   I will leave the first winding as modified, I've taken one apart (unwound one phase), and I'll mod this 3rd, which leaves me 1 unchanged as a reference, and one with an empty set of slots that I can try an alternative winding on if I'd like.   

What I think I will do, as you suggest I think, is take each AaABbBCcC group AaA, and take it from 10T of 5, to 50T of 1.  Then as you say, I have the option of runing the three of them in each group as parallel, or series, and delta or star.   I think first I will go parallel star.  Makes it quite flexible.   I may be able to 3D print a nice little terminal block to sit inside the thing and give me perfectly consistent wire lengths. 

 I will have to decide on, and stick to a length for the interconnections.   I think the poor waveform from these (under low load) is due to the unequal star connection length.   I think I made it worse by using inconsistent lengths in my 10 to 50T conversion.

Does that make sense to anyone else, or does it seem like I'm making a bigger deal out of it than I should?


I'm glad I got the batch of 4.   Great value for 120+30 shipping CAD$ 

I suspect that all this will not be needed with the right controller.   Some of what I'm reading on RC ESC controllers is suggesting that the regenerative brake features on some of the off road racing involving, technical downhill crawling competition with fairly heavy remote control cars , has the features needed, including adjustable braking curve, and dynamic adjustment as well as data logging and even bluetooth control.   It can also use the built in Hall sensor to adjust the drive for efficiency, and can get take very high torque at low rpm, and charge the RC battery, similar to what I am doing, but with fancy electronics that allow you to leave the higher current capacity in place.

Interesting details,

The Magnet array,  I believed as did Adriaan, that there must be an embeded iron ring/bell hidden.   

I do not think so now, I think this is a quite sophisticated Halbach array.   I've examined a gap between magnets, under some magnification, in a spot where there is a 0.25 mm or so gap, as assembly clearance I guess, or perhaps there was a slightly undersized magnet in this array..  It is clear the whole inside face is machined aluminum.   There is not change in the surface from the outside edge to the inside.  The distance between the inside face and the screws that attach the flange is only 1.3 mm, so any steel ring embedded would have to have been perfectly positioned.   Also, the magnetic viewing film shows a very strong inward facing field, almost nothing facing outward.  Iron isn't attracted at all to the outside of the rotor, and the aluminum case measures to be approximately 12.5mm thick, from inside of magnets, to what used to be the bottom of the inner tube well.

Mounting, no ideal flange, but some possibilities.

The outer bearing plate mounts with 8 x 4mm stainless screws.    This plate could be replaced with a made to suit version, thicker.   Also, If you do as I did, it's quite easy to use the wheel as a lathe, with an RC controller, and cut down one, or both flanges, so you have a smooth drum, or a drum with a single flange, as I have done.   An important detail is that the outer surface of the magnet rotor/rim is flat, the inside, where the wire bead seats, isn't quite flat.  I may machine mine flat though.   It's pretty easy to do with a file while the thing is running.  I'm going to drill the edge of the rim, where the bead used to seat, with 32 holes, and have an interesting structural arrangement planned, so that my 3D printed rotor survives the side pressure.   Too complicated to get into here, now. 

I wonder if you saw this

[Adriaan Kragten]

I have tried the link but it isn't working for me. Strange that the magnets are glued directly in the aluminium housing. Now there is a large air gap at the outside of the magnets and this air gap will reduce the strength of the magnetic flux which is flowing through the stator significantly. Another reduction is caused because the magnets are mounted very close to each other and so there will be a direct flux in between the sides of the magnets. But the resulting flux which is going into the stator lamination might be large enough for the wanted torque level. Here in the Netherlands we can buy motors for two different wheel diameters 6 1/2"and 8". I have asked if both motors are identical but I have got no answer. If both motors are identical, you will get two different maximum driving speeds for a certain battery voltage.

[oneirondreamer]

I have tried the link but it isn't working for me. Strange that the magnets are glued directly in the aluminium housing. Now there is a large air gap at the outside of the magnets and this air gap will reduce the strength of the magnetic flux which is flowing through the stator significantly. Another reduction is caused because the magnets are mounted very close to each other and so there will be a direct flux in between the sides of the magnets.

I can put a piece of iron on the outside of the wheel hub, only 13mm away from the magnets, and there is no attraction.   I have used my magnetic field viewing film, and there appears to be effectively no field going outside the ring.   Either there is an embedded iron ring, or the magnets are specially pulsed and arranged to be in a Halbach array.   I've built a small Halbach array myself, out of  5 , 1/2x1/2x1” block magnets.  It was a pain to glue together but it has effectively 2 poles on one side, with almost no flux anywhere but that one side.  https://en.m.wikipedia.org/wiki/Halbach_array .   

Halbach array's have been used in wheelmotors since the 90's, with solar race cars using them because they allow lower magnet volume/mass to do more work.

In the past it was generally assumed the groups of magnets were required to be assembled into Halbach Array's, however it's my understanding that magnets can now also be pulsed to form blocks for an array like this, and that this is common technology in China.    These blocks would appear on their own to have a pole on one side, and the other pole split into the two small edges,adjoining the neighboring blocks.   The next block would have the same magnetic geometry, but with reversed poles.   

[topspeed]

I paid nothing of my generator...as bicyclist often discard their bikes to repairshops when the control unit brakes down which is pricier. 

[oneirondreamer]

I paid nothing of my generator...as bicyclist often discard their bikes to repairshops when the control unit brakes down which is pricier. 

I'm curious as to it's geometry, do you know the pole and stator tooth count?  What sort of windings?   Many of the bicycle wheel motors seem to have low pole counts compared to these hoverboard motors with 30 magnets in only 105mm circle.

[Adriaan Kragten]

I have tried the link but it isn't working for me. Strange that the magnets are glued directly in the aluminium housing. Now there is a large air gap at the outside of the magnets and this air gap will reduce the strength of the magnetic flux which is flowing through the stator significantly. Another reduction is caused because the magnets are mounted very close to each other and so there will be a direct flux in between the sides of the magnets.

I can put a piece of iron on the outside of the wheel hub, only 13mm away from the magnets, and there is no attraction.   I have used my magnetic field viewing film, and there appears to be effectively no field going outside the ring.   Either there is an embedded iron ring, or the magnets are specially pulsed and arranged to be in a Halbach array.   I've built a small Halbach array myself, out of  5 , 1/2x1/2x1” block magnets.  It was a pain to glue together but it has effectively 2 poles on one side, with almost no flux anywhere but that one side.  https://en.m.wikipedia.org/wiki/Halbach_array .   

I have read the article about the Halbach magnet orientation and this might be an option to prevent an iron outside ring. But this requires special magnets which are magnetized using a special pattern such that the field lines coming out of the inside of the magnet make an inside bend of 90° to the sides of the magnet. So half of the field lines coming out of the inside bends to the left side and half of the field lines bends to the right side. If you would have the same flux density at the inside and at the sides of the magnets, it means that the magnet thickness must be halve the magnet width. But if I look at the photo of your magnets, the magnets are much thinner than halve the width. So I still think that there is an iron ring cast inside the aluminium housing but that the width of this ring is somewhat smaller than the magnet height and that therefore you don't see it. If you break away one magnet you should see it. If there is nothing, it must be the Halbach orientation but in this case, I don't understand why the magnets are that thin. A problem with the Halbach orientation is also that gluing of the magnets in the aluminium housing is rather difficult because every magnet is pushed away by its neighbours.

[Adriaan Kragten]

I paid nothing of my generator...as bicyclist often discard their bikes to repairshops when the control unit brakes down which is pricier. 

I'm curious as to it's geometry, do you know the pole and stator tooth count?  What sort of windings?   Many of the bicycle wheel motors seem to have low pole counts compared to these hoverboard motors with 30 magnets in only 105mm circle.

Some years ago someone lent me a bicycle front wheel motor to see if it can be used as wind turbine generator. I can't remember the exact pole number but is was rather low, may be six or eight. A big problem was the rather large peak on the sticking torque and my conclusion was that this peak was too high to use it as a wind turbine generator. In the bicycle you don't feel this peak as the inertia of the whole bike is rater big and once you are moving, you only have to supply the average torque which is rather low.

[TechAdmin]

@oneirondreamer fixed some of your links and attachments, be careful with the BBCode

[topspeed]

I paid nothing of my generator...as bicyclist often discard their bikes to repairshops when the control unit brakes down which is pricier. 

I'm curious as to it's geometry, do you know the pole and stator tooth count?  What sort of windings?   Many of the bicycle wheel motors seem to have low pole counts compared to these hoverboard motors with 30 magnets in only 105mm circle.

I opened it day before yesterday, but cannot recall the amount of magnets..it looks a lot deeper..maybe the magnets are bigger ie longer ?

[Gary]

The inadequate bearings and the challenge of the tire may be overcome easily.  I would think of mounting the blades on a shaft with two spaced pillowblock bearings first.  Then mount the wheel motor/ generator behind it with shaft's end to end.  The tire could serve as half of a flexible shaft coupler.  A disk on the end of fan shaft with a few pins on it's perimeter engaging in holes drilled through the tire.  Set it up loose and have a nice shaft coupler

[oneirondreamer]

Update, hoverboard alternator machined down and installed in my 3D printed, ceramic reinforced kinetic hydro turbine

https://www.instructables.com/id/Micro-Kinetic-Hydro-Power-System/

[oneirondreamer]

Hello All,

I've made some progress in a few area's.   

I've reconnected one of these alternators as was suggested by Adriaan.   

The first one was 10 turns of 5 in hand, I converted 10 turns to 50 turns, by wiring the 5 in hand into 5 in series.   This worked quite well, but gave me some strange waveform output.   The original winding had all of the star winding legs, very different lengths, I left it like this connected in series and the output was odd, not a clean waveform.   The original waveform wasn't super clean, but this went from 30 pulses per rev to 90, with 1/3 of the pulses aprox .5V lower than the others.

Second reconnection, the stator consists of groups of 3 posts, connected in series with 2 other sets of 3 posts, for 9 stator posts per phase.   What I've done this time is connected each of the groups of 3 posts in series, then connected these in groups of 3 in series.   This is 3x more connections than doing it the way I'd done it, however it also allows me to choose to run the 3 stator pole groups in series, or parallel, or possibly even series and parallel.  I hope to get it tested in next day or two, then if it works as expected I'll probably reconnect it with the 3 coil groups in parallel rather than series for I think 1/3 the voltage?

This has been a pain, but it makes me see that rewinding one of these into coil groups of 3, wouldn't be to difficult, and you could use 6 in hand, so that you could rewire in even more ways, both as high speed and low.   

I'm really curious to see if the RC motor controller will run it this way.    I'll let you know.

[oneirondreamer]

Success!!   It lives, first run was 60 rpm, 12V, 4.4 W.    A much cleaner and more consistent waveform, not too far off from the output of the previous setup (simple series of 5 in hand) in terms of volts output/rpm, but I think a bit higher amps (not sure what the total resistance load was, but it was low).   I could set up all three and test side by side, but I'm not sure it's worth my time right now.   

I think Adriaan's rewiring suggestion helped quite a bit.   I am curious about the effect of a short leg to the star point, in the earlier winding.   I may fix that and retest, but dang, who's got the time....