Rewinding old 3-phase stator: Non-symmetrical layout

[Kjell Amundsen]

Hi all.
I wonder if anyone has any thoughts regarding my DIY rewinding project as follows:
*Motorcycle stator off an old Yam RD350 (known as RZ350 in the US) needs rewinding after some of the coils were "fried".
*The stator has 3 larger high voltage coils used for the CDI system - and is left intact.
*The "charging" system consists of 14 coils, arranged as a star 3 phase system, of which several coils were damaged.

My findings.
I would have expected the coil layout to be symmetrical and the number of coils to be a multiple of 3, but as mentioned it consisted of 14 coils. Ie 2 phases had 5 coils and one phase had 4 coils. Unwinding these, i found that 2 phases had ~160 turn/windings in total, and 1 phase had ~180 turns.

This was achieved by combinations of coils with ~40 and ~20 turns respectively. Additionally what puzzles me is the spacing - as the 3 ignition coils fully "occupies" their own sector of the coil, leaving the rest to the 14 coils of the 3phase and somewhat non-symmetrical system. The 14 Three phase coils were arranged symmetricallyfrom the start as 1-2-3 then 1-2-3 and finally ended with 1-2.

So when rewinding, I wonder if I should aim at winding all 3hree phases with the same number of coils ~160 each, assuming this must be symmetrical for an even load between the phases and that there must have been an error With 180 turns for one phase only?
Or, could there be a thought behind this from the factory as a form of compensating for the non-symmetrical layout as described?

Any input towards understanding the nonsymmetrical nature of the cloil/windings layout, and also what strategy to apply for rewinding would be appreciated ...ie blindly copy the original or not.

link to Picture of stator https://1drv.ms/f/s!ArCvt5JQQSh2sxvPb2xKDuGqkqA8
linkt to ppt descripiton of what I found https://1drv.ms/i/s!ArCvt5JQQSh2sw5E_tyEp2B_FqT8

Thanks for Reading
Kjell

 

[Baluncore]

Welcome to PF.

Does the alternator have permanent magnets in the armature or does it have field windings?
All pictures I see show the outside of the armature, how many magnetic poles are there in the armature?

You can attach or insert a .jpg or .png picture by drag and drop onto the post.
Third party sites and power point makes it hard to get the images.

 

[Kjell Amundsen]

Hi and thanks for reply.The external flywheel has embedded/hidden magnets so it is hard to tell how many poles, but I expect it to be multiple.

-Perhaps that can be investigated by using another permanent magnet along the perimeter and count the pole changes?
Further, if it is a multi pole magnet rotor, does it really matter if the coils does not make up a full circle as long as the phases alternate in the correct sequence 1-2-3 and so on?

Is it the harmony between the permanent magnet poles and the phase-spacing of the coils that is the key here when the coil does not make up a cirlce. I mean if the spacing and rhythm is timed correctly can one emulate a correct sine wave output?

 

[Kjell Amundsen]

Flywheel fitted externally to the coils seen in first post. Permanent magnet are encapsulated along the inside of the flywheel/rotor.

 

[Baluncore]

I mean if the spacing and rhythm is timed correctly can one emulate a correct sine wave output?

Yes. But not knowing how many poles are on the armature makes it difficult to understand the alternator stator turn counts.

Take a steel screwdriver and run the tip along the internal magnetic surface, put a mark wherever it wants to settle, they will be the magnetic poles and will alternate NSNSN... Counting the equally spaced marks will tell the number of poles.

Based on the three HT coils that appear to be a magneto, I would guess 4 poles in 90°, that suggests 12 PM poles on the armature.

 

[jim hardy]

We are so trained by textbooks and lecture notes to expect symmetric and balanced 3 phase windings that we (at least I ) recoil from anything different ie unbalanced.

I remember in the 60's when GM introduced the little Buick 90 degree V6 engine with its slight 'lope' at idle , but they held up well. And nothing could more unbalanced than a single cylinder engine... or the single charging coil on old outboard motors...

So there's no reason to insist on perfect symmetry and balance in any machine even if it reminds us of "Modern Art" . I see so much of myself in the OCD TV detective "Monk" that I've come to laugh at myself .

So my advice is put it back exactly like you found it. I'll take a look at your drawings this evening , been sans internet a couple days so hadn't seen your request for help.

The usual cause of burnt up alternator windings like you describe is somebody replaced the battery and got polarity reversed which made some sparks. So they immediately swapped the wires back to their correct position. While polarity was backward it shorted one or more pf the rectifier diodes, allowing reverse current when battery was reconnected correctly. The battery then runs itself down through the shorted rectifier and the alternator windings frying them in the process.

 

[Kjell Amundsen]

Hi
I did ask a stator aftermarked supplier about this non symmetric layout.
"This is usually
compensated for by winding slightly less turns on the 2 coils with
more poles to get about the same amount of current"

I understand "coil" here as the total windings distributed between several "poles" making up a phase. Thus indicating that total and near equal number of windings is the goal even if the geometrical distribution is a little odd. Or do I get it wrong?

 

[jim hardy]

I think you're right.

It'd be interesting to know how the outer rotating magnets are arranged.
I have been looking for a magnetic assemblies supplier site to no avail - i think Google must be mad at me for being conservative, they used to hand me good solid technical sites but now it's all fluff and retail ads.

Anyhow
perhaps with a boy scout compass and a skinny screwdriver you could figure out how that rotor is magnetized.
I suspect the magnetic flux path in the stator is something like this, six magnetic poles,

The phase having one less coil needs more turns to make up for the smaller total magnetic flux that it sees.
That's because voltage e = N_turns X (rate of change of flux)
So rewind it as close as you can to what they had. Matching direction of turns is VERY important.

It's counter-intuitive to me that you found turns ratio of 160/180 turns, a ratio of 0.89(to just two figures) not same as the the flux ratio 4/5 = 0.8
But square root of 4/5 is also 0.89(to just two figures as well)
so it looks like they chose √(flux ratio) for turns ratio, flux proportional to turns^² .

Inductance is proportional to N_turns^². I'll have to mull that over . @Tom.G ? .(@ not working again btw)

anyhow - "Put 'em back the way they was".

okd jim

 

[jim hardy]

Thus indicating that total and near equal number of windings is the goal even if the geometrical distribution is a little odd. Or do I get it wrong?

There's a concept of "Flux Linkages", product of magnetic flux (Webers) and number of turns encircling it. That's most likely what they tried to make equal. As i said I'm still mulling over the square relation you observed.

 

[Kjell Amundsen]

Thanks. I will get back with the permanent magnet setup of the rotor/flywheel within a couple of days.

 

[Baluncore]

The beauty of a three phase star configuration without a central neutral connection, is that a circle can always be drawn through the three points. So the phasor diagram has a slight phase error but will be nowhere near as bad as it could be.

It looks to me like the 3 well insulated stator poles for the ignition are different sizes. The smaller pole, the most clockwise of the three, looks like it is part of the original 3PH alternator pitch, making 15 stator poles, with two ignition coils or magnetos. The two larger coils are on a different pitch, maybe to line up with a N-S pole pair. The small insulated coil is not as simple as it might appears. It would be interesting to know the way the insulated coils are interconnected.

There were several alternative alternator designs for the same engine. Some had no PM, but had a field winding for regulation, others had less poles. The pole count of a PM armature can be found quickly with any steel screwdriver, polarity is irrelevant. It is hard to work out what is really happening without that pole count.

 

[Kjell Amundsen]

It looks to me like the 3 well insulated stator poles for the ignition are different sizes. The smaller pole, the most clockwise of the three, looks like it is part of the original 3PH alternator pitch, making 15 stator poles, with two ignition coils or magnetos. The two larger coils are on a different pitch, maybe to line up with a N-S pole pair. The small insulated coil is not as simple as it might appears. It would be interesting to know the way the insulated coils are interconnected.

Actually the 3 separate ignition coils are not part of the 3ph at all. They are split into what is reffered to as a high and low voltage output. I believe this is to ensure proper Volt and Amps at both idle and full power respectively. Ie between 1500 rpm and 10.000 rpm range. This will also be unregulated AC fed directly to the ignition control unit which is independant of the bike's DC regulated system.

 

[Tom.G]

It's counter-intuitive to me that you found turns ratio of 160/180 turns, a ratio of 0.89(to just two figures) not same as the the flux ratio 4/5 = 0.8
But square root of 4/5 is also 0.89(to just two figures as well)
so it looks like they chose √(flux ratio) for turns ratio, flux proportional to turns^2 .

Hmm, interesting approach...

(think... cogitate... think)...

I'm more inclined to consider the windings the same you would a transformer, Volts-per-Turn. Could it be there wasn't enough room on the winding bobbins for more turns and they just added what would fit, and called it 'Good Enough'? (Wouldn't be the first time reality got in they way!)

 

[jim hardy]

I started my thinking along this line:
"Consider anyone coil. It sees alternating flux so will have AC induced in it. So will its two adjacent neighbors, and if indeed a single magnet spans three coils the three will enjoy that nice three phase relationship we've all grown to respect and love.
The single coil we considered is wired in series with another coil in the next set of three, occupying the same respective position in that set.

Were it not for the high voltage coils that pattern would continue clear around the core and we'd have a traditional three phase machine with eighteen windings..
That they leave out a few coils does not change the fact that 3 phase AC is induced in those fourteen charging coils .

Flux will be relatively balanced in most of the stator. Over by the ignition coils it will be probably less so. I agree with @Baluncore they left out four of the charging poles and put in their place three ignition coils with different sized pole faces..

So the charging circuit consists of two phases of five coils in series and one phase of four coils in series , number of turns per phase adjusted to provide equal terminal voltage .

We're accustomed to nice symmetrical machines but clearly symmetry is not a "physics must have" but a "psychological must have " . Therein lay my difficulty accepting this machine. When i get my hands on one i will study it, meantime TIL nobody said a machine has to be magnetically symmetric.Now, these machines employ armature reaction to limit current in the charging coils and control flux through their cores . Not knowing what is connected to those ignition coils we can only speculate what is flux there.

A most interesting machine ! The humble motorcycle / lawnmower alternator has many lessons for us. I look forward to further study of them.
Clearly symmetry is not a requisite.

https://www.briggsandstratton.com/c...n_us/Files/FAQs/alternator_specifications.pdf

Meantime - rewind it exactly as you found it .
Check your voltage regulator for shorted rectifier diodes before reconnecting the battery. old jim

 

[Kjell Amundsen]

Hi again. Regarding the magnetic flywheel. I did the screwdriver thing suggested earlier and found 12 "equilibrium" spot.

I am a little uncertain on how to interpret that into generator jargon, but are we talking about 12 magnetic poles then?
If my findings had been that all three phases had the same total number of windings despite beeing unevenly spread geometrically, it would have been much easier to accept. Still I am struggeling to understand why one of the phases with 5 coils should have 20 more turns than the other two.

In light of that; does it matter for all practical purposes whether you have 40 windings in one coil or 2 x 20 windings spread between 2 coils, given that the phase-rythm matches the coils in series?
If yes, then the strategy would have been to have equal total number of windings for all phases, spread between the 14 coils as discussed earlier. But was not the case... or now I am beginning to doubt if I counted wrong somewhere.
I will have to decide if the "total number of windings will the best strategy!

Red marks on flywheel at the spots where the screwdriver seemed to settle in the photo below.

Stator seen from the back side, and added winding Counts next to the coils

]

 

[Baluncore]

Regarding the magnetic flywheel. I did the screwdriver thing suggested earlier and found 12 "equilibrium" spot.
I am a little uncertain on how to interpret that into generator jargon, but are we talking about 12 magnetic poles then?

There are 12 poles, alternating -N-S-N-S-N-S-N-S-N-S-N-S- each pole is connected to it's two neighbours with opposite polarity.
If the stator was fully populated it would have 18 poles. The difference generates the 120° phase shift between the three phases.

 

[jim hardy]

Could it be there wasn't enough room on the winding bobbins for more turns and they just added what would fit, and called it 'Good Enough'? (Wouldn't be the first time reality got in they way!)

Wouldn't be the first design by trial-and-error either.

 

[jim hardy]

If the stator was fully populated it would have 18 poles.

so individual charging coils are spaced twenty degrees apart

There are 12 poles,

His picture sure looks like a rotor with 12 magnetic poles

that's thirty degrees apart
over twenty degree coils... ( i missed a little - yellow lines should be dead center on one pole face and midway between next two)

 

[Tom.G]

Using your image with the turn counts labeled (copied below for concenience), let's build a table with the Number of turns for each phase.

Starting from the left and moving clockwise, for convenience let's label the coil phases as:
A, B, C - A, B, C - A, B, C - A, B, C - A, B (this could be incorrect, can't trace most of the wires in the photo.)

Coil turns for each phase:

      PHASE
   A     B    C
T  20    40   40
u  40    40   40
r  40    20   40
n  20    40   40
s  40    20
-----------------
  160   160  160

The winding totals look identical to me, or did I mis-read the image somewhere?

Stator seen from the back side, and added winding Counts next to the coils

]

 

[jim hardy]

Coil turns for each phase:

Wow it gets curiouser and curiouser. Phase C has only four magnetic cores so ought to come up short on N X dΦ/dt .

 

[Tom.G]

Wow it gets curiouser and curiouser. Phase C has only four magnetic cores so ought to come up short on N X dΦ/dt .

Yeah. Phase C would be present for only 4 out of 5 complete cycles; a loss of 6.7% of total output. Some clever folks did that one! (Edit: If my image interpretation is correct.)

 

[Tcarey]

Did you notice that the magnets are not equally spaced? On one side there are 5 and the other there are 7. How that would compensate for the asymmetrical distribution of the coils is more than I want to think about.

 

[Kjell Amundsen]

I checked my notes and reviewed the photos once more.
It for sure sums sup to something very non-symmetrical like this:

Phase A has 5 coils With 160 turns total
Phase B has 5 coils With 180 turns total.
Phase C has *4 coils With 160 turns total.

I the picture phase A starts at 9 o'clock and continues clockwise "A-B-C-A-B-C-A-B-C-A-B-C - A-B"
If the total number of coils were equal as Tom.G suggested in the table I would be happy to accept it as good enough. But it is that 180 turn phase that does not fit quite in with the rest.
And therefore I am inclined to adopt a strategy for rewinding that at least leaves me with equal number of windings pr phase.

That being said - will one coil of 40 windings give the same V/A as 2 coils of 20 windings -connected in series and timed correctly with the magnets?

Ie would it matter voltage wise if 160 turns is distributed along 4 vs 5 coils in the phases?

And could that 180 turn phase be an error that ultimately led to the failure?

 

[jim hardy]

That being said - will one coil of 40 windings give the same V/A as 2 coils of 20 windings -Connected in series and timed correctly With the magnets?

When unloaded the voltage would be the same.
When loaded to any given value of current, i'd expect the two 20 turn coils to experience less voltage drop than the single 40 turn coil. That's because the amp-turns from load current oppose the permanent magnet field which reduces magnetic flux, and voltage follows. Half as many turns halves that opposing mmf. That effect is known as "Armature Reaction" .

Ie would it matter vltage wise if 160 turns is Distributed along 4 vs 5 coils in the phases?

Again , not in absence of load current. But when load current commences so does armature reaction.

And could that 180 turn phase be an error that ultimately ld to the failure?

That would surprise me but i get surprised all the time.
It'd sure be easy to leave out that most CW phase B winding .What test might one devise to see if load current is shared equally among the three phases ? Do you have a clamp around ammeter or three little pass-thru CT's for testing ?

 

[jim hardy]

I might be overly worried about balancing amp-turns. Could be they trial-and-error arrived at a design, or could be somebody on the assembly line slipped in a wrong coil .

How many windings did you find burnt? Were they all in same phase ? How does the rectifier check out ?

 

[Kjell Amundsen]

I might be overly worried about balancing amp-turns. Could be they trial-and-error arrived at a design, or could be somebody on the assembly line slipped in a wrong coil .

How many windings did you find burnt? Were they all in same phase ? How does the rectifier check out ?

I might be overly worried about balancing amp-turns. Could be they trial-and-error arrived at a design, or could be somebody on the assembly line slipped in a wrong coil .

How many windings did you find burnt? Were they all in same phase ? How does the rectifier check out ?

It was coil nr 4 from the start at 9 o'clock on the picture that was most severly scorched and the wire was burnt off. Secondly the two neighboring coil was suffering from overheating and was loosened on the coil in addition to burnt wires. When measuring continuity of the 3 ac output wires they all showed open line.
I suspect I have been running off one phase only for a while before total failure, as the charging was definately poor.

Rectifier/regulator I haven't checked yet. It is also a combined unit, so I am not sure if the Connector pins can be used to measure rectifier diode bridge correctly. I will look into that when rewinding is done.

Further thought:
What if I simply choose to use 3ph x 4 = 12 coils - fully symmetrical - leaving out 2, and put 40 turns on all 12? It cannot be much power to be lost that way? I could even probably squeeze in a couple more turns extra on each of the to achive let say 170 on each phase.

And what made the manufacture think it was such a good idea to go "creative" I Wonder.

(excuse the Capitalization of many of the Words in my posts. I think it is me writing English on a Norwegian computer and unknown Words are probably being classified as names and thus automatically capitalized)

Regards
Kjell

 

[jim hardy]

And what made the manufacture think it was such a good idea to go "creative" I Wonder.

I've wracked my brain on that question to no avail.

What if I simply choose to use 3ph x 4 = 12 coils - fully symmetrical - leaving out 2, and put 40 turns on all 12? It cannot be much power to be lost that way? I could even probably squeeze in a couple more turns extra on each of the to achive let say 170 on each phase.

Were this mine i'd try that just to see whether it works. I think it will.
I know you don't want a research project. I would find a way to instrument the leftover poles and estimate their flux.

But the real objective is to return your motorcycle engine to service.
So find a way to check the rectifiers before you connect a battery..

old jim

 

[Tom Graysopn]

So when rewinding, I wonder if I should aim at winding all 3hree phases with the same number of coils ~160 each, assuming this must be symmetrical for an even load between the phases and that there must have been an error With 180 turns for one phase only?
Or, could there be a thought behind this from the factory as a form of compensating for the non-symmetrical layout as described?

First comment I have here is that there will be a reason for this. With mass-produced Motor vehicles (and the old RD350 was a popular machine in its time), the manufacturers have spent a lot of time getting it right, so stick to the original scheme, no matter how much it doesn't make sense to you or me.

Secondly, If the Yamaha is the same as My HONDA VFR800 the three-phase AC is fed directly into a rectifier and becomes DC, so any imbalance can only have an effect of the actual windings that you are repairing.

A Question, DO you know what fried the stator windings? Was it a failure somewhere else, and you have found and repaired it, OR was it just a time thing?. My experience is that the regulator rectifier and the stator windings go hand in hand, if one fails you should be examining the other to be sure it is ok before starting it all back up again.

Some background on the Regulator-Rectifier . (R/R)
I took a quick look at an RD350 Forum and it appears that the Regulator rectifier works on the same principle as used by a lot of other motorcycles. The Field poles are permanent magnets so there is no control over the output voltage. It changes with engine RPM. In order to get useable output at low RPM the Magnets must be specified to give at least 13 to 15 Volts at idle, and as the Engine RPM Increases the voltage will increase above this, and can get to a couple of hundred volts at Max RPM if the R/R is removed.
To keep the votage down to the correct level the R/R will start shorting the output of the windings through its semiconductors, to "Pull the volts down" . This is why there is a substantial heat sink on the R/R .
If the R/R is turning on incorrectly or Failed short, this could be the reason for the fried windings.
Just a thought.
I notice, like on other motorcycles, Third parties have started marketing phase controlled Rectifiers to remove this seemingly senseless "Shorting", in a bid to make life a little easier for the stator windings.

Forgive me if I have gone a little Off Topic. Tom

 

[Baluncore]

This is the evolution of the current system as I now see it.

The external rotating PM armature has 12 poles, each separated by 30°. As the armature rotates, the passing NS field repeats once every 60°, that is six times per revolution.

The fixed 3PH internal stator would have 18 poles if fully populated. That makes 20° between poles. Each repeat of the 3 phase coils will cover 60°. That is again six times per revolution, obviously a necessity rather than a coincidence.

The ideal simple 18 pole alternator. Series connect poles to make each phase.
Phase A. = 1, 4, 7, 10, 13, 16 = total 6 poles.
Phase B. = 2, 5, 8, 11, 14, 17 = total 6 poles.
Phase C. = 3, 6, 9, 12, 15, 18 = total 6 poles.

But the last 3 poles are missing from the stator pressings, where the two larger HT poles are now placed. The two larger HT ignition poles are separated by 30° degrees, NOT by 20° like the 3PH poles. The two HT poles therefore produce voltages that are 180° out of phase as the rotor PM poles pass. I don't know how the HT coils were or are now connected. Maybe they were once magneto coils supplying the spark to the two cylinders. It has evolved since from mechanical points to electronic timing and CDI.

It is clear that the smaller of the three HT poles was originally used to complete a symmetrical 3PH system along these lines.
Ideal 18 pole alternator, but missing three poles.
Phase A. = 1, 4, 7, 10, 13 = total 5 poles. 5 poles x 40 turns = 200 turns.
Phase B. = 2, 5, 8, 11, 14 = total 5 poles. 5 poles x 40 turns = 200 turns.
Phase C. = 3, 6, 9, 12, 15 = total 5 poles. 5 poles x 40 turns = 200 turns.

At some point, one of the 3PH power poles was pinched for use as an HT ignition power or timing winding. This was when things got messy.
An 18 pole alternator, but now missing 4 poles, (never used).
Phase A. = 1, 4, 7, 10, 13 = total 5 poles. 5 x 40 turns = 200 turns.
Phase B. = 2, 5, 8, 11, 14 = total 5 poles. 5 x 40 turns = 200 turns.
Phase C. = 3, 6, 9, 12, - - = total 4 poles. 4 x 40 turns = 160 turns.

To balance the phase voltages, half the turns were removed from the phase coils on either side of the HT poles, conveniently phases A and B, which made more space for the HT insulation and went half way to fixing the phase voltages. It seems the original plan was to do the same to another pole on A and B which would have balanced the phase voltages.

Alternatively the removal of 5 turns from every full A and B phase pole would achieve the same result, as follows.
18 pole alternator, missing 4 poles, but with full phase voltage balance.
Phase A. = 1, 4, 7, 10, 13 = total 5 poles. 20t + 35t + 35t + 35t + 35t = 160 turns.
Phase B. = 2, 5, 8, 11, 14 = total 5 poles. 35t + 35t + 35t + 35t + 20t = 160 turns.
Phase C. = 3, 6, 9, 12 = total 4 poles. 40t + 40t + 40t + 40t = 160 turns.

But pictures of replacement alternator stator assemblies now show full windings, except at the ends, so it appears that voltage balance has now been forgotten and the manufacturers have lost contact with the design engineers.
That suggests the present 18 pole alternator, missing 4 poles, but with partial phase voltage correction is as follows;
Phase A. = 1, 4, 7, 10, 13 = total 5 poles. 20t + 40t + 40t + 40t + 40t = 180 turns.
Phase B. = 2, 5, 8, 11, 14 = total 5 poles. 40t + 40t + 40t + 40t + 20t = 180 turns.
Phase C. = 3, 6, 9, 12 = total 4 poles. 40t + 40t + 40t + 40t = 160 turns.

To verify there is an error we would need reliable counts of the turns, or to measure the RMS output voltage of the 3 phases relative to the star common centre connection.

Only during starting or at low RPM may there be insufficient voltage on phase C. Phase voltage balance is not super critical when there is no common neutral connection. It would make small amplitude differences and phase errors over each full cycle. There may be no need to fix a phase voltage difference that has now become fossilised by the marketplace.

 

[Baluncore]

Tom Graysopn, welcome to PF.

A Question, DO you know what fried the stator windings?

My suspicion is that the burned coil developed a shorted turn(s). That then acted like an electromagnetic brake which cooked that coil until it went open circuit. Other windings on the stator and the 3PH rectifier diodes would have been out of the loop and so survived. The local heat can bring down neighbouring windings.

The 3PH alternator probably operated on two phases for some time before a second phase suffered the same problem which cut two phases and so stopped battery charging. Alternatively, the alternator diodes may have failed later and revealed the older independent problem.

 

[Tom.G]

I checked my notes and reviewed the photos once more.
It for sure sums sup to something very non-symmetrical like this:

Phase A has 5 coils With 160 turns total
Phase B has 5 coils With 180 turns total.
Phase C has *4 coils With 160 turns total.

Our discrepency seems to be the B phase coil at the 2 o'clock position. The photo is unclear there and it could be labeled either '20' or '40' turns. My assumption of 20 turns was based on three observations and an assumption.

Observations:

• In the photo, the count label seemed closer to '20' than to '40'
• Phase A clearly has two 20-turn windings
• The photo may show that the coil in question has more of the bobbin end-piece showing than the 40-turn coils (this is highly questionable though)

Assumption:

• For load balancing, all phase voltages should be identical; a reason for equal turns

Perhaps @jim hardy, the local power guru, can come up with some 3-phase waveforms showing the two different interpretations. This would require accounting for the field rotor passing both the non-symmetrical physical layout and the differing turns counts. (Sorry Jim)

It may turn out that everything evens out in one of the implementations.

(Wow. A couple good posts while this was being written. Lots of various ideas.)

 

[Baluncore]

Phase A clearly has two 20-turn windings

That may be true of the faulty example, but it is not the case with all advertised replacement stators.

 

[jim hardy]

Perhaps @jim hardy, the local power guru, can come up with some 3-phase waveforms showing the two different interpretations.

My alleged brain hasn't been able to work this one mentally.
I was struggling to come up with something like baluncore wrote .

Until it's verified the regulator rectifiers are intact we have the unresolved question "Did the battery run down backwards through the winding and wreck it?" That happens a lot .

Shorting the windings doesn't hurt anything because only enough armature current flows to cancel the permanent magnet's field , that's how they regulate voltage.
However in normal operation those those armature amps flow through all 40 turns on a coil.
Should a single turn on one coil short, current in that turn will be many fold higher so will overheat that turn and damage adjacent turns on same pole. .Your photo suggests that , but it does not explain why the other phases open circuited too.
That's why i keep beating the "Look at the rectifiers" drum. It's a single point failure that can affect all 3 phases through a neutral tie.

Sorry @Tom.G , google is still rebelling and only serves me advertising.

old jim

 

[jim hardy]

I've been more distracted by the 30 degree magnets vs 20 degree coils. Sounds like "Fractional Pitch" and 2/3 pitch reduces harmonic content.
https://www.ntnu.no/c/document_libr...205-cb04-4928-9ba1-c3eed0b1f62e&groupId=27629

Anyhow - it's an experiment. My advice is build a trial with balanced windings that'll be easy to take back to original if it turns out necessary.

We don't know what goes on in the ignition coils magnetic circuit . Amp-turns in adjacent charging coils can put the squeeze on flux over there.

Nice observations by all here.

old jim

 

[Baluncore]

Here is a plot of three phases, two with peak amplitude ±1.0, one with amplitude ±0.9, plotted relative to the star common = zero.
Once the floating three phase star is perfectly rectified and referenced to zero, the output voltage varies between 1.4 and 1.7321 = √3.
Notice the minima of the rectified signal has a slight phase shift as expected.
When the phases are balanced, all with amplitude ±1.0 then the perfectly rectified output varies between 1.500 and 1.7321