8 magnet 6 coil single rotor

[josephbird1]

I'm attempting a 300-400w machine with 8 2"x1"x0.5" n42 neos. The test coil is 2 strands of #15 guage and 50 turns.

I saw in hugh's book an 8 magnet 6 coil machine but the coils are touching and mine are spaced apart.

The reason for the space between coils is that my hub has a 4" bolt circle so I had 11" diameter plates cut for the rotors. Even at 11" I only have about 1/2" clearance between the spinning threaded rod and the bottom of the coils.

I made a test coil and the following pics are the resulting waveform.

Unfortunately the neos I bought have a hole in the center, could this be the trouble?

The scope is directly connected to the coil, no rectifier

[joestue]

Thats just a 3rd harmonic, caused by a nonsinusoidal distribution of the magnetic flux,

[Flux]

Yes there is indeed a problem with space at the centre if you try to use a standard hub.

Increasing disc diameter is a way round it. Few people look at the waveform on a scope and would never worry about what you see. I think your waveform is rather worse than usual because of the large magnet spacing and small coils. Things will be better if you wind bigger coils so that they touch as in Hugh's version. In reality it gives you more winding space. I suspect the waveform is likely to be worse with single magnet rotor especially if the coils are thick.

There is normally a lot of third harmonic in a single coil waveform ( and of course in the phase voltage). It is uncommon for the harmonic to exceed the fundamental, but it doesn't matter that much. The line voltage of a star connection will be much nearer a sine wave as in theory the 3rd harmonic cancels.

Try making your coils bigger by winding on more turns until they touch and then just live with the waveform. Such a bad waveform makes it a bit difficult for you to decide on what to take for your cut in speed. Some strongly advise the use of a scope but I find it adds confusion. I use a mean reading multimeter scaled rms ( the normal cheapo type multimeter) and just take its reading as though it was rms ( not accurate but works near enough). When you rectify that lot the waveforms you see will be pushed into insignificance.

There are a few combinations of coil size and magnet spacing that get reasonably sinusoidal, but usually there is a large 3rd harmonic component, either in phase or out of phase depending on the spacings. I really don't think it is important.

You probably need to keep near Hugh's number of turns but use thicker wire to make bigger coils or you may do better to use a similar size wire and make the stator thinner to get fat coils. Thick stators are not effective with this single rotor type magnet.

The hole in the magnet will have no effect that you will ever see.

Flux

[snowcrow]

 Are you using stainless steel threaded rod, nuts, and lock washers?

[josephbird1]

Ok thanks for responding!

The test coil was a bit over 1/2" thick. I rewound another coil with 2 #15 awg and 40 turns and it was 1/4" thick but the coil was much wider and the sinewave did improve to a noticeable sinewave. but now its dangerously close to the spinning threaded rod.

Do you think it would be ok to leave it as it is without widening the coils? or would that odd sinewave give me a poor machine? would it affect efficiency?

[Flux]

I don't think the poor waveform will have much effect except for making it difficult to decide on the number of turns for cut in speed. The 3rd harmonic will be cancelled to a large extent with the star connection and in any case the rectifier will mess the waveform up beyond recognition beyond cut in.

As this machine is not likely to be pushed hard you could probably stand a bit of increase in resistance with no ill effect.

Use as much of the space as you can without running foul of the threaded rod. You can improve waveform by making the hole in the coil bigger and retaining the leg width or you could make the coils more triangular or trapezoidal. Both of these two options will result in longer turns and more resistance. Either way I don't see any problem, I have done much the same thing but not increased the magnet spacing as much as you have done. I have no idea of the waveform and have little desire to find out. Just don't even think of delta connection and it will be fine.

Flux

[finnsawyer]

It looks like you should have moved the magnets 22.5 degrees relative to the bolt holes.  That would have allowed somewhat larger coils.  Too late to drill a new set of bolt holes?  It's too bad people don't figure these things out ahead of time by doing some modeling.

You don't give the time base for the traces.  The differences in the large peaks is a matter of some concern.  While it doesn't look like there is any interaction with stray 60 or 50 hz you need to rule that out.  Poor connections can let it in.