500 W motorconversion finished

[DamonHD]

You were clear enough: what you have looks great!

I was attempting whimsical humour with rather higher THD than your unloaded output...  B^>

Rgds

Damon

[dinges]

Ran a few more tests of the genny, this time of the waveshape as it was charging both a 6V and a 12V battery. Decided to upload the pictures here as not everyone may have access to an oscilloscope and see those images first hand. The first three pictures were made at the same settings, vertical scale 5V per division.

First image, unloaded generator just connected to the rectifier. No battery connected:

Next image shows the generator charging a 6V battery. It can easily be seen that the tops of the waves are clipped off:

Also notice the discontinuity at the zero-crossing, even though I measured waveshape at the AC inputs of the rectifiers...

Next image shows it charging a 12V battery:

It can be seen that much less of the original sinus is clipped off. One could probably compare the non-clipped off area of the sine with the 'lost' space in the cylinder of a combustion engine. Lowers the efficiency and doesn't really contribute to output, but is a necessary evil. (and without that 'lost' power the genny probably wouldn't start up decently in low winds).

The next two images show the frequency domain, after a Fast-Fourier Transform (FFT). It's a way of telling how 'pure' the original sine waveshape is. The peak on the left is the fundamental frequency of 21 Hz. The 3rd harmonic is at 63 Hz and much lower than the fundamental, -33 dB down w.r.t. the fundamental. I consider this to be very good. This spectrogram was made of the 'nearly unloaded' waveshape, that is, no battery or rectifier was connected, but a small 18V/0.1 A lightbulb was:

With a completely unloaded output (i.e. no battery, no rectifier, no small lightbulb as load) the spectrum looks a lot dirtier; the 3rd harmonic (note: different scale compared to the previous spectrogram) is now only -10 dB down:

So, all in all, the sine is both cases reasonably sinusoid, with the 3rd harmonic being at least -10 dB down. With a slight (linear) load, the lightbulb, the harmonics become much smaller. This last result of the spectrograms is the only thing I don't understand; how such a small load can have such a large effect on the cleanliness of the spectrum/waveshape. Not that it matters much in the reality of charging batteries and windmills, but still an unexpected observation (to me).

Peter.

[oztules]

Dinges,

Is it possible that the coil is ringing in the unloaded (last) photo, and that the small light globe acts as a snubber?

The harmonics seem very linear in their"fading away" from the  principal freq.

It is not what I expected, but upon thinking about it, I guess this is the price of having the teeth letting go as they pass the magnet, and experiencing a "twang/shock" in the magnetic field as the flux lets go of one tooth and realigns onto the next. This would have to show up in the impedance of the coils.

Even thinking about it, I would not expect to see this in the axial flux design.

.........oztules

[joestue]

It's not at all surprising, look at the current in each phase

Do you have a dc choke after the rectifiers? It should significantly reduce harmonics.

[SparWeb]

Your picture of the loaded sine resembles the green curve:

I think the voltage curve was clipped, too, depending on where you measured it.  I don't have an appropriate image of that uploaded, at the moment.