On Thu, 19 Jan 2012 15:34:24 -0000, "Ian Field"
wrote:


"John Larkin" wrote in message
.. .
On Wed, 18 Jan 2012 23:57:55 -0500, "P E Schoen"
wrote:

"John Larkin" wrote in message
...

I took one of Marcel's alternator models, ran it through a 2:1
stepdown transformer, and rectified that. On the sensible parts of the
frequency curve, the resulting DC current increases, about doubling
the no-transformer current towards the high end. Impedance
matching works!

Of course with the transformer you get less output on the low end, but
then that's what the added battery is for, to power the lights at low
and zero speed.

I just found Rick's data for the dynamo, and it appears that the RMS value
is about 1/4 the P-P voltage, instead of 1/2.8 as would be expected for a
sine wave. So I'm curious about the waveshape.

What I've seen is sort of spikey, unloaded. It will probably look a
little more sinusoidal loaded. But there are all sorts of alternators
out there.

And of course impedance
matching works, but I think it would be better to boost the voltage and
then
rectify it so the forward drop of the diodes has less relative effect on
efficiency. Perhaps the voltage doubler would be most efficient.

I think there would be plenty of voltage just rectifying and
filtering. But you'd need a switcher that works over a wide voltage
range and doesn't blow up when the alternator voltage is really high,
100 volts or maybe more. That's not all that difficult, but a Digikey
transformer is a lot simpler and more reliable.



Once converted to DC, impedance matching can be best attained by PWM and a
buck converter. And the step-down transformer (or any transformer) would
need to have a lot of iron to work at the lower limits where the frequency
is less than 20 Hz, although the voltage is about proportional to
frequency,
so that may not be significant.

It's only a few watts, so the transformer would be small. Use high
enough voltage windings that nothing saturates.


These days there are much better ways to achieve impedance matching and
maximum efficiency than by using a simple transformer.

Sure. But the transformer is simple and rugged.


And solid state
components are much lighter and cheaper than iron core magnetics,
especially
since they need to be specially wound to work optimally at the low
voltages
produced by the hub generator (2-7 VRMS).

The power produced by the bike hub generator would seem to be proportional
to speed, but it may drop off at higher speed because of the higher
frequency and the inductance. This may result in a relatively constant
power
output and a crude form of regulation that results in relatively little
variation in brightness over normal speed range.

What flattens out brightness is the constant *current* (into a load,
over the speed range) of these alternators. Both voltage and frequency
increases with speed, and there's a lot of inductance.

John


Were there spice/LTspice models somewhere or were you referring to the
graphs & tables?


Marcel's graphs, and others I've seen, are consistent with the simple
model of a voltage source (voltage tracking frequency) in series with
a resistance and an inductance. After all, that's what's actually
inside the hub.

If the pole pieces have funny shapes, the waveforms can be
non-sinusoidal. In as much as the designs usually attempt to deliver
constant current into a load, waveform distortion is equivalent to
higher frequency, which only helps.

Loading will make the current waveform more sinusoidal than the
open-circuit voltage waveform.

If one desires to rectify and filter, and then switch down, the
lightly-loaded capacitor voltage could be very high if the distorted
waveform has a high crest factor... you could get higher DC than the
RMS voltage suggests.

John