"Greg Mayman" bravely wrote to "Mike Ross" (08 Sep 03  09:00:00)
 --- on the heady topic of "VEHICLE LED'S"

 GM> If the cycle wheel has a circumpherence of about 7 ft, it will
 GM> turn about 3 times a second at 15 mph. Let's take someone's
 GM> stepup ratio of 35:1 for driving the dynamo. So the dynamo is
 GM> turning at about 100 times a second and generating an output at
 GM> 100Hz.

Okay...


 GM> 4.2 mH will have an inductive reactance of 2.6 ohms at 100 Hz,
 GM> and if the lamp current is 400 mA it will drop 1.06 volts. This
 GM> is NOT insignificant for a 6 volt bulb.

You are using the "inductive reactance" of 2.6 ohms as if it was a real
resistance but it isn't. You have to find the impedance to get the real
voltage across the lamp. Let's assume in this case the lamp presents a
constant 12 ohm load, then Z=12.28 ohms. The voltage across the lamp
will be I * RL = (6/Z) * 12 = 5.86 V and totally not like 6V - 1.06V!

Okay, according to the circuit simulator (N.B. I added a 0.2 ohms
internal winding resistance), I get 5.77 volts without the capacitor,
and 6.04 volts with the 29uF capacitor.


 MR> That's true but if it is a load critical value (i.e. XL=XC=RL) then
 MR> there is no resonant frequency and the effect of the capacitance is
 MR> spread out over a much broader frequency range.

 GM> Back to your books, laddie! It is resonant whenever XL=XC,
 GM> regardless of R.

Of course, it is resonant any time XL=XC, but I assumed you understood
that maximum power will be transferred from the generator to the load
when XL=XC=RL. That's why choosing the capacitive reactance equal to XL
"AND" equal to RL satisfies the desired unity power factor condition.


 GM> What I was objecting to is your comment that the capacitor will
 GM> cause the current flow through the lamp to occur over a greater
 GM> part of the cycle of the output waveform. IT WILL NOT!!!!

Why not? In the above circuit simulator example we can see how a
specific critical value of capacitance becomes useful in pulling up the
output voltage across a broad frequency range rather than simply at a
single resonant frequency.

Try the calculation at 50Hz, then at 200 Hz. You will find the lamp
voltage stays about the same assuming the generator is an ideal 6 volt
source. (My simulator gives 5.94V at 50Hz and 6.4V at 200Hz.)


 MR> There won't be many harmonics unless the pole steel behaves in a very
 MR> non-linear fashion and saturates.

 GM> Rubbish! Saturation is not needed.
 GM>
 GM> Apart from hysteresis effects, the purity of the waveform
 GM> depends mainly on a sinusoidal buildup and collapse of the
 GM> magnetic flux in the pole pieces.

 GM> This buildup and collapse is NOT sinusoidal in most generators
 GM> unless they have been specifically designed for this, and is
 GM> especially non-linear in units that are designed for maximum
 GM> output at smallest size, and when purity of the waveform is
 GM> irrelevant.

 GM> Check it on your oscilloscope.

Okay, I will try to find a longggggg extension cord... 


 MR> I've tested it on the circuit simulator and on the contrary the
 
 MR> The difference being that the circuit simulator didn't have the proper
 MR> generator model

 GM> In that case the results may not be anywhere near the truth.

That's a possibility but however improbable what remains is the truth.

 Mike
 ****

... I worked hard to attach the electrodes to it.
--- Blue Wave/DOS v2.30