Greg Mayman wrote in a message to Roy J. Tellason:

 -=> Roy J. Tellason said to Jasen Betts
 -=> about "VEHICLE LED'S" on 08-29-03  20:01.....

 RJT> Another easy fix for that is to use a current-regulator.  Something
 RJT> like an LM317 (?) and a resistor,  in line with the lamps...

 GM> Let's see... you'd have to drop 1.25v across the resistor, and
 GM> there needs to be at least 2 volts input/output differential, so
 GM> you'd have to generate more than 3+ volts above the bulb voltage. 

Hm.  Perhaps they make lower-voltage chips of that sort?

 GM> Personnally I think a constant voltage regulator would be better.
 GM> You would only have to supply about 2 volts above the lamp voltage.

The common 7800 series parts wants more like a 3 volt difference, 
otherwise the regulator just "drops out".  They _do_ make some
lower voltage parts,  though,  some requiring a volt or less.

 GM> There is another problem, that most cycle generators put out AC, so
 GM> you have to add in the drop across a rectifier to the above losses.

I wasn't aware of that.  Ok,  a schottky diode (lower voltage drop) too.

 GM> Even with a capacitor input filter, the impedance of the usual 
 GM> cycle dynamo is too high to charge the cap to the peak voltage, as
 GM> you can do with a mains supply to offset the diode drops.

How have you determined that?

 GM> The best form of regulation would be a shunt regulator to bleed off
 GM> some of the current when the voltage got too high for the bulb.
 GM> Unfortunately this would increase the load on the dynamo. 

Yeah,  I imagine that it would.

---