Hi
I have an old school Royal Enfield motorbike which uses a blackbox rectifie r regulator unit driven by a 4 wire alternator (permanent magnet rotor, two separate pairs of coils). By default one set of coils drives the headlight s via a shunt regulator and the other set charges the battery via the rec-r eg unit.
The consensus among enthusiasts is to switch to a DC headlight circuit, by connecting the alternator coils in parallel and using them only to charge t he battery. Many people successfully do this, this avoids running separate AC wires to the front of the bike and lets the headlight work brightly even at low RPMs in the city.
However I don't like the fact that they use a shunt regulator : The alterna tor produces about 180 watts and maximum of about 50 volts - at full RPM, t he shunt will be dumping all that wattage and it will dissipate as heat in the alternator coils, inside the primary case which is already burning hot from the clutch and primary drive friction.
I want to make a simple series regulator with minimal component count, and also have the ability to run without battery if need be (this is a points b ased bike). Perfect regulation is not critical, it's more of an overcharge protection.
This is the circuit I have in mind :
Note that the alternator will produce upto 6200 Hertz, peak voltage is abou t 50V, maximum current output is about 15 AMPS - peak intermittent load fro m the bike (horn, lights, ignition) may be upto 20 amps.
I wonder if maybe the 10K is much too high or whether a 2n3055 can take tha t voltage and current. Do I need a darlington pair? The output should be 13.6 volts for a lead acid, as far as I know.
How would one build a similar circuit around an IRF540 instead? Does it make sense to put the capacitor as shown? My intention is that if t he battery goes kaput (it happens often!) , i can simply disconnect it and the capacitor should be able to handle the ignition load at least.