This page discusses the circuit that I built to regulate the power output from my homemade DC generator that was at one point installed on my bike. Please note that these days, this circuit would be considered obsolete. A modern version would use a power MOSFET device to switch the power, and a PWM (pulse-width-modulated) switching mechanism instead of analog. However, the circuit described below does work, and is simple to understand.
If you are only using a normal 3 to 6 watt bicycle generator, you don't need this circuit. Instead, you should look at building one of the much-simpler Unregulated Battery Charging Circuits I also describe.
|Regulated Charging Circuit Schematic|
(please note, this schematic based on recollection only) It can be built fairly compact (mine is about 1 inch by 2 inches by 3 inches) and a lot of that is a small heat sink for the power transistors. By adjusting the variable resistor properly, I have adjusted it to put out a regulated voltage that is about 7.5 - 8 volts. This allows for fairly rapid charging of the battery, without frying the other circuitry on my bike. The regulator, with the components I used (I can't remember individual component numbers off the top of my head) is probably capable of working with loads up to about 30 - 40 watts. Of course, this will depend on the components you use and how big your heat sink is). I originally built mine to work with a 40+ watt homemade generator and loads that were similar. I eventually replaced that one with a 15 watt homemade generator. The regulator on my bike actually has two PNP Ge power transistors in parallel
This circuit works by increasing or decreasing its own electrical resistance to keep the voltage across the load (in this case, battery plus whatever accessories are currently running) constant at 7.5 to 8 volts, while the input voltage varies (due to the generator running at different speeds) and the load varies (depending on the how charged the battery is and what accessories are turned on). It works because, when there is little or no voltage across C2, then there is not enough voltage across the base of Q3 to turn it on. That means that transistor Q3 is off, (like an open switch between its collector and emitter) and that means that all the input voltage (mediated by resistors R1 and R3 (if there is any input voltage) is applied to the base of Q2, turning it on. When Q2 is turned on, it will pretty much cause the base of Q1 to be shorted to ground, which will turn it on as well (PNP transistors have the opposite polarity of NPN transistors), causing all the power that is available to flow through the regulator circuit, charging C2 and also powering whatever load is present. This, of course, causes the voltage on C2 to go up.
As the voltage across C2 goes up, the voltage across the variable resistor (VR) will go up as well. The voltage on the centertap to the VR will climb too, and eventually start to turn on transistor Q3. As Q3 turns on, it will start to act like a closed switch (between collector and emitter), and this will cause the voltage across the base of Q2 to drop, until Q2 starts to turn off. When Q2 starts to turns off, the voltage on its collector (and the base of Q1) will go up, because of R2. Thus Q1 starts to turn off when the output voltage of the regulator reaches a certain level.
Since transistors are (mostly) linear, as opposed to simple on/off like I described above, the regulator will reach some equilibrium where Q1 is turned on just enough to maintain the proper output voltage across C2. C2 is drained mostly through the load, and the bigger the load, the more Q1 will turn on to maintain the proper output voltage. This way, the output voltage gets regulated. The output voltage can be adjusted using the VR. The closer that the centertap is put to the negative side of the circuit (- on generator) the higher the output voltage will have to be to turn off Q1. the closer the centertap is to the other side, the lower the output voltage will be.
The other components that I have not mentioned are also (of course) necessary. The first diode (D1) is there because, since I built the generator out of a DC motor, spinning it backwards will make it put out the wrong polarity voltage. This can damage the regulator. D1 shorts out the generator when it is going backwards, keeping it from doing bad. (One could also put D1 in series with the Generator, but this introduces another diode drop that the generator has to overcome, reducing efficiency.) R4 prevents too much current from going into the base of Q3. D2 prevents the battery from draining itself through the regulator circuit, when it is not running. C1 is there because I thought it looked pretty. (and because it helps to mediate 'noise' coming from the generator, which would otherwise mess with my fine radio reception.). I chose Q1 be a PNP Ge transistor, despite its supposed inferiority to Si transistors, quite purposely. The main reason is because Ge Devices have a 'diode drop' of only about .3 volts, while silicon devices have a diode drop of around .7 volts. In other words, The regulator is more efficient with the Ge power transistor. Ge transistors tend to be a lot 'leakier' than Si transistors, but that does not matter for this application.
The regulator, as I have drawn it, can only accept a DC input voltage. However, you can make it work with an AC generator (this is not necessary for regular bicycle generators) by running the output of the generator through one of the Unregulated Battery Charging Circuits and then using the output of that circuit to power the regulator. The only change that would need to be made to the unregulated charging circuit that is to be used is that the battery or batteries must be replaced with large value electrolytic capacitors. (Try 2000uF or so, and make sure that the voltage rating is high enough to handle the generator running full-blast with no load) Keep in mind that the components in the unregulated charging circuit must also be able to handle the same amount of current as the high-power parts of the regulator. The best circuit to use for this would be the bridge rectifier.
There are all kinds of 'canned' voltage regulators that are available (try Radio Shack). The only ones I can think of off the top of my head are the 78XX and 79XX regulators, which are 3-pin packages where XX is the voltage they regulate to. The problem with these regulators is that they can only provide about 1 amp ( 5 or 10 watts) of power, which is not enough for my application, and their voltages are not adjustable. There are devices where you can control the voltage, but I don't know about their power dissipation. And, of course you can build a regulator from discreet components like I did. There are simpler (and undoubtedly more elegant) ways to construct a regulator from discreet components (I could get rid of Q2 and some of the resistors if I had used an NPN power transistor, for example) But I built it the way I did because of the lower voltage drop for the PNP Ge transistor. Like I said about all the unregulated charging circuits, specific details are best left up to you, since your situation would most likely be completely different from mine. If you do build a regulator, make sure that all the parts that will be carrying large currents (the diodes and power transistor(s)) are rated to handle at least twice the loads you expect to be putting on them. In other words, if you expect your regulator to supply 2 amps of current (16 watts if its output voltage is 8V), make sure the diodes and power transistor are rated at about 4 amps. Doing this ensures the parts won't overload and fail. Also make sure the wiring that you use is heavy enough to handle the current. (I have melted a few wires that were too fine gauge for their applications.) It is best to make sure that the power transistor has at least a small heat sink on it, too.