You need a high voltage stepper motor. A 2 or 3 volt stepper motor would have to spin incredibly fast, faster than it would in a hurricane, in order to generate enough voltage to charge a 12V battery. A lot of amps at 3 volts will do little to charge a 12V battery without a dc/dc converter or something to boost the voltage.
For servos, to a point, the higher the voltage the better. I've read where people use 30V motors successfully, but I think even then, depending on the nameplate rpms, it would likely take a pretty high rpm to make 12 volts. You want something that you can build blades to match in both the rpm and the power. I tend to think that a high tsr machine would put more wear and tear on the bearings.
For example, if I understand correctly:
If your servo motor can make 3 amps at 15 volts at 300 rpm, then you need to be able to make blades that can approximate this. The power your taking out would be about 45 watts. If you are trying to make this at a 12MPH wind speed then, if we can expect to extract about 1.75 watts per square foot of swept area at this speed, based upon roughly average performance of machines posted on this site, then we would need to sweep about 26 square feet. We would need to have a rotor diameter of roughly 5 3/4 feet. The tip speed would need to be about 300 * pi * 5.75 / (5280 feet/mile)*(60 min/hour) = 61.6 mph this would be a tsr of 61.6/12 = 5.1, which is, I believe, what we would plug into Ed's blade calculator or Hugh's formulas to calculate the dimensions of the blade required.
The power out is generally less, often significantly, than the nameplate power in at a given rpm so emperical data is best. If you are looking at a 3 to 5 foot machine, and you calculate it out to require about a tsr 5 to 7 blade, then I would think it would be a pretty reasonable motor.
Take all of this with a grain of salt as I only have limited practical experience applying these principles. I would be curious to hear others perspectives.
Rich Hagen