Hi Lexx -
Fred, and others have put fourth some good comments, and Ill add a bit.
Search this board for things like "coil shape" etc... to find lots more thoughts on the matter*.
I wonder how big your prop is? If you know that - and you know the approx TSR that you want the prop to run at, then you should be able to determine the approx rpm that the alternator needs to cutin at.
For any given stator, and any given rpm - then the power available will be directly related to the magnetic flux through the coils. If the magnets are too small... the prop will overpower the alternator - itll overspeed, the stator might overheat. There are more, and less "efficient" stator designs when it comes down to the flux required to get the right result. It's a bit tricky I think!
I usually take the approach of using a lot of magnet (because I sell them and have lots and lots available ;-) ) and then build my stator and often find I have to open up the airgap so that the alternator doesnt overpower the prop.. causeing the TSR to fall off and the blades to stall. Or... sometimes this allows me to run a thinner line from the batteries to the windmill I think. There has to be a certain amount of resistance in these sytems - too little, and the prop will stall - too much, and the prop will overspeed - and, the stator might heat up too much. The bigger the magnets you use (the stronger the flux) then the lower the resistance will be in the stator. You can choose to either have resistance in the line, or in the stator... (there will always be some in both). Using xtra large magnets means lower resistance in the stator - and you may have to add some to the line. Xtra $$ spent on magnets = less $$ spent in the line perhaps. (am I sounding like a salesman??... these are my current thoughts anyhow). If your running at low voltage, and the distance is long - saving a wire gage size or two in the line can add up.
That wasnt exactly your question though...
"there must be a way of deciding how long to make your coils to get the proper voltage."
Voltage is directly related to rpm, and it's directly related to the number of windings in the coil. Since you should know exactly what rpm you need to cutin at - then you shoot to make coils that will do that. When I start with a new design, I layout the magnet rotors Ill be working with. Then I draw a template that is exactly the shape that I think the coils should be. Some will say that the hole in the middle of the coil should be the size of the magnet exactly... or some will suggest slightly larger. My thought is the hole in the middle should be exactly as high as the magnet is tall, and slightly less in width. This is after a few fairly unscientific test I've done and I could be wrong. I think the rules may change a bit between disk, and bar magnets. Basicly though - the hole should be about the size of the magnet. If the magnets are well spaced, Im of the opinion that the OD of the coils should be such that they pretty much touch each other in the stator.
So once I know exactly the right shape and size of the coils, I would take some wire (usually Ill use a fairly fine gage but it doesnt matter too much) and wind one coil to exactly the right size. Then, poke it in between the magnet rotors, and spin them at a known rpm - and measure the output of that one coil. You'd be measuring single phase - probably RMS (depending on your meter) AC voltage. Actual voltage in your case then, since youd have 4 coils per phase - and youd probably be wiring in Star, would be the output of the coil, X about 1.4 (this gets you about the peak voltage instead of RMS), and then multiply that X 4 (that would be voltage for 4 coils in series) and then multiply that times 1.72 (that would be about final AC voltage youd expect to see from the alternator with 4 coils per phase in series, wired in star.
and thats just 1 test coil. Well call 'A' the number of windings you need, and 'X' the number of windings in your "test coil"
Since voltage is directly related to rpm - then you can figure how many volts your test coil would've produced at cutin rpm from whatever rpm you tested your coil at.
We'll assume your cutin voltage is 12 volts, for a 12 volt system.
So A/12 = X/whatever voltage your test coil makes at cutin rpm.
(since we know what X and whaterver voltage your test coil makes at cutin rpm then it's easy to figure out 'A')
(I think maybe Im making a very simple idea sound too complicated)
Wire gage is double the area every time you go down 3 gages in size. (or darned close). So... if you fit 200 windings of #20 gage in your test coil, and the results suggested you needed 50 windings, then you'd know that you could fit about 50 windings of #14 wire.
In a nutshell - I find the right size/shape for the coil,I wind a test coil exactly that shape/size, from that I figure the right number of windings (or try...) and then I figure what thickest possible gage of wire will fit that number of windings into the given space. Once you've picked a given size magnet and built your magnet rotors, the goal for me to to achieve the desired cutin speed with the lowest possible resistance (thickest possible wire that allows for the correct number of windings)