Reworking the rotor was the hard part. Assemble the rotor, stator and case, thread your red, white, blue and black stator wires up through their respective holes and re-install the rubber terminal insulators. Rotate the rotor by hand to make sure that it isn't dragging on anything, and if it is take things apart and correct the problem.

You should notice that the rotor is a bit harder to turn since the rebuild, and you may want to put the pulley back on, at least temporarily. This increased resistance is from eddy currents in the stator laminations that are produced by the now much stronger magnetic field. To get an idea of exactly how much more current this machine is now capable of, get a friend to short all the stator leads together while you turn the rotor by hand. You should be very pleased to find it MUCH more difficult to turn while the leads are shorted, even at this very low RPM. That is because when the leads are shorted current is flowing, and that takes power to produce.

If an oscilloscope is available, connect the black stator lead to the ground of the oscilloscope and any of the colored leads to the input. Give it a spin and you should see a nice sine wave on the scope. Do each of the colored leads in turn and you should see a similar sine from all of them.

If you find that shorting the output leads makes the rotor much harder to turn, and all of the colored output leads give consistent sine waves, and you can test each stator lead against the case ground with a multimeter and you find no shorts, then the odds are good that you have just made yourself an alternator with a permanent magnet field. Congratulations! You aren't done yet, though.

There are actually several ways to use this alternator, but let me describe the simplest. Replace the diodes in the alternator (I'm assuming they are good), but before you do you need to decide if you want to go to the bother of isolating the output from the case. At this point the stator leads should be isolated (you tested this, right?), but the diodes will short the negative output to the case if you just bolt them back on. In most cases this isn't a problem, but if it is for whatever system you envision, you can slip some insulating material between the diode bridge and the mounting bolts, or simply mount it in some other fashion. I leave that problem to you. Assuming, then, that you are not concerned with this issue, mount the diode bridge as before and replace the metal cover on the back end of he alternator. This part is now done!

If you now mount some blades on this alternator, put it up on a pole and attach a battery you will be open for business . . . assuming you have some wind. As long as the alternator is rotating at a rate too low to overcome the diode voltage drop and the battery voltage there should be very little load on the alternator and it should spin up quickly. As soon as the output voltage from the alternator exceeds the battery voltage and the diode voltage drop, current will flow into the battery and the alternator spin rate will be constrained by this load. Thus such a system becomes self-regulating with one exception: if the battery becomes completely charged the battery voltage will rise and the energy from the alternator will start to break down the water in the battery into hydrogen and oxygen, thus "burning" it off. One solution to this problem is to use a voltage comparator to detect when the battery voltage exceeds, say, 13.8 volts, and close a relay that applies an additional (sometimes called a "diversion") load to the system sufficient to absorb this extra power. Something useful would be good to use for that load, like a hot water heater, some additional lights or a pump. An implementation of this solution is below, but first a better one . . . perhaps.

Using an Arduino for System Control

Arduino Boost Circuit

(This circuit diagram may be downloaded in pdf format here.)

(The Arduino sketch I used may be downloaded in txt format here.)

I have long been annoyed by the fact that winds too low to generate enough voltage to charge a twelve volt battery directly are effectively useless in most cases. So I decided to look into designing a boost circuit to correct this problem. I had an old dead switching power supply from a PC that had what looked like useful inductors in it, so I picked the largest and stripped it out to experiment with. I had also bought an Arduino controller to play with, and was very impressed with the possibilities of using the Arduino to control a windmill output. My understanding at the time was that designing boost circuits was very tricky, and I'm sure this is true, to do it "correctly," but in this case there was little to lose because of that "nothing" power output below twelve volts. Somewhat less than perfect is still a lot better than nothing.

So, why not try it?

With this sophisticated philosophy firmly in mind I did the following. I wired one end of my inductor to the drain of a MOSFET transistor I had lying about, and grounded the source. The gate I wired directly to an Arduino PWM output because this particular MOSFET has a logic level input. I then wired the positive end of an ordinary "D" cell battery to the other end of the inductor with the negative end to ground and fired up the Arduino in PWM mode. The result astounded me! An oscilloscope probe on the drain connection of the MOSFET was showing a train of 70 volt pulses!!!

Seventy volts from 1.5 is definitely boost. Thus encouraged, I added an output diode to the MOSFET drain and connected a mostly-charged lawn mower battery to this output. So now I had the circuit indicated by Q1, L1, and D3 on the diagram above, plus a "D" cell feeding inductor L1 and a lead-acid 12 volt battery attached to D3.

I started up the Arduino again and watched as the 12 volt battery voltage slowly climbed up, and the 1.5 volt "D" cell voltage went down. This very simple circuit sucked the "D" cell down to 0.5 volts and moved this power into my 12 volt battery. It was continuing to move power when I shut it down from boredom.

I'm putting this information here for anyone to use in any way consistent with open-source ethics and licenses. That is, if you use it, you must share what you develop with everyone else at no charge for the information part of it. It is my hope that someone or some group of someones will run with this and develop a usable controller for the rest of us. That controller could have some very useful features, such as intelligent battery charging and intelligent loading control for the windmill. The remainder of the circuit above is simply two voltage and current monitoring circuits for input and output.

Those of you who have read some of my books or rants, especially Hubris Ark, will know that I have a very dim opinion of the United States government, and a deep concern about climate change and what now looks like an inevitable extinction event. If our species is to survive it will because people like you advance the science of alternate energy sources in spite of our coin-operated government whores, and this could be one way of many that that happens. Our government will only advance the interests of the fossil fuel industry, because that's who pays them the most. The fossil fuel industry does not want you to develop your own energy sources. They would force you to buy the rain, if they could. Which is reason enough to do it, all by itself.

A Diversion Circuit

Below is a circuit diagram for a simple voltage regulator diversion circuit you can construct from Radio Shack parts and an automotive relay. I haven't tried this circuit myself, so let me know if you try it and have any problems with it so I can make any necessary corrections (and I'd like to know if it works too).

voltage regulator

Here is a list of the major parts, all available from Radio Shack except the relay. You should be able to get a relay from any auto parts store.

  1. LM339 Quad Comparator, RS cat# 276-1712 $1.29
  2. TIP42G PNP transistor, RS cat# 276-2027 $1.59
  3. 2N3909 PNP transistor, RS cat# 276-1604 $2.59 for pack of 15
  4. 1N4733A Zener diode, RS cat# 276-565 $1.29
  5. 1N4004 Diode, RS cat# 276-1103 $0.79
  6. 10 Kohm 15-turn trimpot, RS cat# 271-343 $2.59

In addition to the above, you will need some resistors as shown, a circuit board to build the circuit on, a box or whatever to put it in and protect it from the elements, and if your relay coil has a DC resistance less than about a thousand ohms you will need a heatsink for the TIP42G.

After you build the voltage regulator circuit you will have to adjust it to operate the relay at the voltage you choose, probably 13.8 volts. If I didn't screw something up here, the circuit is supposed to work like this: The top of the 5.1 volt zener diode is a voltage reference for the positive input of the LM339 voltage comparator at pin 9. As the battery voltage increases a fraction of that voltage will be present at the wiper of the 10 Kohm trimpot, because the trimpot and the 5.6 Kohm resistor in series with it form an adjustable voltage divider with a range between ground and roughly 2/3 the battery voltage. The pins on the LM339 number counter-clockwise from the top left on the chip (looking down on it) so the trimpot wiper is connected to LM339 pin 8. Once the voltage at LM339 pin 8 (the inverting input) exceeds the voltage at pin 9, the comparator output at pin 14 will change state to a "low" value. The output of this comparator is actually an open collector type, which means that it only has an active low, no active high. Anyway, when the output at pin 14 goes low it will pull the base voltage of the 2N3906 transistor toward it's collector voltage through the 6.8 Kohm resistor thus turning it "on." The 2N3906 transistor is connected between the Base and Collector of the TIP42G, so when the 2N3906 becomes conductive, that allows current to flow from the Base to the Collector of the TIP42G as well, so it also becomes conductive. This configuration is known as a "Darlington pair," by the way, and is used to increase the current gain. When the TIP42G is conductive, current will flow through the relay coil, thus energizing it. The 1N4004 diode serves to snub out the inductive kickback from the relay coil when the TIP42G turns off.

This circuit can be set by powering it up with a DC voltage source set to the set point you are trying to achieve, for example 13.8 volts, and adjusting the trimpot until the relay just operates. A convenient source of 13.8 volts (or something close) is a fully charged automobile battery with the engine running to maintain that charge.

As designed above, the inherent and very small hysteresis of the LM339 comparator is used, but this may not be enough. You will know this is the case because the relay operates too often. This insures the highest accuracy, but wastes energy. Anyway, the fix for this problem, if it arises, is to add two more resistors, with one replacing the wire from pin 9 of the LM339 to the top of the Zener diode and the second connected between pin 9 and pin 14 of the LM339. I would make the first resistor (between pin 9 and the Zener diode) about 1 Kohm, and the second one about 1 Mohm as a first guess.

Here's what you are trying to do. Pin 14 is the output of a comparator in the LM339, and pin 9 is the "positive" (non-inverting) input to this same comparator. Pin 8 is an inverting input to it. The way it's set up in the circuit above, as the voltage on pin 8 increases (inverting input) over the voltage on pin 9 the output goes low, operating the relay. The load from the relay draws the battery voltage down slightly, dropping the voltage at pin 8 below the voltage at pin 9 and the relay is released. The two new resistors make a voltage divider because they are in series, with the node between them being the input on pin 9. Now assume that the input voltage on pin 8 has just gone above the voltage on pin 9 and caused the output to go low, operating the relay. This output voltage (now low) will be applied through the new voltage divider to pin 9 and pull it down somewhat, thus making the magnitude of the voltage on pin 8 greater relative to pin 9. Since the difference between these voltages is now greater, a greater change in battery voltage will have to occur to cause the system to cycle. The amount it will be greater is determined by the ratio of the two resistors, which in this case is 1 to 1000. So, if the battery voltage is right at 13.8, a thousandth of that is 0.0138, so you've just added 0.0276 volts of hysteresis (0.0138 volt higher when the output is high, and 0.0138 volt lower when the output is pulled low). If this is not enough hysteresis just reduce the resistance of the 1 Mohm resistor or increase the resistance of the 1 Kohm resistor until it is. Either way will change the ratio.

That's all I have so far on this subject, I hope you find it useful. If you spot something stupid that I have done please use the contact link below and let me know so I can fix it! Also, if anyone wants to offer to jump in and provide machining or make blades or mounts or whatever for those not blessed with the right equipment, let me know and I will add that information to the "Resources" section. What I have in mind is your name, address, contact phone, what you want to do (like machine the rotor claws) and how much you want for the service. I'll also make some space for user comments on the services you offer to provide feedback for others.

I've been having quite a bit of fun with this, I hope you do too!

Identifying the Appropriate Alternator.


Modifying the Stator.

Modifying the Rotor.

Assembly and Voltage Regulation.

Resources and Services.

Purchase our Special Magnet.

Seal 2.png (57,716 bytes)

Moon Bug
Alternate Power
Surplus Equipment