12 Watt Turbine

Design and construction of a 12 Watt wind turbine

The following pages describe the construction of a small wind-turbine suitable for trickle-charging a car battery.

The Goal: 12V, 1 Amp (12watts).

The 12 Watt Turbine

The first incarnation is based on a 100mm rotor, like the prototype except I beefed everything up somewhat. Using such a small rotor does limit the amount of power that can be generated but a baby step is better than lying on the sofa eating pies.

The Frame

After trying many configurations, it seemed the easiest way to mount the turbine was to use a couple of Pillow Block bearings on a frame made from steel box section. The first incarnation of this is shown below. My father-in-law came up with this design and put together this frame for me. I guess it could be made by just bolting things together but it was quite fun grappling with an arc welder for 'sticking' bits on here and there which I had to do as the project progressed.

It may look strange that the axis is mounted off-centre but there is a method in this madness, to do with auto-furling i.e. a means of turning the blades out of the wind if the wind force gets too strong. More on that later.

The Shaft

I decided to use a 10mm shaft to mount the rotors and managed to get hold of a 10mm silver steel shaft just for the purpose. 10mm bearings in pillow blocks are a little hard to come by (as they are quite small) but I did get a couple so was set to move forward.

The Rotors

After my disapointment with the performance of the prototype, I reasoned that as voltage generated is proportional to the rate of change of flux, having more magnets will will give a greater rate of change. I managed to squeeze 12 magnets onto each disk rather than the original 6.

As for the rotor disks themselves, I used a couple of scavenged 100mm angle grinder tile-cutting disks which are thicker and more robust than what I was using before. The disks also have a rim at the edge which helps keep the magnets in place (after my unfortunate 'garage-incident' earlier. As these disks have a large 22 mm hole in the centre and I was intending to use a 10mm spindle, I bolted a steel sprocket (with a 10mm hole) onto the rotor to allow it to be mounted.

This is the plan of the rotor. Each magnet being 19mm in diameter. You can see that the rotor is pretty small by comparing it to a 20 pence piece.

The picture below shows one of rotors on the shaft and attached to the steel frame. The magnets have been super-glued in position.

The sprockets do not have a method to attach to the shaft and they are often glued as the clearance is pretty tight on the shaft. I drilled and tapped the sprocket to allow adjustment with a grub screw for initial testing.

The Stator - Coils

The coils need to be pretty consistent in size and shape so it is worth knocking together a coil winder. Then there is the problem of deciding the wire gauge to use. The thicker the wire, the more output current , but you can fit less coils in the available space. Going for more turns gives more voltage but the wire will need to be thinner so can produce less current (and hence less power).

I opted for 24 SWG wire as this has a maximum current carrying capacity of about 4 amps which is four times more than my planned maximum. It also allows a greater voltage to be generated as more turns can be fitted into the available space.

Due to the small size of the machine, I was limited by the size and number of coils I could use. I opted for six coils and decided to make a single-phase design as the amount of voltage generated for each coils would be quite small so they would need cascading in order to produce 12 volts.

Here is a picture of one of the finished coils after taking it off the coil winder.

I actually made two stators. The first one used 100 turns of wire per coil which seemingly just fitted in the space available. I was disappointed with the results of this stator so made another one with 300 turns/coil which is the one described here.

The picture below shows that these coils are really squeezed to get them in space available. In fact, I squeezed the coils in a vice to make them more ob-long than round so they would fit together better.

In order to get some idea of amount of voltage that would be produced by the machine, I decided to test several coils on a test jig scavenged from my original prototype design. This allows me to tape coil(s) to it and attach a rotor and measure the voltage produced. I used one coil and one magnet on a 100mm rotor (modified from the prototype). These are the results obtained when rotating at about 220 rpm:

These are not exact figures but show that the voltage generated is proportional to the number of windings. When I added a second magnet, the figures doubled. Adding a second rotor doubled the figures again.

Referring back to the plan I had made for the rotor, it became apparent that could just squeeze 6 coils with 300 turns onto the disk, or 5 coils of 400 turns. Unfortunatly, as the coils become larger, several magnets (of opposing poles) pass over the magnet at one time and some flux cancalling starts to occur meaning that ten magnets did not give ten times the output of 1. For the 300 turn coil, the relationship held true.

I decided to go for six coils of 300 turns which should give:

12 magnets * 75mv = 900mv per coil per rotor.

With 6 cascaded coils, this will give:

6 * 0.9v = 5.4v rms

so for two rotors (ie 24 magnets):

total output at 220 rpm = 2 * 5.4 = 10.8v rms

This is in the right ball-park as once the voltage is rectified to DC, it will be in excess of 12 volts.

10-Jul-2007