S4w's Wind Generator Notes

When one is building an alternator for a windmill there is always a compromise that must be made in order to get the desired results. One of the most important decisions that one must make is to choose a suitable cut-in speed for the propeller that is going to be used.

If the cut-in speed is too low, the alternator will stall the propeller and not much power will be gathered from the wind. If the cut-in is too high then a lot of power is gained at high winds but none will be gained in low winds. The following graphs contain theoretical power curves into the batteries. They are all represented as straight lines. In reality the power going into a battery bank most of the times is an exponential function but for illustrative purposes straight lines will be enough.

Below is a graph of a setup that would stall the propeller:

As you can see, the alternator output power curve is above the propeller's power curve. If the cut-in is too low, the alternator will try to extract from the wind energy that the propeller can not supply thus the result is a stalled propeller.

A better choice is shown in the picture below:

This propeller will not stall as the alternator power curve is just below the propeller's power curve and will produce satisfactory results in light winds. It can be achieved when the alternator has considerable internal resistance or when additional line resistance have been added. The drawback of this
setup is that the propeller overpowers the alternator. This situation is very disadvantageous in high winds because not even 50% of the power is being gathered and not to mention very dangerous too. The alternator can burn in high winds and leave the propeller without load probably causing a quick self
destruction.

In the other hand a fast alternator (high cut-in speed) is not likely to burn out in high winds. It will be more capable of gathering much better the power available in high winds as shown below:

The only bad thing about this setup is that there is no power collected in low winds and that kind of wind speed is more abundant. That is the reason many people choose to wind a slow alternator completely ignoring the other facts.

I have come with an implementation to alleviate this problem from one of Flux's ideas. The device is basically a boost converter combined with a high cut-in, efficient alternator.

The boost converter will increment the voltage of the mill by a set amount selected by the user. As the current entering the batteries grow to a set point also selected by the user, the booster starts to decrease the PWM gradually when it hits this current set point. From this point on the output current will remain constant until the boost converter finally shuts down and normal windmill operation is resumed. This is all a smooth transaction for the windmill, not like star to delta switches.

An example of this is shown below:

The point where the alternator power intersect the booster's output power should be where the booster turns completely off. This ensures max power into the battery bank. Now you have the benefits of both worlds less the boost converter efficiency in the low side.

The slope of the boost converter's power curve is mainly determined by the chosen current set point. Less current will cause the output power curve to have less slope and more current will cause a steeper curve. For this purpose one wants the boost converter's slope to be less steep than the alternator's slope. If not, one would be forced to select a higher cut-in thus loosing a bit of power in the low end.

There is no replacement for wind power here. This booster will NOT make you any more power if its not available in the propeller. NO MAGIC HERE. Note that all the power lines are always BELOW the power available from the wind.

Maybe there is one case where you want to stall the blades and this is at very high winds when you want to protect the windmill from damage.

I made a few boards to prototype this device and the whole idea works very well with efficiencies up to 96%@24V. The boards for sale should be in by May 2nd 2008 and ready for sale starting May the 6th.