Designing, and Building My Own Wind Turbine
Field Notes from Irricana, Alberta, Canada
by Steven Fahey



24 September
2008
"Lights ON"
24 October
2009
"Repaired"
14 November
2009
"Battery Test"
15 January
2010
"Capacitors"
15 February
2010
"Solar Panels"
5 June
2010
"Water Pump"
31 December
2011
"Maintenance"
HOME
PAGE

DIARY ENTRY FOR 10 JANUARY, 2010

Boosting the Turbine Output in Light Winds

Lately, I have been interested in ideas concerning ways to tease a bit more energy out of wind turbines. I used to believe that adding anything to the generator would just cause problems. Reading about these ideas on the internet has made me consider their merits. Plenty of claims, a bit of evidence, not much proof.

Gordon's Cap Doubler Circuit
from "The Back Shed Forum"

Others have tried similar things, like
user: DamonHD on Fieldlines.com

and
user: Spacejunk on Fieldlines.com

There seemed to be some promise, and more than one person claiming practical benefits, so I didn't give up on the idea. I also realized that some people don't have the time or presentation skills to put together an explanation that others can easily follow. That stuff was drilled into my head in college, so maybe I can offer some value by trying to describe the concept and principle myself. I am motivated to work on this by the rather lack-luster performance of this generator, replacing the old one. I don't have another replacment generator, so I'd like to make the best of this one.

I started by drawing circuit diagrams. My trouble with reading other people's circuit diagrams, I don't really understand them unless I draw or build it for myself.


A "Doubler" circuit, like Gordon's.

A "Tripler" circuit, based on the doubler.

Spice Simulations of Capacitor Boost Circuit

Spice is a program that simulates electric circuits for you, and gives you detailed results of the current flow, even voltage spikes. Spice allows me to draw any circuit I can imagine and see what it does without heating up the soldering iron. There are different versions of Spice. I found "5Spice" which suits me fine.

[RANT="ON"]
Caution: When using computer models, the law of GIGO applies. If you don't believe in "Garbage In = Garbage Out" then go ask Airbus, CERN, or NASA. They, too, can tell you that even with the most sophisticated computer models, reality can bite you in the A**. And there are people working for them that are smarter than you and me!
[RANT="OFF"]

Anyway, I did start with Spice models of capacitor doubler and tripler circuits. I drew them "in between" the existing wind turbine and 3-phase rectifier that I already have. Here's what they looked like:

Spice diagram of a Voltage Doubler, using single phase bridges intended to boost the current charging batteries by a 3-phase wind turbine. (see the full size diagram)
Spice diagram of a Voltage Tripler, using single phase bridges intended to boost the current charging batteries by a 3-phase wind turbine. (see the full size diagram)

The diagrams were getting crammed on the screen, as you can see. The free version of Spice allows only this much space for a diagram. I'd have to register the program to use larger circuit diagrams. There were several benefits to using Spice to model the circuits before building them:

  • I can vary many parameters to find good values of capacitors to try,
  • I can look for voltage spikes,
  • The program can graph numerous values on the same chart, telling the "story",
The little "hexagons" in the diagrams show where the voltage and current measurements are made. To "turn off" the capacitors, I disconnected the wire between "Boost" current measurement and the battery. The generator is modeled as 3 voltage generators, each out of phase. I selected 28VAC peak and 3.333 Hertz to equal roughly 200RPM on my generator. This is just about the cut-in speed. The resistance and inductance are the values I measured with the generator on the bench. I'm sure that the stator resistance goes up as it gets hot and the inductance changes, too as the current flow causes back EMF and the magnets apply magnetic fields in the stator, preventing linear inductance to occur in a simple coil. Nothing's simple. This is just a starting point.

I ran the simulations with as much numerical accuracy as possible (1msec intervals), allowing me to see the little voltage spikes in these circuits. One thing I wanted to see in Spice was how much the diodes switching ON and OFF would cause voltage spikes, compared to the simple battery-charge circuit. It turned out to be true, there can be surprisingly high spikes.

Once the circuits were drawn up I was able to run the simulations again and again with different values of capacitors. The first runs were actually with the boost circuit disconnected, to get a "reality check" with normal conditions. Then I reconnected the boost and each subsequent run had different values for all 6 capacitors:

Or you can see them in sequence by clicking here (509 kB).

I think you'll agree that it looks more like a mess than an increase in power. Without the caps, the battery charging current was relatively steady, with a ripple you can discount considering the long wire runs and so on. Small capacitors did not have a measurable effect on the current, while the higher values of capacitors produced large fluctuations in current every cycle or more. Even by integrating under the curve, I find the same average current in most situations (about 2.5A) whether it's the doubler or the tripler.

Noise and spikes are coming from the rapid and repeated switching of current through numerous diodes. I can only imagine how much RF interference this would produce. In addition, the voltage spikes that I expected are there, prominently. I found some higher than 100 volts in this 28-volt battery system. Whenever I disconnected the output of the capacitor circuits and ran the sim, I found that the open-circuit voltage across some terminals was higher than 50VDC. Whatever I build, it should be enclosed in a metal box!

The simulations of both the tripler and the doubler have failed to show a cause for an improvement in power collected at low wind speeds, as reported by their proponents. There is no reason to give up yet, though. Spice is a computer simulation, and as I said above, failure of a computer model doesn't mean it doesn't happen in the real world. Perhaps these Spice simulations are total garbage in the wind turbine generator situation.

It sure would have been nice to have a picture of what's going on inside, though...

I decided to test the concepts for real on my turbine to see if there was any point pursuing these ideas any further.