Flyback converter for wind power

oops.

I think I must have had a rush of blood to the head at the time I came up with the flyback converter idea!  A better alternative is to use a boost converter while the alternator output is below the battery voltage and simply shut it down once there is sufficient voltage available for battery charging directly from the alternator.  There will be fewer  losses that way, instead of trying to pass all available power through a converter all of the time.  The boost converter also has the advantage of no spikes at the switching transistor as it's these that are caught and provide the output, it also has one less winding on it's transformer.  The alternator voltage can pass straight through the boost inductor and output diode to charge the battery once the voltage reaches the charging level - at which point the converter can be shut down.

Prawncrackers,

That is a useful circuit for powering a high power led from mains voltages.  If you need a simulator then I can recommend LTspice, it is free, and works very well. Saves a lot of tinkering time.

From here: w w w .linear.com/designtools/software/?gclid=CMOE6frq5rkCFWOWtAodcHQA_g#LTspice

You'll need remove the spaces!

I may need a snubber as the transistor I'm using is fairly low breakdown voltage.  I didn't model the leakage inductance of the transformer.  I was just going to try and have a look on the scope to see how bad the spikes were.

you can try this self-osc converter for a LED

http://danyk.cz/izled.png

scale parts for lower voltage higher power operation particularly the 470k. batteries connected parallel to the tant. caps, the LED replaced by a suitable resistor divider to provide feedback for 13.8v output. the snubber circuit is i guess optional. this circuit looks simpler...

i dont have a simulator so i wouldnt know hahaha

This is the circuit that I'll be building except the switching transistor will be a 2SD882 because I have a lot of them retreived from junk boards.

I know visitors can't post links but :-)  w w w.facebook.com/photo.php?fbid=611368665597778&l=2794c21bd7

(ok, so I find bugs in computer stuff too!)

It's on my Facebook cover pictures (always public)

i did a quick skim on SGS application note. it looks like an interesting circuit and you can even strip this down into something more utilitarian.

like direct voltage feedback, you dont need (galvanic) isolation from the primary anyway. minus a transistor, minus the optocoupler, minus the TL431.
IMHO a simple resistor divider at the output can feed Q2. and youd only need one secondary.

eliminate the overcurrent limiting circuit (R7?)

cheers

I take your points prawncrackers.  There is a problem with the efficiency which would probably limit the use to low powers making it only suitable for charging batteries and not providing main power.  I will build my circuit and see how it performs with a rectified stepper motor driving it (driven by a drill for testing).  I can't see it being destroyed, but will try anyhow.

The charge pump idea sounds interesting and pretty simple to implement (there are some designs here on sm0Vpo's pages).

Trying to get power from a varying supply voltage looks like it may be a problem.  Problems are there to be solved however...

hello Al,

im not an SMPS expert but i think that is the downside of wide-input converters, most commercial designs are even only rated like 50% of their dissipation and stuff... and to think many discrete self oscillating-designs are more prone to dysfunction when input voltage varies.

i have a friend, who fancies [has the money] this free energy thing, although solar, uses a charge pump rather than a converter. i dont remember the exact circuit but its a simple voltage doubler that activates when the source voltage dips below, say, 12v. the losses are just about two diode's drop.
he doesnt mind the efficiency though the panels only reach about 20v during the brightest of the day.
if wind turbines output pulses then maybe it would make charge pumps easier to implement.

Sorry for the spammy username of power1, you can call me al if you wish (that one is already taken).

I'll explain why using a flyback converter.

1) can convert from higher or lower voltage than the output.

2) Only 2 active devices needed.

3) All amateur wind generators (that I've seen) use a permanent magnet alternator - voltage varies with speed of turning, useful power still available at low voltage outputs.

4) power throughput rises as input voltage rises (perfect for getting as much power as possible from varying wind speed)

I've recently designed a small Flyback converter using LTspice IV and it simulates well - a few changes need making to compensate for the parts not available to use in the model.  According to the simulation it will be able to provide charge to a 12V lead-acid battery from a voltage input of between 5 and 17 Volts at a source impedance of 5 ohms (quite high).  The maximum input voltage could be raised by using a more suitable transistor than the 2SD882 that I have available, and no problems should be encountered with voltages of up to 35V on the input.

There is no need for a pulsating output, just a diode - the battery being charged taking all the current that it can.

The only downside is that it is only 65% + efficient so some power is wasted, wind energy being free this should not be too much of a problem but, of course would like to improve on it.

For anyone wanting to design a self oscillating flyback converter a useful reference is SGS-Thomson app note CD00066269.

Can't access Harry's homepages at the moment (24/11/2013), seems like the net connection is down somewhere.

Hi Power1,
you should decide to use either the buck or boost topology of the converter, i.e. the input should be always more resp. less volts than the battery. Consider a generator giving (say) from 18 to 60 V. If its voltage drops under the lower limit, the circuit may switch to another converter, producing the short current pulses.
The common solution is to control the excitation current of the rotor to keep the output nearly constant. E.g. the alternator in a car gives nearly constant voltage from 800 to 6000 r.p.m. or more. The produced current does not flow through any converter.
IMHO the circuitry can be designed using discrete parts only.

BR from Ivan OK1SIP

I've been experimenting with making alternators and would like to harness some wind power for charging batteries / running lighting (180 Watt max).  Having been through everything I can find online it seems like a not too difficult problem to design a suitable windmill.  However due to varying wind speed the output of a wind generator will likely be 'all over the place'.

My idea is thus: a permanent magnet alternator that outputs between 10 and maybe 60 volts (depending on wind speed) coupled to a constant power flyback converter - the idea being that the output will always be at (say) 13.8 Volts (and a suitable current) if there is sufficient wind energy to provide it.  If there is insufficient energy then I would like the output to still be available in short pulses (rumoured to help de-sulphate lead acid batteries).

If there is a lot more energy available then the input (voltage) to the converter will rise but the output remain constant.

Does anyone think that this idea is any good? Is it possible to make a converter using commonly available parts - only transistors / resistors / capacitors and a home wound transformer - no Ic's or processors (for reliability and ease of finding parts).

Feedback greatly appreciated.