Go Down

Topic: Wind turbine Savonious Charger based on Joule Thief - advise for a charger? (Read 6724 times) previous topic - next topic


I built the simple Joule thief.

- Transistor : 2N2309 - 1 pc
- Rezistor: 1 K/ 0.5 W - 1 pc
- LED: white, 5 mm - 1 pc
- torroid core, 1 cm diameter, white ferrite - 1 pc
- cables- 0.5 mm D: 2 pcs @ 30 cm each, blue and white
- minibreadboard - 1 pc
- Battery NiMH  AAA, 1.2V - 1 pc (alternatively: pile AA, exhausted)

Now it comes the strange part: a slightly larger blue torroid core, of 2 cm D, 80 cm wrap same wires did not work. I went down to 40 cm wraps - no chance LED did not light.

In one example it was a picture of a blue ferrite ring - looking quite the same I used and did not work.

Searching for ferrite cores I found tons of info about a zilion types of ferrite rings but nothing clear about what the colors mean. Somewhere I found that some types have more permeability than the others, which sounds somehow trivial.

Supposing I will end up with a suitable more efficient circuit, then how I can be sure I  can order the same type of ferrite cores?

Thank you very much for any tip, kind guidance or pointing to a more friendly web-site!



Somewhere I found that some types have more permeability than the others, which sounds pretty trivial.
Absolutely not!

The permeability, loss vs frequency , saturation magnetising force etc vary enormously with different materials.

Unless you know these characteristics you can't design the inductor.

I don't know the colour codes for various ferrite materials - where did you get the cores  from? does the manufacturer publish data which would help?




I fully agree.

The seller knows nothing more than the size and the color.

The Joule thief is a demo, so no critical parameters.

But still, how can I order the same ferrite cores from some other seller or from the same seller later on, if they cant tell me more than the above?

It would be wonderful to differentiate by color, at a minimum. But would that suffice?


An alternative might be to use transformers in place of the torroids or to find somewhere a circuit using linear inductance. I would avoid that since it will certainly lead to  more complex circuits.


I am trying again.

I have no idea why it did not work first time. Just do not understand.

Do you know what the black dots mean, they are more important than core type/permeability ?
They mark beginning of the winding, I think you connect wires not correctly.


Yes, I've gotten away with just using toroids designed to be used as RFI filters, when making Joule Thiefs.  I know very little about designing inductors, and I've had little problem getting Joule Thiefs to "work".  Do they run at optimal efficiency?  Who knows.  When I'm doing this stuff, I'm in full-on hobbyist mode  :)

So, like Ted said, you probably didn't get the wiring right.  Just switch the wires on the primary, and try it again [or on the secondary].  I.e., make sure it's hooked up like the diagram Ted posted.  Also, I've found it seems to work best when the turns ratio is 1:1.  Example: if there are 10 turns on the Primary, then make sure there are 10 turns on the Secondary.

Also, it helps if the toroid is meant to to be used for RFI filters with a max current flow greater than the kind of current the Joule THief will be running at. If it's just driving a simple White LED, then we're talking around 20 to 30mA.  Also, the toroid should be able to support frequencies up to a few MHz.

Now, I'm sure my more inductor savvy piers are going to jump all over me [and go ahead], but I'm just trying to give you "seat of the pants" hobbyist pointers [you know, like, "hey, this toroid that I bought from Electronic Goldmine [with no discernable parameters other than it's for making RFI filters] looks big enough for this application, lets give it a try!"
"It's a big galaxy, Mr. Scott"

Please DON'T PM me regarding what should be part of the Public Conversation -- Let it ALL hang out!!
Unless, of course, it's to notify me of a mistake.


I wrap the 2 wires (blue and white) together, in the same way in the case of white ferrite and blue ferrite.

I connected the white wire of one side to the blue wire from the other side - this is the common end. Then I used the remaining wires to connect to the circuit as in the schematics.

I unwrapped and wrapped again starting opposite direction. I used grey ferrite core. I switched the ends. No way. Silent LED.

One moment I thought there is very low inductance in blue ferrite (for unknown reasons), so perhaps more winding would help.


The method is here (but it is detailed in many other web-sites)


and a more kids friendly desccription here:



Here it is the pic:

- background: the circuit which works ok
- grey ferrite ring - which does not work (wired)
- blue ferrite ring which does not work

The length of the wires in blue and grey cases were 80 cm and then 40 cm.

The length of the wires in the white ferrite ring 30 cm.




Thank you for pointing me to the tutorial!

It is about a fixed H-rotor (giromill). Effective in high speed wind and the pitch can be modified automatically with the wind-speed. Some designs cant ensure starting on its own in low speed wind.

The person in the video used a direct coupling to the generator, which can be beneficial in terms of efficiency when winds  are high-speed.

The interesting point is how the author shaped the wings to mimick an aerodynamic foil.


Can I use a 4.7 uF cap?


@ted - genial idea! Thank you very much!

My 4.7 uF cap got lost. So I use a 10 uF cap.

Mysteriously, the LED lighted! (she happy, me happy :-)). Using both the white ferrite ring and the blue ferrite ring.

With the cap, the white ferrite circuit seems to be less brightly than without the cap.

I didnt measure the current in both cases (with cap and without cap), so maybe is just an illusion.


Ok with the classical and simple Joule thief, magnetic field etc. I can explain that, first to myself.

But I cant figure out what the cap does in this case. My first thought is that it changes the frequency to fit what the transistor can commute. But I maybe wrong.


Phase two: generator.

Firs fail: the toys motors cant supply the current I need to light the LED in the Jule thief  circuitry I made.

I mounted a small propeller on a 3V motor and blow it by a hairdryer (a big one, really). Although the propeller spins very rapidly, like in the case of the motor being feed by energy, the LED did not light.

By multimeter to motor, the current supplied was up when the motor starts and stops - which I suppose has something to do with current pulse /induction.

I could not reproduce this behavior with the LED, however.


Next attempt: PC fans. I shall check when I will find the fans I brought in my closet.

A more sophisticated attempt in plan: brushless motors. Using my dear Shotkeys to get DC. :-).

Go Up