Big Joule Theif

[slayer007]

Hi Slayer007 thanks for response. My JT did not work don't know why. I'm using a 2" torroid as you did but nothing happened.
I was using electrolyte polorized 10uf but nothing happened. Maybe it's not the capacitor problem it's the winding problem for JT.
Any help please?
Thanks

Guruji. if it's not working at all you might have your coil wired wrong.

Or I should say the wires comming off your coil might be wired in the wrong way.

Did you notice there should be a dot or a letter by one end of the coils in the circuit.

The dot should be the beginning of the coil the other side with no dot is the end.

You will notice the beginning of one coil will go with the end of the other coil.

So coil 1 with the dot should go to positive of the battery.
Then coil 2 with no dot should go to the positive also.
Coil 2 is the one that should have the 1K resistor on it.

I hope that will help I don't know if I explained to well.

EDIT.

Thanks for showing your circuit Tectalabyss it looks interesting.

[Xenomorph]

@Guruji:
If you havenīt marked them somehow prior to winding the coil, you can determine the wire ends that belong together by measuring their resistance with a multimeter.
If it shows a low resistance then the ends must belong to the same wire.
If there is no reading (-1), then the ends probably donīt belong together or have been broken maybe.
Then you just have to follow the schematic wiring like Slayer007 described.
Hope you get it working now!

[Guruji]

Thanks for helping guys. I did the 20 turn coil with two coils twisted then going around the torroid. Is that ok what it should be done?
Thanks

Guruji. if it's not working at all you might have your coil wired wrong.

Or I should say the wires comming off your coil might be wired in the wrong way.

Did you notice there should be a dot or a letter by one end of the coils in the circuit.

The dot should be the beginning of the coil the other side with no dot is the end.

You will notice the beginning of one coil will go with the end of the other coil.

So coil 1 with the dot should go to positive of the battery.
Then coil 2 with no dot should go to the positive also.
Coil 2 is the one that should have the 1K resistor on it.

I hope that will help I don't know if I explained to well.

EDIT.

Thanks for showing your circuit Tectalabyss it looks interesting.

@Guruji:
If you havenīt marked them somehow prior to winding the coil, you can determine the wire ends that belong together by measuring their resistance with a multimeter.
If it shows a low resistance then the ends must belong to the same wire.
If there is no reading (-1), then the ends probably donīt belong together or have been broken maybe.
Then you just have to follow the schematic wiring like Slayer007 described.
Hope you get it working now!

[slayer007]

@ Guruji


Here is a video on youtube that will help you with making a basic joule thief.

YouTube - Make a Joule Thief

Any npn transistor should work.
The led will only light one way to.
The long side of the led goes to the collector and the shorter end goes to the emitter or ground.

Once you get the basic JT going you will have no problem with the other circuits.

[lanenal]

@shlodo:
I would suggest a very rough way of doing this, without much expensive equipment, but the measurement might just be good enough given that the COP is significantly larger than 1.

Simply use a resistor (make sure it has the right wattage rating) as a load to the battery -- this way you can measure the voltage over the resistor every 15 min, then use an excel sheet to find out the work output -- the relative error in this method should be all right: given that the voltages over time (not the rest voltages, but the runtime voltages) does not change more than 10%, the maximum relative error would not exceed (1+10%)^2-1=21% (as Watt=V^2/R) even if you just measure runtime voltage only once. So, if your measured COP is something like 5, you know that the actual COP would be better than 5*(1-20%)/(1+20%) = 3.33 in the worst case (that is, output is off by 20% more and input energy is off by 20% less).

Another easy thing to do: with lead acid batteries, as long as you don't go below 50% of its capacity, the rest voltage changes are almost in linear relationship with its output work/energy, then your initial way of doing business is quite OK to my eyes. See my reference page in the previous post for lead acid batteries. So, if you are using lead acid batteries, play with them using my method above, to validate the linear relationship between voltage changes and energy output.

Hope that helps.

lanenal

@lanenal
You are quite right, I couldnt believe I overlooked that! I realised not long after I made the post. But I have done many tests and they have all been positive. I will have to repeat the test with the same batteries...
I like your idea of finding emperical functions between V0 and V1. Ill have to look into it. Ive been trying to come up with a proper testing method.
I noticed Bedini uses a battery capacity meter which measures in % how many amp hours are in the batteries.
I saw a simple one at Jaycar.com.au for cheap too, but it reads in increments - 70-80%, 80-90% etc.. which I thought wasnt very accurate.

-Has anyone come across these??
-Can anyone out there suggest a proper testing method to compare the work we can do with the charge battery vs run battery??
-To accurately measure current consumed would i need to use a true RMS meter?

Any help would be much appreciated

Edit: explained the math in more details.
Edit: Sorry, seems I have to repost this to get my edit through.

[lanenal]

A quick enhancement: By using sqrt(Vmax*Vmin), the relative error can be reduced by half. Since a discharging battery should drop in voltage, so Vmax is the runtime voltage at time zero, and Vmin is the runtime voltage at time of finish.

So, here are the things to measure:
1. starting rest voltage: Vrest0
2. starting runtime voltage: Vrun0
3. ending runtime voltage: Vrun1
4. ending rest voltage: Vrest1
5. runtime duration T

Suppose you know the resistor R. Let Vrun = sqrt(Vrun0*Vrun1). Then you have: Output(Vrest0,Vrest1)=T*Vrun*Vrun/R=T*Vrun0*Vrun1/R.
The maximum relative error of the estimated Output above should be (Vrun0/Vrun1 - 1), as the actual output is within [Vrun1*Vrun1/R, Vrun0*Vrun0/R].

@shlodo:
I would suggest a very rough way of doing this, without much expensive equipment, but the measurement might just be good enough given that the COP is significantly larger than 1.

Simply use a resistor (make sure it has the right wattage rating) as a load to the battery -- this way you can measure the voltage over the resistor every 15 min, then use an excel sheet to find out the work output -- the relative error in this method should be all right: given that the voltages over time (not the rest voltages, but the runtime voltages) does not change more than 10%, the maximum relative error would not exceed (1+10%)^2-1=21% (as Watt=V^2/R) even if you just measure runtime voltage only once. So, if your measured COP is something like 5, you know that the actual COP would be better than 5*(1-20%)/(1+20%) = 3.33 in the worst case (that is, output is off by 20% more and input energy is off by 20% less).

Another easy thing to do: with lead acid batteries, as long as you don't go below 50% of its capacity, the rest voltage changes are almost in linear relationship with its output work/energy, then your initial way of doing business is quite OK to my eyes. See my reference page in the previous post for lead acid batteries. So, if you are using lead acid batteries, play with them using my method above, to validate the linear relationship between voltage changes and energy output.

Hope that helps.

lanenal

[Guruji]

Joule thief

Hi Slayer I did that and saw that video too it worked for me too.
The thing on the big toroid ferrite core I did it with just coil wire maybe that was wrong I had to use normal wires.
If it's because of that I will do it again it's ok.

@ Guruji


Here is a video on youtube that will help you with making a basic joule thief.

YouTube - Make a Joule Thief

Any npn transistor should work.
The led will only light one way to.
The long side of the led goes to the collector and the shorter end goes to the emitter or ground.

Once you get the basic JT going you will have no problem with the other circuits.

[slayer007]

Hi Slayer I did that and saw that video too it worked for me too.
The thing on the big toroid ferrite core I did it with just coil wire maybe that was wrong I had to use normal wires.
If it's because of that I will do it again it's ok.

The wire should be insulated wire.
The wire I'm using is magnet wire it's insulated with enamel.

The enamel coated copper wire will work better you just need to clean the enamel off the wires to make a good contact at the ends.

It's great you have it going now.

[tjnlsn255]

Toroidal cores?

Where does everyone get their toroidal cores from to make their toroidal coils?

Thanks for any and all help with this search.....

Tj

[tjnlsn255]

Thanks Tec.... looks good!

[Lidmotor]

Solar Garden Light

I have been side tracked working with a similar circuit to the Joule Thief. It was posted by the Daftman at the "Teep" forum. It is like the JT circuit so I thought that it would be OK to post the info on it here. The circuit uses an small inductor instead of a toroid coil to generate the spike that runs the LED. I replicated his circuit and added a 'second battery' charging aspect to it. I then added automatic solar charging to it from a 'solar garden light' circuit that I found here: Solar Light This page at "talkingelectronics.com" gives a very good description of how these JT like circuits work and shows a diagram of the wave form. I found this very helpful.
Here is a video of my "Automatic solar charged Daftman LED/charger light"


YouTube - Automatic solar charged Daftman LED Light

Cheers,

Lidmotor

[Guruji]

Big Joule thief

Hi guys I am still with the problem that nothing is coming to my big joule thief by slayer007. When connecting the 20turn double wire to the 1.5v battery it should make a sound to know that it's working?.
My 480turn coil is about 5ohms and the 20turn ones are about 0.2ohm. Are these ok?
Any help please?
Thanks

[Xenomorph]

@Guruji:
In case you have shorts between the layers of your secondary winding, then the induced voltage will be much less and thereby inhibit the coil singing volume.
Have you put a diode between collector and emitter of the transistor just to check if the Joule Thief part actually works? It looks like you don`t get an oscillation going.

As Slayer007 has pointed out multiple times, check the correct wiring of your primary winding. Look for the dot indicators in the circuit diagram.
Try to reconnect the wires differently, you just have 4 possibilities to actually connect the ends, try them all and see if it works then.

Or connect a small neon signal lamp to the ends of the secondary winding. With 480 turns, the induced voltage should be high enough to reach the firing voltage of the neon at least.


If neither LED nor neon will light then check your circuit. Measure your transistor voltages/current, if you can`t do that use another transistor.
Check if you havenīt mixed up collector and emitter. The transistor is the main part of the oscillator.

Good luck

[Zigis]

Guruji, maybe for testing start with larger voltage, say 9v battery.
If you have neon, solder it between collector and emitter, like in Bedini SG, this is good for transistor protection without load too. If BJT work without load, neon must light. If you have no neon, you can put LED here too.

If you turn on with 9V battery and without anything in output, and neon light - BJT work. If not light, try to change ONE 20 turn coil polarity, only one coil, triger coil only or output coil only.

Once you get neon working, you can start to experiment more with voltages, output etc.
Zigis.

[slayer007]

I have been side tracked working with a similar circuit to the Joule Thief. It was posted by the Daftman at the "Teep" forum. It is like the JT circuit so I thought that it would be OK to post the info on it here. The circuit uses an small inductor instead of a toroid coil to generate the spike that runs the LED. I replicated his circuit and added a 'second battery' charging aspect to it. I then added automatic solar charging to it from a 'solar garden light' circuit that I found here: Solar Light This page at "talkingelectronics.com" gives a very good description of how these JT like circuits work and shows a diagram of the wave form. I found this very helpful.
Here is a video of my "Automatic solar charged Daftman LED/charger light"


YouTube - Automatic solar charged Daftman LED Light

Cheers,

Lidmotor

Great video Lidmotor,

It looks like you been busy very nice setup.

Thanks for posting the video.