Ferrite at work

  • Topic Is Sticky
  • Last Post 13 December 2020
Jagau posted this 12 September 2020

An easy to assemble circuit that works, I have tried it myself and it is stable and functional.

This circuit is not from me it is from a Russian researcher who introduces himself under the nick name of "not a square" in russian  'Не квадрат'' 'on this website you can talk to him.

The circuit has been working for 5 days now and the voltage is very stable. Here is the circuit:


I admit that I am happy with the result and that the ferrite fuel is doing its job. This circuit deserves to be studied and developed.

A view of the oscilloscope on the collector

I started at 2.3 volts and now after 5 days the circuit is stabilized at 2.457 volts



Order By: Standard | Newest | Votes
Jagau posted this 12 September 2020

Hi all

If you make the circuit, you don't need the 0.5 capacitor to connect the 1 turn coil to itself.
I also had to adapt the capacitor from 5uf to 1.8uf unpolarized with low ESR
I also spoke to the author of this circuit, he told me he would come to register on here.

Let’s welcome it as aboveunity.com can do.


Chris posted this 13 September 2020

Hey Augenblick

This is perhaps harder to follow than the original?


Best wishes, stay safe and well My Friend,



Jagau posted this 13 September 2020

Good morning all

Thanks Chris for your help.

I did in schematic with Spice while respecting the wire colors and the start and end of the wires on each of the 2 ferrites.


A spice reproduction of the author's original drawing.

As I said in a previous post, I modified the circuit a little so that the battery charge increases more quickly. From the first day until today (fifth day) the load has stabilized more than it was at the beginning while still operating the circuit and with the two LEDs on.

Modification I made fromoriginal circuit.
I took off the cap. 0.5 uf and connect the loop of 1 single turn without the cap. Adjusted the capacitor from 5uf to 1.8uf which gave me a very quick increase in circuit performance.

I made a second circuit with a TV ferrite that works great too.

Am experimenting with this circuit without any battery, only with preloaded caps and the preliminary results are very good

hope this help


Attached Files

Jagau posted this 13 September 2020

i augenblick

Yes the polarities (phase) of the coils are very important.
Follow the color of the wires as I added on the shematic.
Identifying your coils with start and end marks respecting the place where they are located on the ferrite will be easier.
Need help do not hesitate to ask.

At the end of the fifth day both systems are still working very well, I think the ferrite fuel is doing its job, more scopeshot to come


Nekvadrat posted this 14 September 2020

 с этого начиналось: 

...подключил к подсевшему аккумулятору (1,18В)...напряжение поднялось до 1,236 - держится несколько часов


Chris posted this 14 September 2020

it started from this:


Добро пожаловать, Неквадрат,

Спасибо, что поделились!

С наилучшими пожеланиями, оставайся в безопасности, мой друг,



Welcome Nekvadrat,

Thank You for Sharing!


This thread should be marked AU then?


I agree, done, if needed we can revert.

Best wishes, stay safe and well My Friend,


Nekvadrat posted this 14 September 2020

Jagau posted this 14 September 2020

Привет, Неквадрат
Я очень горжусь тем, что вы здесь, на этом форуме, добро пожаловать.
Я верю, что все присутствующие смогут оценить ваш проект, потому что только вы его автор.

Я еще раз вернусь к информации, которой вы только что поделились.
с большим интересом, большое спасибо за то, что вы здесь.

P.S. На шестой день две системы все еще очень стабильны и функциональны, напряжение не падает, как предсказал Неквадрат. Да, я тоже верю этому А.У.


Hello Nekvadrat
I am very proud that you are here in this forum, and welcome.
I believe that everyone present will be able to appreciate your project, because only you are its author.

I will come back to the information you just shared.
with great interest, thank you very much for being here.

P.S. On the sixth day, the two systems are still very stable and functional, the voltage does not drop as Nekvadrat predicted. Yes, I also believe this A.U.


Jagau posted this 14 September 2020

The circuit I did on Spice is the one to try for now and it's the one that works.
There are only 6 coils in this configuration and the sixth coil is not a coil, it is only a 1 turn winding between the two ferrites, so the circuit I have on my table is exactly as done with spice.

As for the two LEDs, with the start of an analysis of the circuit that has not been completed, they act as a switch and not as a load.
As I could see it with my oscilloscope, the circuit does not saturate well on the contrary it consumes almost nothing and moreover it recharges the battery in another time of the oscillation.
There is interaction between the two ferrites in a back and forth movement like a swing that comes and goes to make an image.

As you may have noticed Nekvadrat made some additional modifications to the product circuit at the start.
I will make the modifications as he proposes with another assembly and I will communicate the results.

@Nekvadrat, don't hesitate to take us back when we make mistakes, I think your comments will be greatly appreciated

@Nekvadrat, не стесняйтесь возвращать нас, когда мы делаем ошибки, я думаю, что ваши комментарии будут очень благодарны


Jagau posted this 14 September 2020

Hi all

A picture is worth a thousand words
The direction of the windings is very important
I took a photo, take a good look at the start and end of each coil.

As you can see the first coil at the top,the white one is a bifiliary winding (2 wires)

Above and below very important, adjust the pot of 100k until an increase in the voltage of the battery.
At the beginning the leds are flashing it's good you have let the system stabilize and you are there.

thaelin posted this 15 September 2020

Здравствуйте: Поскольку Есть много рисунков, я хотел бы идти вперед с тем, который вы даты, как 14.1.2013. Можете ли вы сказать мне, сколько жарких используется и какие ветры рассчитывает для каждого, чтобы не иметь путаницы. Спасибо и теплый прием. таелин

Nekvadrat posted this 15 September 2020

...просто соберите по рисунку, дальше - нет предела усовершенствованию........

Nekvadrat posted this 15 September 2020

схема довольно интересна и имеет перспективы к развитию...

Vidura posted this 15 September 2020

Hi Nekvadrat, welcome on the AU forum. Have you had some success replacing the battery by capacitors? Regards Vidura.

Jagau posted this 15 September 2020

 HI all

A recap is in order
When I started this post I asked '' Nekvadrat '' for permission to bring here on the Aboveunity.com forum the drawing which appears at the beginning of the thread, it is the one I had at this moment in my files.
He kindly replied that there was no objection to my bringing it here.

Following the invitation of Nekvedrat on the forum, I am very happy with his coming, I learned like all of you that there were 2 other drawings that nekvadrat kindly provided to us.
There is an evolution in the first drawing which will lead us to evolve as nekvadrat did, it must have taken hundreds of hours, and I thank him for letting us know about its evolution.

In order to avoid confusion in the 3 drawings, for the moment I would like everyone to focus on the first drawing provided in order to understand it and to understand all the nuances of this first drawings.

Having some experience in building such a circuit it is much better to start with things that can be understood and explained.

I can tell you seriously that the first drawing works and I start to understand the nuances

We can then evolve with the last two drawings as nekvadrat did and ask for his help without constantly bothering him with trivial questions.

This is what I had in mind when I started this thread, thank you for your understanding


Wistiti posted this 15 September 2020

Agreed Jagau. I will try the first one you share and post my results when I will have it working.

Thank you for your support and contributions!

Wistiti posted this 15 September 2020

Hi team!

Here is my replication of the first circuit Jagau have shared with us. 


I know many things is different cause i have to use the things i alredy have on hand...

Instead of a battery i use a Maxwell Ultracap of 2000f (i know it's big!)


I have to said i have to inverse the bifilar connection to make it start.

Will see how long it will run...


Hope to see some other's replication!


Attached Files

Jagau posted this 15 September 2020

Hi wistiti

Well done wistiti

Will follow your result with great interest


Vidura posted this 16 September 2020

Hey friends, I have also given a try to the circuit yesterday, like wistiti I also had to swap the terminal of a coil compared with the drawing (the one on the base of the transistor) to make it oscillate. An old NiMH battery started at 1 v have raised to 1.15v in a few hours. Replacing the battery with electrolyte capacitors @1410uF it doesn't sustaining. Vidura.

Jagau posted this 16 September 2020

Hi vidura

 Do you use straight ferrites like wistiti or round ferrites, because the design is made accordingly of round ferrite (closed magnetic path). But for you too it works well done.

A battery is an active element (ionized chemical product), it does not charge in the same way as a capacitor which is a passive element (displacement current between two plates)
To charge a capacitor, i think, this is not the right setup.
But it's good to try to experiment, all experimentation is good to do.

P.S, almost a week already with my setup and the voltage has not dropped by 1millivolt, the system is self-sustaining, the oscillator is working and the 2 LEDs are on.

L0stf0x posted this 16 September 2020

Hello team, I also had a fast try to it like my friends Wistiti and Vidura..  also had to reverse polarity. 

I have no idea what the LEDs should do, but they blinking.

The capacitor is charging somehow.. I hope I ll find time to experiment more later today and understand how it works.

Wistiti posted this 16 September 2020

Nice job my friend!! Just to be sure, witch coil do you reverse the polarity?

Thanks for sharing!

Jagau posted this 16 September 2020

Hi LOstfox

It's really good to see that it works for you too.
As for the leds blinking, I noticed the same thing as you last Thursday.
The 100k should be set to have minimum voltage drop at startup and I have noticed that when the battery reaches a certain level the frequency of flashes decreases and only a very slow flash is surely a feature of this circuit. I have already built 5 different circuits with the drawing that Nekvedrat provided us.

Well done it seems that it also charges a capacitor that keeps us informed of your results, with the whole team we will surely achieve a good result.
thanks for sharing.

I will invite others to share and experiment, don't be shy this is the first selfrunner I see that works without voltage lost.

Other future observations on this selfrunner.


L0stf0x posted this 16 September 2020

Hey my friend Wistiti! Hey Jagau, thanks for sharing! Sure! I changed the polarity of the upper purple coil (at original drawing), the one that goes to base of transistor. I also burned 3-4 transistors and several LEDs while I was playing using battery with it..

The super capacitor is 10 farad and started with 0.57V and went to 1.764V in few hours and stayed there.. and the LED was blinking faster. I haven't played with 1.5uf capacitor yet.

Another try was with a LiIon battery at 3.95V and when I was connecting earth (red cable) at the negative lead of the LED that is connected to both cores (Led was removed.. positive lead was open) the voltage jumped to 4.40V. Strange eh?

Edit: my earth connection is directly connected to a copper rod buried 3 meters in my garden. It is not coming from the main house ground grid connection.

This need more time to investigate!

Jagau posted this 16 September 2020

hi lost

 Yes I think the 4.4v is the voltage that re-ignites the system, that's why I think he installed a 4.7 volts zener diode in his last circuit to regulate the system (to be investigated)

This circuit like you is really efficient when it is well adjusted, it consumes almost nothing and gives a lot, like you my I'm investigating and it's fascinating isn't it?


L0stf0x posted this 16 September 2020

Video is ok now sorry about it! Sure Jagau this circuit is very interesting ! I wish I could understand how it works! but I have no idea yet!

Wistiti posted this 16 September 2020

Nice video my freind L0stf0x!

When the time permit, would you draw a schematic of your circuit using earth ground? im not sure i understand where it is connected...?


Thank you and happy to see you back!

YoElMiCrO posted this 16 September 2020

Hi all.

I join the replica of the circuit, also its study.
Today later without fail I do it, thanks for sharing it.
As soon as I have an answer about how it works I share it
in this same thread.

Thanks in advance.


Jagau posted this 16 September 2020

 Hi yo

 Thank you for your presence on "ferrite at work"
I very much appreciate your comments knowing in advance that they will be very relevant and informative.



kalleFin posted this 16 September 2020


Hello to everyone!

I took my shot at this wonderful project. Thanks to everyone who is involved in this. I greatly appreciate your openness. My circuit is running currently for the 25th hour using two AAA alkaline batteries in series. That was what I had on hand when I started this experiment. At first I wondered if this really could work as I witnessed the batteries lose voltage, but as soon as Jagau described the tuning, I got to the system up so that the batteries seemed to be regaining some of their lost charge. Now the charge seems to have stabilized after about 24 hours of climbing.

My build uses large ferrite cores. I was planning on using them to power a Barbosa & Leal type of set up. Actually this system has some resemblence to their designs, in my humble opinion.

Thank you Chris for this opportunity to share and collaborate and reach beyond. You have been a great inspiration!



Jagau posted this 16 September 2020

Hello kalleVery nice replica I appreciate your sharing. When you will be able to have rechargeable batteries you will see a big improvement, I have the same configuration as you with 2 small 1.2 volt AAA batteries in series, I find the system to be much more stable at this voltage. Once again a great success.

Yes Chris is surely very proud of his team that I believe in a spirit of cooperation, this forum is a treasure

P.S. My friends, Don't forget to thank Nekvadrat for giving us the opportunity, He is the inventor of this circuit



Wistiti posted this 16 September 2020

Yes, the first thank's goes to Nekvadrat!

Im curious if we can pump it up at a higher voltage, something like 12v or higher...

A really basic and fun circuit to explore.

thaelin posted this 16 September 2020

My first try was a flop. Did the fix and it runs. Has a very high cycle so both led's are bright. But is an amp eater. Does use a 3055 tho and needs to be changed to a smaller transistor. Proof of concept passed with colors.

Will have a better layout soon as back from 4 day work trip. Will need to figure out how to post pics/vids to show how it looks.


and:   tried to run on 3v and will not. Think transistor trigger circuit is too much.

L0stf0x posted this 16 September 2020

Wistiti sorry I missed details.. I repeated the earth effect, to make sure its repeatable.. and it is! Here is where I connect the earth.

You don't need to remove any LED. And do not reverse the coil.. just use the original circuit on that. The LEDS are both off. But the voltage at battery goes high immediately when I connect the earth there. I got a new video but its nothing different than my last one. Same effect. Strange indeed! 

Edit: I think without the Led is working better.

Jagau posted this 16 September 2020

Hey thaelin
I know it's not often easy, maybe just a little phasing error and you'll get it, just keep a small 1.2 volts rechargeable at the start it will be easier.

A very good thread that Chris wrote a few times ago but which will surely help you avoid problems with this circuit. Look to the right of your screen, link is right there :   Some coils buck some dont, 

@ lostfox
wow you discovered a new way with an external mass as indicated, it's good let's keep experimenting.


YoElMiCrO posted this 17 September 2020

Hi all.


I still do not test the circuit, but doing
an analysis by inspection of the same I have
more or less an idea of how it works.
Looking at the oscilloscope image it is evident
What is related to the phenomenon of the curve
negative magnetic permeability of the material
used in conjunction with your way of working.
If you see the video that Zanzal from Avalon uploaded you will see
their similarity.
This weekend I will replicate the circuit without fail,
also that of Loz, to study it better.

Thanks to Nekvadra for sharing.

Thank you all in advance.


Wistiti posted this 17 September 2020

Hi team.

I have restart the experiment cause my transistor fail (cheap one fron Ebay...) I change it for a 2n2222 and rewind on a toroid core.


After a night running on the ultra capacitor, i see a little discharge on the cap. Maybe i have to have a better tuning.

Experimenting will continue...

Jagau posted this 17 September 2020

Hi all

An update
I started this schematic last Thursday 10th with 2.325V and it has been now a week since the circuit works and the voltage is today at 2.54V.
The 2 leds are on and the oscillator is still working.
The nominal voltage of each of the batteries are normally 1.2V, i use 2 of them.
So these are very good results considering that the syteme run 24 hours a day.
Some waveforms taken this morning on the collector of the BT139 is the transistor I use

Take note that on the second photo it is when we stretch the first to see what is inside.


Atti posted this 17 September 2020

Hi all.

Should both LEDs flash or just one?

Jagau posted this 17 September 2020

Hi atti
At the beginning the 2 leds flash but when the battery charge becomes stonger,

a little flicker is sent.

did you manage to do it? if so, well done


Atti posted this 17 September 2020

Only one LED flashes. There will probably be setup issues. It has been operating for three hours without a voltage drop. There is no voltage increase yet. I will perform more tests as my time allows. (my first opinion on the setup: it's like a battery desulfator)

Jagau posted this 17 September 2020

Try to reverse the last coil at the bottom
probably just a phase error.
For battery desulphation there are more efficient systems.
It works for the author and me and also several others just a few adjustments you will get there.

Nekvadrat posted this 18 September 2020

Atti возьмите многожильный провод и соедините "восьмеркой" два тора. От оборота "восмерки" в ту или иную сторону - зависит многое...

L0stf0x posted this 18 September 2020

Well I am still puzzled about the circuit. I tried to charge the cap again using a 2n3904. But with 2n3904 It doesn't want to charge the capacitor. But with BD437 "most times" the cap is charging again. It looks that every movement of ferrites gives different results. So my opinion until now is that the tuning has to be done not so much with the trimmer rather with the position of ferrites and their common loop somehow.

Jagau posted this 18 September 2020

Yes you are right Lostfox, I have built a few of them and they all work differently.
Even on a certain model of breadboard it did not work, too much stray capacitance and stray resistance on a breadboard I think.
It is a very subtle circuit but it works very well to recharge batteries that were not full. And the most interesting is that it does not discharge but by operating the circuit for more than 8 days for mine.

To recharge a capacitor I believe that, for the moment, it goes a little too quickly, we must try together with our beautiful team to find how it maintains the voltage of the batteries as well, when we have understood its operation well we can surely go there. with capacitors.
It would have been too easy just to build this little circuit, but I think it's a great start and there are still several hours of fun to experience on this project.

Greeting Lostfox

Chris posted this 18 September 2020

My Friends,

You Guys and Gals, yes, I believe we have a few female Members which is awesome to see, are all AWESOME!

This is what can happen in a controlled environment, no trolls! You should all be extremely proud of yourselves! Thank You, ALL of you, for being so awesome!

This is awesome to see! People coming together, uniting, for love and helping each other! Achieving what was not possible yesterday, today!

Best wishes, stay safe and well My Friends,


Wistiti posted this 18 September 2020

Hi team.
Just let you know I have stop the experience cause it drained the capacitor...

After seeing the last video of L0stf0x I chose to rearenge my setup...

Will let you know ...

Zanzal posted this 19 September 2020

Hi team.
Just let you know I have stop the experience cause it drained the capacitor...

After seeing the last video of L0stf0x I chose to rearenge my setup...

Will let you know ...

Hello Wistiti, are you using that large Maxwell capacitor as your source? If I may offer a suggestion, you might wish to consider a smaller capacitor, between 100uF and 1F. Preferably something with low self-discharge. Consider how much energy it takes to charge a large capacitor like that then also consider the self-discharge of the capacitor. The self-discharge may be higher than the charge rate and so the capacitor will eventually discharge even if the device itself is using less energy than it is producing.

Atti posted this 20 September 2020



 So my opinion until now is that the tuning has to be done not so much with the trimmer rather with the position of ferrites and their common loop somehow.

I got into a similar situation.
After one day of operation, the result is:
the initial voltage of 2.45V did not increase, but did not decrease. The settings need to be changed...

baerndorfer posted this 20 September 2020

if you can have a circuit with self-oscillation then you can tune your wires so that they meet a sub-harmonic from your resonance frequency. as a result you can have a standing wave in the wire which can provide extra energy to charge your source.

very nice circuit - i plan to try also and will share my experience here.


Atti posted this 20 September 2020

 i plan to try also and will share my experience here.

We welcome you and excitedly.


Jagau posted this 20 September 2020

Hello Atti
Are you using 1.2volt rechargeable batteries??

If this is the case with two batteries the full-charge voltage will be around what you mention 2.45 volts and it is normal that the max is reached, but if the voltage does not drop you have succeeded.

I have approximately the same result as you and voltage do not drop


Jagau posted this 20 September 2020

Hi all

Another important detail:
Whether it is for this project or another, never let your DDM constantly connect to your circuit.
Why; the DDM has impedance and it eats the battery slowly, rather randomly taking time-to-time tests.

The input impedance of my DDM is 10 meghom, some lDDM have a lower impedance going up to 1 megohm, check this when you buy a DDM.

Here in this project we use low voltage.

P.S. On a microwave oven there is a parallel resistance of 10meg. on the capacitor,  it is used to discharge the capacitor when the microwave is stopped.


Atti posted this 20 September 2020


I used a non-self-oscillating oscillator to investigate why you can recharge the battery. Currently this is my opinion .You can also watch in the video.If my time allows, I will try another layout later.

Jagau posted this 20 September 2020

Hi Atti

You are correct, indeed experiment with various options that's how we move forward.
thank you for sharing your research.

Do not try to recharge ordinary batteries, it will not work. Only rechargeable ones work and specifically NI-MH types


Mimo posted this 20 September 2020

Hello team,
special thanks to Nekvadrat.

here is my setup (with the materials I had on hand):

- 2 ferrites dimensions: 36 X 22 X 15 mm (unknown material - old stock);
- 1 capacitor 2,2uF 250 V (C1) unpolarized, reference A50IT42202660K;
- 1 capacitor 0,66 uF (C2) unpolarized;
- 1 capacitor 3,2nF HT unpolarized (added in parallel with D2 for more luminosity on D1);
- 1 NPN 30 V transistor, 120 mHz, 2 A;

- L1, L2, L4, L5 and L6: Litz wire 0.1 mm X 180 strands;
- L3: 0.2mm x 19 strands tinned copper, stranded wire;
- R1 50 kΩ X 2 A (50 is sufficient, besides I did not have 100 k) ;

- V1: 2 AAA rechargeable batteries 1.2 V, 700mAh NiMH in series.

- I needed, as indicated above, to reverse the ends of L3.
- I secured the two loops of L6 with smalls nylon clamps.

I stopped adjusting R1 just as both leds stopped flashing.
D1 is very bright, D2 is much less.
V1 is currently stabilized at 2.280 V, after a gradual increase in voltage (starting at 2.150 V).
The circuit was powered up about 20 hours ago, so it will be necessary to monitor progress.



Jagau posted this 20 September 2020

Hi Mimo
Very interesting, and clean setup, I note that you use litz  wire
I can't wait to see the results with this type of wire.
thank you for sharing, we keep in touch.

P.S. I think we have a good team here,8 members have already tried and the results seem positive to most of you. it is certain that there will get positive results.


YoElMiCrO posted this 20 September 2020

Hi all.

Today I replicate the circuit and submit it to analysis,
This week I tell you what I have found.
Also photos, of course!
Very interesting how it works, it's a bit complicated
to explain, in fact I still don't quite understand it.
But time to time, it's just that.


thaelin posted this 21 September 2020

   Up and going again. Had to resort to the  3055 as I blew the 2222 about 5 mins in. Showing 9.6 mhz and 9.6v on the collector. Battery is showing a rise at this time. Will be loging the voltage every .5 hr for a while tonite to see where it goes. Using a AAA NiMh to start with. Will try and find a set of bigger. Starting volts was 1.257


Jagau posted this 21 September 2020

Welcome aboard Thaelin As you may have surely noticed, the adjustment of the 100K variable resistor is very important.

@mimo,  I am curious to know if Litz's cable is behaving well?

@YO, it looks like a very simple circuit but it has potential. One thing I noticed the 2 LEDs must be there.
If we remove one the circuit stops charging, as if the 2 leds act as a switch. The oscillator is easy to analyze but in the second rear BEMF of the second ferrite this is what I have not yet understood. It is as if two impulses crossed, this is what I see on the oscilloscope


Nekvadrat posted this 21 September 2020

...репликации в большинстве случаев имеют положительный результат. Имеет смысл расположить тороиды определенным способом, к примеру друг над другом ...либо перпендикулярно. (могу и ошибиться, но ведь мы здесь для получения результата...)

Nekvadrat posted this 21 September 2020

....Jagau,...многие знают "переключатель Тесла". В данном варианте Вы можете попробовать свой переключатель, - заряженных АКБ, почему нет?

Jagau posted this 21 September 2020

Very judicious remark Nekvadrat,

I made different assemblies and in one of my set I noticed that the location of the 2 ferrites one in relation to the other seems important and even when they are very close to one of the otherwise it only worked if I removed L6 the coil in one turn, a curious phenomenon to study.
For the Tesla switch I did not experiment but will reread the details on Tesla.
Thank you for your help Nekvadrat

Mimo posted this 21 September 2020

@Jagau, and the team

yesterday I made a second identical device, except I did not use Litz wire, only 0.2mm x 19 strand tinned copper wire. The 50K potentiometer was not enough, I had to add a 40K resistor in series.

Led D1 is brighter with the Litz wire than on the other device.

I have just disconnected the small capacitor that I had placed in parallel with D2 and which made D1 much brighter. Immediately, the tension on V1 rises again.

On the second device, I had placed a small capacitor in parallel with D1 to get more brightness, but I just removed it. Here too, the tension on V1 rises again immediately.

I will observe until tomorrow the tension of V1 on both.
Perhaps it will be necessary to cut the power supply for a few minutes and measure the voltage of V1 at rest to have a better idea of ​​the real voltage. I will do it tomorrow.


Jagau posted this 21 September 2020

Hi Mimo,

We are in experimentation mode and that can help whatever brings us back.

I will also invite those who have succeeded, sometimes it is the small details and observations that you have made that help us all to move forward. Do not hesitate to tell us about your tests.


thaelin posted this 22 September 2020

  Sadly the new one didn't charge. Ran down to .9 volts so called it off. Will have to rebuild and see how that goes.


Can someone post the ohms on the pot setting to see if mine was even close?




Atti posted this 22 September 2020

Hey Jagau.

There may be other better solutions. Yes, this is an experimental mode right now. As you say.
At present, I continue to maintain my contention that this is in part desulfurizing.
Why do so many people run out of battery after a while?
I know a lot depends on the settings.
And that doesn’t mean there’s no option in this app.
If the process works, it should work not only with a small battery (1.2V) but also with a larger battery. For me, the question is, why does it work, how and for how long?

If you look at the oscilloscope diagram in the uploaded video, you can see that almost a quarter of the current returns to the battery. Why?
I think this is a forward switching in asymmetric mode.
The circle is not self-vibrating here. So the control.
The diode is replaced by an LED connected in series with another transformer.
When this other transformer is loaded with the correct polarity, it works similarly to this arrangement.
This is Chris' layout.  

One has loosely coupled coils. In the other, the coils are on two transformers.
Thus, if one of the columns is loaded with the correct polarity, the self-induction voltage on the load on the control column and on the control coil will increase.
This is returned to the battery. The only question is how much we will use and how much we can miss.

Jagau and everyone.
If not, could you show an oscilloscope image of the recorded image? Thus, it would turn out that the phenomenon is similar in others.

(i hope the google translation is understandable, but if anyone bothers you let me know)


Mimo posted this 22 September 2020

here is the value of the potentiometer of my 2 devices:
44.2K for the one with the Litz wire and 40.4K for the one with the tinned copper wire. With this setting, the LEDs are very bright and do not flash.
Check the correct order of your connections, because this assembly does not present any difficulty, it must light up as soon as the power is turned on, either in flashing or fixed mode. The potentiometer must be adjusted at the limit of the flashing / fixed LEDs. Remember to reverse the two ends of L3 from the diagram in the first post above. Check the data sheet of your transistor, its base is not always the middle leg. Obviously, scrupulously respect the winding directions of the coils, including L6 placed in 8.

For all, according to my findings:

1 / By adjusting the potentiometer so that the 2 LEDs flash quickly, you obtain a complete self-charge of the batteries, within a time which depends on the initial voltage, but at most in a few hours. The voltage then slightly exceeds 2.4 volts. When the LEDs are on steadily, the complete self-charging of the batteries also operates, within a slightly longer time and the maximum voltage is slightly lower, but still higher than 2.3 Volts.

2 / I brought the two tori together and positioned them in all possible arrangements. I did not see any improvement in the brightness of the LEDs or in the evolution of the voltage.

3 / On my device with the Litz wire, the D1 led is slightly brighter on D2. On the other device, D1 and D2 have approximately the same brightness. The advantage with Litz's thread is very moderate: D1 a little brighter, but D2 a little less than on the other device.

4 / By placing a capacitor of very small value in parallel with D1 or D2 (test with a panel of capacitors to find the right value), we obtain a very significant increase in the brightness of D1, but the voltage of the battery drops .

5 / By placing a neodymium magnet between L4 and l5, North against the outer edge of the torus, and with an adjustment of the potentiometer, I succeeded in lighting 4 leds connected in parallel, very bright. The voltage has dropped to around 1.6V. I did not continue the experiment beyond a few minutes. I removed the magnet and set to classic mode. This morning, the battery had fully recovered and was around 2.4 Volts.


Jagau posted this 22 September 2020

Hi Mimo

You shared valuable information. After having made various tests that you have realized the tension to drop a few times but as you said at the end;

This morning, the battery had fully recovered and was around 2.4 Volts.

You are proof that it works. And several others have also succeeded, just follow the plan as indicated by Nekvadrta.

My potentiometer is very close to your 49K value, it looks the same.
Your advice is very relevant and I believe that several will benefit from it, thank you Mimo.

I will post more photos that can help others succeed.


Jagau posted this 22 September 2020

Hi Atti

I use NIMH rechargeable batteries, not regular batteries. The batteries did not have time to sulphate a NI-MH battery does not contain any sulfite, it is composed of nickel hydroxide for the positive electrode and for the negative electrode use an alloy that absorbs hydrogen. They lose their efficiency by degradation of materials they are not recoverable.

When you run a project in selfrunner mode like this, the batteries must constantly power the circuit and recharge at the same time.
The beauty of this project is that it works, we have several that have been successful and it's up to us to try to find out how.
This is what we are all doing together.


@ all

On another Russian forum I found this pdf which can be helpful, it is in Russian but you can translate it easily with google translate. See attach pdf


Attached Files

Wistiti posted this 22 September 2020

Hey Jagau.

It's me or in the .pdf you shared L4 and L5 are also bifilar?

Thank you!


ps: witch free translator do you suggest?

Jagau posted this 22 September 2020

Hi wistiti
The pdf provided is not from Nekvadrat, it is that another one who made it is a kind of summary that he took everywhere on different sites.
If you reread carefully he says it clearly and I quote him:

To facilitate the perception, I drew transformers on the cores.

This is to facilitate understanding it is not as the author did.


I use a BD139 transistor


Wistiti posted this 22 September 2020

Ok thanks for your reply.

Im not looking for a transistor suggestion but a translator.

Jagau posted this 22 September 2020

On another Russian forum I found this pdf which can be helpful, it is in Russian but you can translate

it easily with google translate


thaelin posted this 22 September 2020

Google failed. I used Bing and had to translate page then copy to text file. Was going to just add the text file but see no link to add file. So will just insert it here.

Good time readers, subscribers, and all those who just wandered into this page! Today we will talk about an interesting scheme,
repetition, which for a few days was just a record! And even in the open mode, which in itself is an event in the world of Free
Energy Seekers. The author of the scheme is a free energy seeker from Ukraine, known in many forums under the nickname"Not a square".
It's actually a diagram.

Схема рекуперационного фонарика Автор "Не квадрат", Украина
Scheme of Recovery Lantern Author "Not Square", Ukraine

... don't get me wrong, I've been looking for a long, long time. Today, of course, the experience of 2014, it is difficult to
remember, but the scheme is even a working layout, photos and oscillograms ...

This finder went further and increased the size and tension of the source. This replicator has two stable devices, with different
parameters of components. This means only one thing - the working principle of the scheme is already repeated.

A few more searcher posts under the name "Transistor"

For the fifth day the generator assembled as a night light is not squared. During the day the battery rose to 3.8 volts, and then
the controller akb centalized the charge by 3.8 volts, along with the load - 3 LEDs. It's a joy. The forums are full of stagnation
and no progress. You have to start somewhere. If you have something interesting, share it. Criticism to do it is more easy than to
take soldering iron in hand.
What did you say, "transistor"? I transfer in one day the recovery system (reverse impulse), charged the powering AKB from the mobile
phone, which is constructively built into the mini-controller level of charge. This node does not allow the voltage on the battery to
rise in the six 3.8 volts. For dummies and investigators I will explain: the voltage of the pulse of the charge is above the level
of 3.8B with a current component sufficient to charge the battery.

Another post given by the seeker

... Without it, the brightness of the glow depends on it. Guys how can you talk about the work of the scheme without collecting it
and not experiencing it? That would collect it and fix a couple of hours. The layout on armored cores works worse. The diode stands
according to the author's scheme ...
... Well, about the layouts, I do not have cool devices except the arrow and electronic tester can do it well. As a result, 10 days
I even got used to and relatives too that in the hall in the corridor and in the basement around the clock there is a light. And when
I ran them there and there were no three volts. What is interesting when adjusting the shift in the direction of the increase in
current increases the current of the transistor and the frequency of the generator as well as the charging current. I'm already
working on a new generator in the case of the factory lantern 5 LED as I finish. If all goes well in the future car akb with a
powerful generator of this type and already with powerful ladder lamps for lighting ...

... Basic and collector windings are not included as blocking.
It follows that without led Led1 the circuit will not start.
Moreover, the beginning of the transistor's opening with a base voltage instantly results in the triggering of the Led1 diode due
to two windings, which are additional supports for it. One of these windings has a constant time, given by the capacitor 0.5. So
the burning time of this LED is determined by this capacitor (in the first approximation). Without it , time is determined by the
fading of current in the turn of communication.
Once Led1 is extinguished, the transistor is shut down by the result of a locking voltage on the base winding of the transistor.
At this point, Led2 takes over the self-induction of the second ring, developed through the bond revolution and the capacitor 0.5.
The charge, which"settlement" at the transition of the collector-emitter at the expense of the OEDS of the first tr-ra, then goes
back to the source through the same collector winding.
The scheme differs from the usual blocking in that the closure of the transistor is not associated with saturation - neither the
transformer (it is saturated), nor the transistor. The closing process is managed by Led1 in the diagram described above.
Also, the current consumption from the source in the opening phase of the transistor here is much more than the usual blocking -
at the expense of Led1. ...
This is the opinion I have cited for the fact that the analysis of the scheme is different for everyone.

There are still repetitions of this scheme of Not Square.

The quote: "He is also very capricious to the choice of details. After all, 12 to 17 MHz is not a hoarse. By giving away the second
ICE you can raise the amplitude with your hand, and even more than with LED. It is better to put a high-frequency transistor. For
example, the same CT3102, but better higher power. I think it's fashionable to make a composite emitter repeater. In general, gene
class, seven feet under the keel! If two people already write that the accu is charging, then probably it is."

Цитата: "Он также очень капризен к выбору деталей. Все таки от 12 до 17 МГц это не хухры мухры. Отпаяв второй ЛЕД можно рукой
амплитуду повышать , да еще и больше, чем со светодиодом. Транзистор лучше поставить высокочастотный. Например тот же КТ3102,
но лучше повыше мощностью. Думаю модно делать составной по типу эмиттерный повторитель. В общем ген класс, семь футов под килем!
Если уже двое пишут, что акку заряжается, то наверное это так."
The quote: "He is also very capricious to the choice of details. After all, 12 to 17 MHz is not a hoarse. By giving away the second
ICE you can raise the amplitude with your hand, and even more than with LED. It is better to put a high-frequency transistor. For
example, the same CT3102, but better higher power. I think it's fashionable to make a composite emitter repeater. In general, gene
class, seven feet under the keel! If two people already write that the accu is charging, then probably it is."


Repetition of the third (nickname "AND")

Цитата: "На общей волне энтузиазма, тоже собрал схему Не квадрата. Действительно хватило 30 минут. Заработала с первого раза.
Транзистор кт3102АМ. Сколько будет работать, посмотрим. Не квадрат, спасибо!"
The quote: "On the general wave of enthusiasm, too, assembled a scheme not square. It really took 30 minutes. Earned the first time.
Transistor ct3102AM. How much will work, we'll see. Don't square, thank you!"

These materials from the website "STRANNIK-2"


My "vision" of the work scheme:

I have drawn "transformers" on the cores (cores) for ease of perception. In design we have: Block generator system (R1-C1-L4-VT1)
which is constructively placed at the core of the T1 transformer, together with windings: Bifilara (L1-L2/1) and 0.5 revolution (L5/1).
In fact, the work of a transistor is set on this core. The second core forms the Transformer Tr2 with windings of the second bifilar
(L3-L2/2) and 0.5 revolutions (L5/2). We have the connection of both cores through the L2 and L5 sections of which form a common
"inductive" connection. How can the scheme work.

When the battery is connected to the circuit, the elements of the circuit have current. The current flows in the L2 winding, through
both sections and led Led 1. If the voltage of the source is sufficient, it will light up if not enough, accordingly will not light up.
To the base of the transistor, through the winding L4, the same comes current. In the L5 winding, the current will only result from the
oncoming currents of L2 and L4 windings. there are still many nuances, perhaps, but for the first understanding is enough. The incoming
current to the base of the transistor opens it. And the current begins to flow through the L1 winding. since this winding has half as
much resistance as the two L2 windings consistently, the current will be larger in it, respectively the core will be remagnetized, and
through the winding L5, will give the impulse to magnetize to the second core. At the same time, the current in the L2 winding will not
change its direction, and the voltage in this thread will increase, which will cause the glow of the LED Led 1. At the same time, the
current in the L3 winding for its glow has the opposite direction (yes I'm not mistaken. And the main current in the L1 winding locks
the current in the control winding L4, which causes the transistor to shut down. (this is an element of the block generator work). At
the moment of switching off, we have a "bloated" magnetic field in both cores, and a promise to self-induction in the system.
Self-induction is if just the occurrence of EDS induction, with the"absorption of your own magnetic field coil"

Due to the phenomenon of self-induction in the electrical circuit with the source of EDS at the circuit is installed not instantly,
but after a while. Similar processes occur in the case of a chain-breaking,and(with a sharp blur) the magnitude of EDS self-induction
can at this point significantly exceed the EDS of the source.
We are interested in the processes precisely when the chain is blurred. But we should not forget that the current in the chain will
flow according to the rule of the Right Hand (in our case " for thesolenoid"). It should also be taken into account that the current
caused by the magnetic field will form in the thread in which conditions for it will be more favorable. This is a closed circuit and
resistance to the general circuit. There are a lot of controversial moments at first glance, but first we will try to understand. We
will deal with the rule of charge of the AKB.

The main charge is current and voltage. In order for the current to start moving towards the battery, the voltage of the "external
source" must exceed the battery voltage by N times. Also resistance elements should also be appropriate, naturally and the potential
of the field. If the voltage of EDS self-induction depends on the frequency, magnitude of magnetic induction and the length of the
conductor affected by this magnetic induction (and in our country it has a waning character), which we can roughly calculate according
to the formula:


EDS(Volts) )


As we see the power of current in our formula of definition of EDS is not! The current measure should be considered, according to
another formula - according to the law of Om for the full chain, with a little clarification.


I (A) - EDS - U bat. / R + Rn +r0

The current measure will depend on the difference in voltages between EDS self-induction and battery voltage, divided by the amount
of resistance to the recycling circuit. So, our charging current, this measure of EDS is not up to zero, but before the battery voltage.


To meet the sufficient condition, the guaranteed battery charge is suitable for us only a thread with the windings of L2 (L2/1 q L2/2).
It remains only to clarify whether the current flows through the LED, in the opposite direction? Let's turn to the omniscient network
with a request and find such material.

*The voltage of the LEDs is applied in the opposite direction.

The 25 volt voltage applied in reverse voltage will disable the crystal, with a voltage of 12 to 15 volts in the opposite direction
completely safe for the LED. The working voltage of the LEDs app in the opposite direction does not ignite and does not disable the

As we can see, this is also not classified data. The current of self-induction into the battery in the non-square scheme comes exactly
by this way, and the LED in this switch register does not burn. At the same time, L2 windings, connected consistently, act as if
separate components absorbing each of its core field by increasing tension. In this switch register

It turns out that the work, short closed winding L5 through both cores remains unclear. At the moment of inclusion of the L1 winding
with the appropriate direction of the current, in the winding L5 there is a current of the opposite direction, which excites the
magnetic field of the second core. Probably this combination does not give the cores, do not achieve full saturation. During the
register of induction pulse induction, the direction of the current in the L5 winding will be the opposite direction of the current
of the L2 thread self-induction, which will have a beneficial effect on the process of utilization of the stored magnetic field energy
in both cores. For each battery voltage, you need to calculate the parameters of the device personally, but these are the details of the

In my opinion. it is appropriate to supplement this scheme with two elements: VD1 Diode and Stabilizentron VD2. This is if the scheme
is done at great strain. Well and the led chain Led1 add a matching resistor, so that the reverse diode has dimed the LED with a

Not a square smart, that's why he passed by and did not develop further do not know. The only thing left is a power option, which I
will consider with the participants of the "Source" project. This scheme can be assembled by anyone. who little-meaning with circuitry.
Soldering is not enough to hold, you also need to think.

Wistiti posted this 22 September 2020

Hi team! Just let you know I have build the variant with bifilar on both toroid (like my question to Jagau above) and it's also work. I use 2 x aa nimh battery's for the source. Start 1 hour ago at 2,136v and I'm now at 2,156v.

Really interesting circuit.

Jagau posted this 22 September 2020

Hi wistiti
A second try and it works for you, perseverance always pays wistiti and let's go.

As several seemed to have problems for the last document submitted, I translated it to you in English with google translate

in pdf format

Et en francais pour toi Wistiti


Attached Files

Wistiti posted this 23 September 2020

Hahaha! Merci Jagau, c'est bien aimable.

Project still climbing. 2,159v it is slow but it goes!

thaelin posted this 23 September 2020

Is anyone having troubles burning out the D1? I have now taken out 3. Installed a standard diode and now system seems to be charging as it should. Started a log just now to see where it goes. Will leave it on all night and check again.


   I am now seeing the same hump wave form you posted about the 17th with the internal form the same too. I sure hope this is finally a working unit. Have made 6 now and was really getting frustrated with it. Time will tell.


L0stf0x posted this 23 September 2020

Wistiti my friend good to hear its working for you too!

I also run it for several days I use a Li-ion Battery started at 2.651V and slowly went to 3.554V and stayed there for almost 1 day. I will let it run..

Now its time to try different things that will make it more efficient. to try bifilar on both coils like Wistiti did successfully,

or double turns at coils ., also to build a 12v version with more LEDs, etc

Today if I have time I will build one using my best nano-perm cores with extreme permeability to see the differences. 

Jagau posted this 23 September 2020


Good morning all

Yes you're right Thaelin, it seems easy to build but it requires attention during assembly. If you have the same waveform as me, the ferrites are at work, well done you did.

@Lostfox, I see you are a real experimenter, you have several ideas for the design and different ways to test it, I will follow with interest and thank you for sharing your knowledge with all.

I invite those who have succeeded and who follow from outside the forum to join Aboveunity.com and thus be able to share your knowledge with us. It's free and as you've probably noticed, there is mutual respect among the members and a lot of mutual support.


L0stf0x posted this 23 September 2020

Jagau first of all thank you for the circuit, it is really interesting! Secondly I am just a simple experimenter, at the lowest level than the other guys in here,  but thanks for your encouraging words wink

Indeed as I see it Nanoperm cores gives faster charging!! I got this little Li-po battery from 3.40v to 3.94v in few minutes and raising as we speak!! I know the battery has low capacity so faster charging is normal, but I will check Li-ion as well to be sure.

I ll try some more variations I have in my mind .. I will keep you updated guys.


Jagau posted this 23 September 2020

I think we use the same ferrite Lostfox.
The one I have is the Ferroxicube 3E27 epoxy coated
Magnetic permeability of 6000 and with that everything has been working wonderfully for 14 days and yes 2 weeks now. The tension does not drop and I am as you say fascinated by this circuit which requires that to be studied. I have several theories now and plan to talk to you soon.

After 14 days Losfox I think we have a real selfrunner in our labs, what do you think?

Jagau posted this 23 September 2020

Another drawing made by Rakarskiy all in colors which could help to make the circuit of Nekvadrat.


To note on the base of the transistor the capacitor 0.1uf in parallel with the resistance of 5 ohms, probably to give more stability to the circuit, therefore to try.

L0stf0x posted this 23 September 2020

Nice drawings Jagau ! 

I don't remember their type but they are not ceramic type, they have plastic cup on outside and nanocrystal tape inside. Their permeability as far as I remember was from 40.000 to 70.000. But yes the best core material will be more productive here.

It indeed looks to be a self runner! I will try to run it with 12v using 3 li-ion batteries 18650 - 3400mAh each. So the changes should be the transistor to be high power one and maybe a resistor 100 ohm 1-2w resistor in series with the potentiometer may also required. Also 3-4 Leds in series will substitute the single ones. Mosfets and IGBTs also need to be tested.

Rakarskiy posted this 23 September 2020

Доброго здоровья всем! Я добавлю немного материала по схеме "не квадрат"! У меня есть собственное видение того, как работает эта схема. В любой энергетической отрасли важным моментом является индукция. Феррит не имеет хороших показателей. Сердечник должен быть выполнен с хорошими показателями инерции и индукции поля. В этом случае добавьте диод в противоположном направлении к светодиоду. Но самое лучшее решение-рассчитать схему по параметрам.


Исходя из логики, я упростил схему, но в этом варианте нужно более корректно подходить к расчету схем, лучше делать это на железном сердечнике.



Jagau posted this 23 September 2020

Yes lostfox  a great challenge to take up in 12 volts to be followed with interest.

And also welcome to the thread, rakarskiy and thanks for being there too to give some good advice.
Here we are only at the beginning of the experiment and you have a lot of lead over us on the Nedbvadrat circuit.
Many members here are interested in knowing it better and I believe in developing it.
As for the different calculations to be done, fortunately today on the net there is a lot of availability which can save us a lot of time.

When you have time, I would like to read your theory on this circuit and I believe that it will interest several.


when you say:

  The core must be made with good indicators of field inertia and induction

 Is it related to the magnetic permeability ? Maybe with Metglass it will be better than ferrite?

have a nice day


Rakarskiy posted this 24 September 2020

magnetic permeability for cores has two levels, absolute and permeability. The relative permeability indicates how many times the magnetic flux in the core has increased compared to the magnetic permeability of the vacuum. If we have a solenoid with a length of 30 mm (0.03 meters) without a core, with a certain number of turns (100), then when it is closed into an electric circuit with a current of 1 ampere, we get the magnetic field strength, which is calculated by the formula (Н = I * w/ L).

H(Am) = 1 * 100/0.03 = 3333 Am (Ampere per meter) ; To find out the magnetic induction of this field in Tesla, it is enough to use the relation (1Am = 1, 256 х 10 (-6) Т).  3333 Am * 0,000001256 = 0,004186248 T  

Usually, unclosed ferrite has a relative permeability of 600:  600 * 0,004186248  =  2,5 T.

Unfortunately, we will not get this value in reality. The reason is in the limiting possibilities for material saturation, for ferrite it is within 0.5 T. 

In fact, with this value of the strength of the electromagnetic field, we will heat up our core. There is one more feature that, in our opinion, is not related to field radiation, the value at the end of the core pole will already be in the region of 1/2 of the real saturation. Steel has a maximum saturation field of up to 2T compared to 0.5T ferite. In addition, electrical steel can withstand high values ​​of the strength of the original electromagnetic field. Academic sources indicate that the frequency limit for steel is around 5 kHz. But this requires very thin sheets.


L0stf0x posted this 24 September 2020

Rakarskiy welcome and thanks for sharing!

I tryed your circuit. But its charging the capacitor (5.4V - 5F) only if I remove R1 with led and place a simple diode there. Actually the two diodes have no perpuse to be there at all..

With the LED in circuit, the Cap is discharging slowly. I put more turns just to fill up the gap.. turns was 80- 160 -80.


Jagau posted this 24 September 2020


Hello Rakarsk

Good discussion;
This is a theory that seems valid I think but I think it could probably only be used to increase power with a different circuit than this one. The fact of using iron as a material greatly decreases the magnetic susceptibility which in my opinion has a very great importance that must be taken into consideration with this small circuit.
The ease with which the ferromagnetic material becomes magnetized (susceptibility) with very high inductance factor of 6000 and for these 3E25 ferroxcube allows this circuit to operate at low power . At the moment iron does not have this possibility
so with a very very very low current I can make these ferrites oscillate with a saturation well below the data sheets  

Probably Nekvadrat who is the author of this circuit could add some.

Don't worry, this is just the start of my theory, we still have a lot to say.


Jagau posted this 24 September 2020

Hi LOstfox

Nice demonstration, you are now the inventor of a new circuit to charge a capacitors, you are amazing
Thanks for sharing


Rakarskiy posted this 24 September 2020

Now take it and wind it a little differently. "Dumbbell winding" should be performed with a larger cross-section, but with fewer revolutions. Then the led and resistor will be justified. The reverse stroke that this winding performs must have a lower resistance and a length sufficient for the self - induction EMF to be dimensionally greater than the battery or source voltage by 1.618- 2 times. then the condition of the charge current pulse according to the Ohm's law is fulfilled.In General, this scheme was planned to compensate for the battery charge, but it is too simple.The development is already a commercial project, but the solution was suggested by a non-square scheme.

The formula for the maximum current pulse calculation according to Ohm's law is the same as for a traditional generator:

Current (Ampere) = Self-induction EMF (volts) - Battery charging voltage (volts) / dumbbell winding resistance (Ohms) + battery resistance (Ohms) + diode resistance (Ohms)

The charge pulse will be, not increasing, but decreasing. The current will only be up to the EMF = battery voltage level


link to the stimulation process. I recommend taking a simple digital generator. Unfortunately, the simulator can't calculate the actual field. But the chain's algorithm works fine. I used the capacitor to simulate the discharge of the source. In this version,there is a discharge, but this is natural, taking into account the primary and secondary parameters.



Wistiti posted this 24 September 2020


L0stf0x,  my freind, since I see you jumping in this, I know it will be a great journey! 😊


For you last experiment about the cap charger if you can draw a quick schematic it will help to replicate...


About the 12v version I already give it a try with the only 2 green led a have and as you can imagine I burned them... so I have ordered some more.  Also a bigger transistor with a heat sink is needed. There are some variant to be found to switch this circuit to a higher voltage...


Anyway is great to play with it.


Take care everyone!

Rakarskiy posted this 24 September 2020

the relative initial permeability and saturation induction of the ferrite are different. Saturation is the maximum field limit that a material can create. Ferrite has a small size. 


L0stf0x posted this 24 September 2020

Lol Wistiti, you know me well ..yes I like simplicity in circuits and this one got my interest indeed wink

Rakarskiy I use "Dumbbell winding" at the last circuit. But you are right about winding resistance!, with fewer revolutions! I will check it as you suggest!

Wistiti here is the Cap charging circuit as I did it. Coils winded as shown. I hope its correct

*The turns need change according to Rakarskiy's suggestions for better efficiency.



Chris posted this 24 September 2020

Hey Guys,

Just a heads up, those Caps are Super Caps: 2.7V 10F and 5.4V - 5F, they are super Caps.

Under certain conditions they will be ok, but give them a charge from a battery, they could charge enough to run or months by them selves. Something to beware of.

Best wishes, stay safe and well My Friends


Wistiti posted this 24 September 2020

Hey L0stf0x, thank you my freind. I will try it.

Chris you are right about the cap but if it charge there are no doubts!

Rakarskiy, thank you for being with us!

Chris posted this 25 September 2020

Hey Wistiti,

Yes, absolutely!

We have see this before. It was the same with the Batteries Charging in the Split the Positive Circuit you and L0stF0x mastered. We have the equation:

Q = C x V


W = ½ C x V2



  • Q = Charge, in Coulombs
  • W = Energy in Joules
  • C = Capacitance, in Farads
  • V = Voltage, in Volts


As an example:

  • 10 Farad's Capacitor
  • Start Voltage 1 Volt.


This means we have Q = 10 x 1 = Q = 10 Coulombs of Charge on the Capacitor.


After 24 hours of Run, we have:

  • 10 Farad's Capacitor
  • Measured Voltage of 2 Volts


This means we have Q = 10 x 2 = Q = 20 Coulombs of Charge on the Capacitor.

We have a 2x gain in the Capacitor of Energy, Energy is Coulombs of Charge on the Capacitor. 1 Coulomb of Charge is equal to 1 Joule of Energy. EG: J = 1 Joule, C = 1 Coulomb, then V will equal 1 Volt.

1 Volt = 1 Joule / 1 Coulomb = J / C


Best wishes, stay safe and well My Friends



P.S: Welcome Rakarskiy, good to have you onboard!

Rakarskiy posted this 25 September 2020

This combination is very interesting, the first such placement of windings was performed by Nikola Tesla. All in the concentration of magnetic lines of force at the poles. You should get acquainted with another project on the Internet. A coil for a pulsed unipolar/ bi-monopolar motor was also developed at the time. Everything is based on the topology of correct operation with the magnetic field of the core. With respect.



baerndorfer posted this 27 September 2020

thought about this circuit for days now. think the idea behind is brillant..

because the switching is done on the lowside there is a massive pulse in L1. L2 gets this energy (with some gain in voltage) and transfers it into the second 'stage' where another gain can be achieved - but this time we can gain some more current.

Ch1: Input pulse (from Transistor)

Ch3: This is L2

Ch2: This is the second toroid - we have 2 points for extracting energy! Ch2 is measured under load (25W resistive load)

Ch4 (DVM): superCap charging point.

this circuit can be used for charging batteries with little voltage (it acts like a boost-converter) for example. this is very useful i think. and maybe there is aboveunity somewhere.

the energy consumption from DC source (supercap) is under 6Watts.

thank you guys!




Wistiti posted this 27 September 2020

Hi team! Nice work Baerndorfer! I have some little time this weekend to experiment a bit. I try the last circuit sketch by L0stf0x for the super cap. I don't know why... but again I have to reverse the coil on the base of the transistor. It is an interesting way to use POC. I will try this in other experiment to see the effect compared the way I usually connect them. Anyway I have not try on super cap but it works on 6 v lead acid battery! Thanks you and will love to see others experiment!

L0stf0x posted this 28 September 2020

Wistiti! great to see that works for you too my friend!!

I want to add few points about the original circuit posted by Jagau, that I think will help the replicators to adjust it easier.

First of all do not underestimate the small cup C2 that is shorting the one loop coil L6. As I see it, the L6 purpose is to synchronize (tune) the two ferrites with the main frequency of the circuit. The other coils that suppose to be equal, must be equal, and the length of their leads should also be equal length and diameter of course. Or you can use an induction meter to trim them equal accordingly. If the circuit oscillates ok, the LED1 will light up anyway. But to get to the magic point you will have to light up the LED2 as well (Maybe will not light up with the same brightness as the LED1) but what you need to make sure is that they oscillate with the same rate without interruptions or delays from LED2 following LED1. This may not happen if the capacitor C2 is wrong value. But maybe you are lucky enough and the closed loop without any capacitor will much. Or you can try different width wires or by adding more loops at L6 and checking loop by loop where you get the wanted result.  C1 and potentiometer are there to give the desired range of frequency only. All the game is played at the correct (balanced) coils and C2. Again these are my observations and conclusions from testing it and I maybe wrong.



thaelin posted this 28 September 2020

I just got a shock tonight, while checking how the thing was running, I measured the amp draw again and it was not 50ma as I thought, it was .5 ma! Is no wonder that this thing will run long time on a D cell NiMh. And to boot, burning out led's too. Sadly tho, it still is not charging as it should. Will follow LostFox's idea and rebuild it again with his things in mind. Seven is the lucky number, right?



thaelin posted this 28 September 2020


    Could you share with me the layout of the charger you show a couple of posts up. I wound up the coil as per the pictoral view but it seems lacking to me. I would like to unleash this on a 12v slab removed from a UPS to see if it will have the same results as you. I have some Tip series transistors that should work great for this. Thanks much




Wistiti posted this 29 September 2020

Hi thay. I'm away of home right now but when I get back I will do it with pleasure! And will check the voltage of the battery too!!

Wistiti posted this 30 September 2020

Hi Team!

After more then 2 days my 6v battery goes down to 6,048v. It is still over the 6,038v before running it at the beginning.


I will let it run to see if it continue to drop.


Thay here is the way i connect it.


Im curious how other experiment are going..?



L0stf0x posted this 01 October 2020

Hi Wistiti, I have the same problem. with the charging circuit and the Jagau circuit. The effect is the same but the speed of changes, is depended from the capacity of battery that is used. It charging at the beginning..  after some hours/days will keep the voltage stable... and after some hours/days the voltage decrease. This reminds me the magnet/plasma circuits experiments.., this seems to happen with such circuits.

Maybe a solution would be to alter the circuit to charge another battery, and swap the two batteries to keep going. At the form of split positive for example. When I get some free time I will try to do that..



Atti posted this 02 October 2020


I’m not saying I did everything right. It's just a simple one
I'm a replicator.
Unfortunately, I still can’t report good results. Over time, the batteries slowly discharge. NiCd NiMh and lithium. I took a picture of the shape of the charging current. The preparation was based on the given parameters.I also tried several layouts.
The LED flashes as told but the camera image is a delay.
I wonder what the shape of the charging current is for others.


Rakarskiy posted this 02 October 2020

About working in long-term mode. Try connecting a load to the battery for a short time to discharge it a little. This will change its internal resistance, which should cause an increase in the charging current. (in simple terms, what you have collected, take it to the side). If the system restores the charge again, this will be the algorithm of the self-propelled device. And without battery operation, the system is balanced, so it is discharged and it is looking for its optimal balance point. Respectfully, my thoughts.

YoElMiCrO posted this 03 October 2020

Hi everyone.

I have already done several tests of the circuit.
The operation itself is a saturation auto-oscillator
core, the series resistor that biases the base of the transistor
as is logical it will apply a current in it which in turn
multiplied by the beta will create the collector current.
It is because of this that it will be necessary to adjust said resistor so
saturates it.
The capacitor on the other hand, the main function will be that said
transistor work on common emitter.
As the feedback is positive it will oscillate
at a frequency that will be dictated by the core material,
turns, load, collector current, etc.
If you want a more in-depth explanation, let me know.
This is just an approximation, just that.
On the other hand the apparent battery charge will stop
when its dielectric absorption enters equilibrium
with the energy possessed by the peaks generated by the leakage inductance.
The best way without fear of doubt to affirm that the circuit charges
the battery would be the circuit in this image.

Thanks in advance.


Jagau posted this 05 October 2020

 Hello experimenter friends, back from a long trip. I will read with interest your experiences, a lot of late reading for me,

A small update that will interest many, the starting circuit is still in operation and therefore after more than 21 days, it works perfectly, the voltage is still maintained at 2.5 volts and does not drop.

I would like to know if if mimo and kallefin also have good results too?

Well done Atti and
Thank yoeimicro for your comments

Mimo posted this 05 October 2020


current results:

16th day for my first device with litz wire: 2.246 V.
14th day for my second device with stranded wire: 2.268 V.
However, I had a gradual dimming leaving the light visually fixed. To get stability, I was forced to switch to fast blinking mode.

You get a great result that I would love to reproduce. What are the dimensions of your ferrites (if it's already published, I may have missed it)?


Jagau posted this 05 October 2020

Hi Mimo

Already the sixteenth day for you it's excellent thank you for the feedback Mimo.
My ferrites are Ferroxcuce 3E25 I put the pdf attached, they are the ones you see at the beginning of the thread with a small black tape on top to better hold the cables tight.
We continue it is going very well and I believe that the configuration of the beginning of the thread is the best proof is there.

Attached Files

strape posted this 07 October 2020

Hello everyone .. We replicated this circuit but still without success .. after switching on the voltage on the battery increased but only for about an hour, then everything turned and discharged .. we had it built on a cup core .. today we ordered a toroidal ferrites .. so I'll see when he arrives and remake the whole device from the beginning ... I didn't even have curves on the oscilloscope like Jagau had. Either it all vibrated and the diodes lit up differently but there was a large current consumption from the battery or it pulsed with flashing LEDs but the battery was still discharged .. the only thing we observed was that the frequency of LED flashing increased ..
How nice will the ferrites arrive, I will add further information, including photo documentation ... What is your experience of others with this connection ...?

Jagau posted this 12 October 2020

Hello friends
Back from yet another week of travel.
Still a lot of material to read, I will give you some good news soon.

P.S. hi strape round ferrites are absolutely necessary, keep us informed of your results when you have received your ferrites.


YoElMiCrO posted this 13 October 2020

Hi everyone.


This is the datasheet for the torus i use.



Jagau posted this 13 October 2020

Hi  all experimenters

The next step in this wonderful little circuit is to find out what its consumption is in Watts and you have to bear in mind that this measurement is taken after more than 4 weeks of operation.
Using my DDM we can read a consumption of 4 milliampere or 0.004 A. and this on 2 batteries of 1.2 volts, which gives us a total consumption of 0.004 X 2.463V = 0.009852 watt or 9.852 milliwatts.

wth 2 463 volts

As the precision with a DDM is not as precise as with the oscilloscope here is a measurement taken with my oscilloscope.

taken through a precision resistance of 0.1ohm see mean voltage, we can see that the measurement is even more precise, 4.62 milliampere. 

So now we have 4.62ma X 2.463v = 11.379 milliwatt

When you calculate how long with this type of circuit the two 1.2 volt batteries would discharge

Battery used 300ma 1.2 v X 2

you can see that after runtime of 64.56 hours or near 3 days later the batteries would be dead.

And this circuit is still working after more than 30 days and no sign of weakness.

A further study of this circuit is therefore required, slowly but surely by taking a lot of notes and photos as Chris advises.

The next step will be to find out what is the maximum power it can consume without draining the batteries.


Jagau posted this 17 October 2020

Hell all
A fifth week without failure.
Tell me Mimo is it still working for you?

I think there is also KalleFin who had good results, keep me informed of your results.


Mimo posted this 17 October 2020

@Jagau, and the team

30th day for the one with the Litz wire : 2,07 volts, in fast flashing mode. Slow decline, but 30 days anyway.

The other copper wire device is very weak, but still on in fixed mode. I did not measure it. The best luminous performance is the litz wire (same ferrite toroids).


Jagau posted this 17 October 2020

Hi Mimo

Hello Mimo, I am very proud to hear you say that.
I think there is something we need to investigate for sure, thank you so much for answering I was feeling a bit lonely, but it gave a boost of energy  , I think we are making the future of this thread and what is coming.

If others have good news I look forward to hearing from you.


Vidura posted this 18 October 2020

Hey jagau and all following. I'm glad that some have good results with the circuit. In my first experiment it performed well, rising the battery in a few hours, then I made some improvements, without success, and Couldn't find the same parameters again. I used two toroidal cores out from common mode chokes. Anyway I want to tell that I have not spent much time with the circuit, as I am working on a couple of other projects and don't want to disperse too much. Also my free time is limited, a lot of work to do and difficult economic situation. Maybe at a later moment l will recover to investigate this circuit. Best wishes, Peace for all. Vidura.

strape posted this 18 October 2020

Hello everyone.
I'm still failing. I already have new ferrite toroidal cores. We did everything according to the original plan at the beginning of this thread .. then we turned the winding of the L3 coil. there is always a drop in battery voltage .. we tried to measure on an oscilloscope and I can't find the same curve on the collector as my Jagau .. so what to watch out for? Yesterday I rewound the coil L3 in the same direction as the L1-L2 bifilar and turned the diode .. everything is the same, the voltage still drops. as i found out, the frequency that arises on the oscillator cannot be influenced, the potentiometer can only accelerate the number of pulses until full oscillation occurs .. I will try to make some pictures ..

Jagau posted this 18 October 2020

Hi all

@ Vidura, your comments are always appreciated here my friend. You have done a lot for this forum with practical experiences.

Have you tried this easier to read diagram?

To help you try with a voltage of 2 batteries 1.2 volts at the beginning it is easier to debug.
With your images I may be able to help you better.


strape posted this 18 October 2020

Hi everyone, Jagu

This is what was not clear to me if the winding on Tr2 is also bifilar.
Nowhere did we find out the number of turns. At the beginning it is written on all windings 8 turns and on Tr2 windings with LED 12 turns .. So how is it?
We browsed the internet and saw many pictures of this connection but nowhere is it written koik to turn the threads ..
Thanks again for the provided picture .. I'm going to remake it .. :-)


Jagau posted this 18 October 2020

Hi strape
as you can see in an image already posted on this thtead,


For Tr1 the bifiliar white wire, clockwise you have to adjust for your ferrite, between 6 and 8 turns

for  Tr1   the black wire counterclockwise 8 turns

for Tr2 the yellow wire clockwise goes to the LED and has 12 turns.

for Tr2 The green wire has 8 turns counterclockwise.

I hope this will help


Jagau posted this 19 October 2020

Hi all

I built a second device with the same parameters as the first one and instead of AAA batteries I used a capacitor with the same energy in joules as the 2 batteries1.2v 300 ma.

It works, I have the same effect on my oscilloscope, but the capacitor discharge very slowly.

One observation, LED2 does not blink when I connect the capacitor and if I switch the batteries everything goes back to normal. Since a capacitor does not have the same internal resistance and does not charge and discharge in the same way as a battery the impedance Z of the circuit changes in the same way.

I am investigating the process that will allow it to operate with a capacitor.


Mitza84 posted this 18 November 2020

Hi Jagau how is your setup is still working ? is still keeping the voltage ?i am very curious ,i replicate it and it holds the voltage very long time but is still droping slowly i try diferent ferites and diferent wave forms i have 2 potentiometers one like in setup and one in the base of transistor i try diferent combinations my setup go someware around 300-600 uA consumtion with leds flicering but still not selfcharging i dont know where i am rong. maybe you can help 



Mitza84 posted this 18 November 2020

this is my setup is not conected now i play for days but the voltage still drops slowly dont know why .



Jagau posted this 18 November 2020

Hi Mitza
The circuit should be built like this:

Why do you have 2 potentiometers? only 1 in the schematic.

Did you forget the loop between the 2 ferrites? look at Yellow trace.


Mitza84 posted this 18 November 2020

Hi Jagau thank you very much for the advice i put the loop also if i connect the loop in a way the consumption go high and if i connect the other way the consumption go down and the wave i am getting is similar to yours i will do it the way you showed me now and i will be back with news thank you again Jagau .


Mitza84 posted this 18 November 2020

i think i found the right pulse Jagau  the voltage is rising slowly for now il keep watching for now thanks for the help .


Jagau posted this 18 November 2020

Yes well done
you have the right waveform, I can predict a nice trip of more than two months with this setup, well done.
Another great achievement, you are a guy very attentive to details

And if you expand your waveform you should have like the lower one here

thanks for sharing.


Jagau posted this 19 November 2020

In this type of configuration the only problem I could see is the number of charge and discharge cycles of the battery in fact it is the standard life cycle of a rechargeable battery. At a given time and it depends on the age of the battery it is no longer able to recharge itself,

Mine lasted 64 days and I tried to charge them with a standard charger and it had reached their end of life cycle, by placing new batteries the setup is started again and it still works, the ferrite fuel is there for a few things. I am currently working on a version with capacitor, it is not finished.

Mitza84 posted this 19 November 2020

Yes is the same when expanded many small like almost sinusoidal waves .


Jagau posted this 19 November 2020

Very well Mitza,

in this forum there is a lot to learn and we share what we know.
You just have to know how to listen, there are some who are able to do it and others not.
Why, because you followed the recommendations.

Several other threads here have some fascinating experiences to try.
There are still other great things to learn in this forum. More to come.

P.S. Another observation, take different waveforms at different point in the circuit with your oscilloscope, you will have a surprise and yes we are in another world.


Mitza84 posted this 19 November 2020

Hi Jagau i tried to push the circuit with two 1.2v 40mA small bateries from solar led lamp for garden but over night in 12 hours it drop from 2.6v to 1.2v because i wanted to see if is really charging ,on bigger bateries it last for ages from what i can see i put it on 1000 mA batteryes in the morning and it stayed at 2.5 v in 8 hours and is still 2.5v so did not drop but it has a consumption of 0.5mA from what i can measure in blinking mode for me . what i need to know from you is you keep on blinking mode or you push it a bit more to have constant light ? and for you worked on one battery or you use also 2 batteryes in series?  i will let it run with these 1000mA to see how long will last . And if you change the capacitance you can alter the wave by making it longer or shorter .i used 1uF and 2uF It goes well with bouth.


Jagau posted this 19 November 2020

Hi mitza84

 First of all I use 2 rechargeable batteries 1.2 volts in series of 300mah for more stability.
The critical point of this setup is the pot adjustment.
When the batteries are charging at the beginning the 2 leds flash, the more the batteries charge you will then notice a slowing of the flashing until only one led flashes weakly and the other is stable. As you have the correct waveform at this stage your batteries will be fully charged.
Normally on a different setup other than that one, the batteries only last about 4 days according to my calculations.


Mitza84 posted this 20 November 2020

Hi Jagau ,yes you are right this is indid a very long runner .i will keep under observation longer time to see what happen .thank you for the infos. i hope i can make it on capacitors that would be really cool .but maybe you do it faster .


Jagau posted this 09 December 2020

Hi Mitz

I tried it with a capacitor of the same capacity as a battery that I use and it lasts about 2 days.
A capacitor has no chemicals like a battery.
It is a passive component i.e. it cannot produce ions which produce the recharge of a battery by changing the chemical state of its electrodes. It cannot be discharge and charged at the same time with low voltage.

We can clearly see in Akula's schematics that he uses another process for recharging a capacitor at same time it is used. In his diagrams as I understand it, he uses the capacitor in a first phase, it is the discharge and in a second phase he recharges the capacitor. Thus we produce the charge and discharge in 2 phases

These are the type of experience that I am reproducing now. 


vita_cell posted this 13 December 2020

Hi Mitz

I tried it with a capacitor of the same capacity as a battery that I use and it lasts about 2 days.
A capacitor has no chemicals like a battery.
It is a passive component i.e. it cannot produce ions which produce the recharge of a battery by changing the chemical state of its electrodes. It cannot be discharge and charged at the same time with low voltage.

We can clearly see in Akula's schematics that he uses another process for recharging a capacitor at same time it is used. In his diagrams as I understand it, he uses the capacitor in a first phase, it is the discharge and in a second phase he recharges the capacitor. Thus we produce the charge and discharge in 2 phases

These are the type of experience that I am reproducing now. 



Jagau, hi, do you remember me? Hi again!! Will not write too much, but I have something to answer. As I born in USSR, I speak russian. Akula clearly says in one of his videos (from his Youtube's channel), that you should use high voltage capacitors, above 50v, he says that 63v are good enough, and those are made differently. He says, it is for catch high voltage peaks better (or something like that).  Don't know if it true or not, but I must to say it.


Hope you success running Nekvadrat's circuit on some capacitor (or supercapacitor). I am not expert, but personally, I don't recommend you to use supercapacitors, because of vloss (leak). I have some good flashlights on 10000uF-25v (Jackcon), sharges instantly and works with good bright light for 5 minutes (well it lasts more but light dims) (no overunity or something like that). I did tests some caps of this type, and those aren't equal. Those from Jackcon have very low leackage, while some of NONAME/NOBRAND chinese capacitors of this capacitance have much more leakage while my Jackcon holds charge for days and has very small leaks. My experience with supercapacitors is, when you chare them until it draws no mA and disconnect them, you instantly will lose a lot of voltage. I recommend to testing with good brand capacitors, and well, if it higher voltage, the better.

Jagau posted this 13 December 2020

Yes I remember you Vita cell we met on this site in russia

I have already entered the web address of this site at the beginning of this thread.
Ok thanks for the good advice, I take them all.
For supercapacitor, you are right they are not very efficient and especially they only work between low voltage.
Precisely the high voltage capacitors seem more efficient, always in search mode.
My greetings 

хорошего дня, твой английский хороший и ясный


vita_cell posted this 13 December 2020

Here is some explanation about "добротность" (looks like is "Q factor" in english) of capacitors. He also do some tests. https://en.wikipedia.org/wiki/Q_factor

At 14:42 he concludes: "when we have a capacitor with so big volts hold/reserve, the capacitor's Q factor increases"

At 15:59 he says: "his (capacitor's) Q-fac looks like it is lower, but it is lower because we charged it almost near to his nominal electric potential (volts)."


Here is Akula's explanation at 0: 33:

Akula says: "Those capacitors that can perceive/catch/capture backward high voltage drop/peak(*оэдс), only high voltage ones are valid. So, 400v, 60v capacitors. Right now I have 63v one and it works fine. For selfrunning (catch voltage peak) those are better. Are better because those aren't paper made, those have bigger distance/space between plates. Those are capable to catch and charge with high voltage peak that are up to 50-60v, while those don't create additional load. All 16-35v capacitors are weak for this, those above 50v catch this backward peak better."


эдс = it is electromotive force


оэдс = not sure how to translate it in english, but it is backward electromotive force drop, looks like one from autoinduction.

Jagau posted this 13 December 2020

Very interesting, thank you Vital cell I am going to study seriously.


I love it a lot, it gives me ideas

Tom Bearden called this massless current transfer


Members Online:
Since 03.08.17 - Widget: 27.11.18
Your Support:

More than anything else, your contributions to this forum are most important! We are trying to actively get all visitors involved, but we do only have a few main contributors, which are very much appreciated! If you would like to see more pages with more detailed experiments and answers, perhaps a contribution of another type maybe possible:

PayPal De-Platformed me!

They REFUSE to tell me why!

We now use Wise!


The content I am sharing is not only unique, but is changing the world as we know it! Please Support Us!

Perhaps a Bank Transfer?

  • Bank: Bank of Queensland
  • Name: Christopher R Sykes
  • BSB: 124-001
  • Account: 21580359

Thank You So Much!

Weeks High Earners:
The great Nikola Tesla:

Ere many generations pass, our machinery will be driven by a power obtainable at any point of the universe. This idea is not novel. Men have been led to it long ago go by instinct or reason. It has been expressed in many ways, and in many places, in the history of old and new. We find it in the delightful myth of Antheus, who drives power from the earth; we find it among the subtle speculations of one of your splendid mathematicians, and in many hints and statements of thinkers of the present time. Throughout space there is energy. Is this energy static or kinetic? If static, our hopes are in vain; if kinetic - and this we know it is for certain - then it is a mere question of time when men will succeed in attaching their machinery to the very wheelwork of nature.

Experiments With Alternate Currents Of High Potential And High Frequency (February 1892).