Step Charging Part 1

Hi Juan,

Thanks for writing! Yes, I have worked with the “Avramenko plug”. Actually, I thought I invented it once. About two years after I thought of this myself I found the same circuit on the Internet by this name.

Below I will try to illustrate some of the circuits I used, when I was “discovering” this thing for myself.

    0        metal ball (6″ copper toilet float)

    |        1 foot copper wire

    3        Tesla coil (240 turns # 22 on 6″ PVC tube each spaced 1.5mm apart)

    3        ”

    3        ”






-||–+–||-   Two capacitors in series

|         |   Connected to

|         |

-|>|-+-|>|-   Two diodes in series


    ///       Ground

This circuit can be made bipolar as well, just make a mirror image below the circuit where it says “ground”.

Although I ran the Tesla coil with a MOSFET circuit I also noticed the capacitors would charge with the battery disconnected from the driver circuit. Presumably this was from external RF reception but I am not sure.

The capacitors were 470 uF 450 Volt and they would build up quite a big charge very quick. If nothing else this circuit is a good way to charge capacitors.

I did find, after exhaustive testing and measurement, that the energy in the capacitors never exceeded the energy spent feeding the MOSFET driver. This was not over unity. It was found that even at resonance, the current draw to the feeder circuit was proportional to the charge on the capacitors.

I will note that fast diodes were necessary in this circuit since it operated at about 750 kHz. I used 1N4948, a fast 1kV diode.

The capacitors would also charge with ground substituted by another metal ball. They would even charge (though slowly) without any ground at all, just one diode in series across the free ends.

I  also used a function generaor’s (15v p/p) output connected where the ground teminal is placed. A 555 timer works well too if it is tuned right. This charges the capacitors too, to over 1000 volts (in series) if you are feeling reckless.

If a 555 is used, two diodes must be placed from the output (pin 3) to +v and circuit common. They will point “upwards” in the schematic (up from ground to pin 3, up from pin 3 to +v), to pump excess EMF (overvoltage) from the Tesla coil back to the battery. This recycles that energy instead of wasting it.

At resonance the back EMF will disappear and the diodes will no longer be necessary to protect the 555. However, due to instabilities and frequency shifts they should not be removed.


If I run a Tesla coil in discharging sparks, I can use a circuit

    0       6″ copper toilet float


-||–+–||-   Two capacitors in series (0.1 uF 400V)

|         |  

|-www–Ne-|   Neon lamp with 47K resistor

|         |

-|>|-+-|>|-   Two diodes in series



This flashes the neon lamp or even keeps it lit. This circuit will work up to hundreds of feet away.

If another Tesla coil is connected like the first circuit above, and is resonant to the emitting Tesla coil, this will mimic Tesla’s patent and transfer energy over a large distance.

Tesla used that setup to light 100 50-watt bulbs (if I remember right, it was several kilowatts) at a distance of five miles from a small coil in his laboratory. There were no diodes to rectify, but just a resonant circuit.

The small circuit also works near a metal plate that carries HF AC.

—–||—-   Capacitor 

|         |

-|>|-+-|>|-   Two diodes in series


     3        Tesla coil (110 turns # 22 on 6″ PVC tube)

     3        ”

     3        ”






     ///       Ground

This works, too, even without ground.

With a high voltage (>2kV) capacitor a xenon flash tube will flash using this circuit. Two of the “Avramenko plugs” can be used, on each end of the coil. This needs no ground at all.

Performance is better if metal plates or spheres are connected to wires wrapped around the tubes. This is not a complete circuit, just an electrostatic effect. Jean has done this same thing and placed it on his website, so we have probably all seen that.

The final circuit I tried was very unusual and behaved strangely. It was an attempt at over unity that did seem to self-run. Sometimes it would last for days but the longest run was four months, obviously more than the battery had.

   +   +——+——–+———                                            

       |      –        –        |                       

       _  |   ^ d      ^ d      |       

   12  –  |   –        –       -+-                      

  volt _  |   |        |       | |                      

       –  |   |   sec  |   |——-|                      

       _  |   +-wwwwww-+    |  4 8  |                      

  lead –  |   | ==T1== |    | lm555 |                    

  acid _  |—–mmmmmm—-+-|3     2|-|                      

       –  |   |   pri  |  | |      6|-|                      

       _  |   |        |  | |1     5| |                     

  cell –  |   –        –  | |——-| |                     

       _  |   ^ d      ^ d|  |vr1/ |  |                  

       –  |   –        –  —–www—-|                  

       |      |        |     | /   =  =                  

   –   +——+——–+—–+—–+–+                                             

                                   c1 c2               

 This is the odd circuit. It used a small lead acid cell, 12 volts. This battery was coated in a layer of aluminum foil, taped together so it would all conduct. The battery fed a 555 oscillator circuit. The output of the 555 was fed through the primay winding of T1 to the al-foil “battery plate”. This battery-capacitor induced a voltage through T1 that was fed through bridge recifier. This is illustrated as four diodes labeled “d”.

This circuit was tuned by vr1 to resonance with T1 and the capacitance of the battery plate. At resonance it would run for a long time, even feeding a flashlight bulb for some months. (connected in series with T1 secondary). T1 was wound with 1:1 turns ratio on a ferrite E-I core.

c1 was 0.1 uF and c2 had to be messed with and chaged to get one that would give frequency of resonance. It was about 680pF. vr1 was 200k-ohm twenty-turn variable resistor.

This 555 arrangement does not use discharge pin 7. The additional pull up resistor necessary to use pin 7 constitutes wasted current which is to be avoided. Using pin 3 instead solves this and makes the circuit more efficient. Also, TC555 or C555 (MOS semiconductor) should be used instead of NE555 (bipolar transistor). MOS is far more efficient. c1 stabilizes oscillator frequency.

I will divulge the secret of this device, that the thing that made it run so long with this circuit at resonance was the magnets around the battery, projecting fields through the electrolyte. I used big ceramic magnets and made a “sandwich” with the battery in the middle.

Magnetic fields alter the trajectory of all charged particles, including ions in solution. I used this effect to the advantage.


Finally, I built one last circuit that puzzled me for a long time. I built it after I started reading Tom Bearden’s papers. It appeared to me in a dream.


         –   |   |

—-|    ^d  |   |

|   * s1 –   |   Ne

=c1  /*–+   =c2 |

|   *    –   |   3

—-|    ^d  |   3

         –   |   |


The left side of this circuit is a capacitor connected across the two open contact points of a SPDT reed relay (s1). The center (toggling) contact of the reed relay connects to the center of the diodes of the “Avramenko plug”. This charges a capacitor that is discharged by the Ne lamp in series with a 47K resistor.

Diodes d can be single 1N4948 but I found it was better to use many in series, like a chain of 10 diodes in series for each “d” in the schematic.

c1 was a 0.1 uF 25kV oil capacitor charged to about 600 volts maximum. The oil capacitor was used because it has very low leakage and will stay charged for a long time.

Diodes d were 1N4948. c2 was 0.047uF 630V polypropylene dielectric.

Brushing a magnet over the reed switch (to toggle it) would make the neon bulb flash. A switched coil that actuated the reed switch much faster than this was used and a lot of energy came out for the size of this circuit. c1 did not discharge very quickly, it became obvious that c2 was getting more charge on it than c1 contained.

While under certain circumstances this circuit will measure over unity, (Tom Bearden says it will do this, too) solid state switches don’t work to substitute for s1. Read Bearden, he can say why better than I can. One needs a physical disconnection/connection at s1 to get this effect. I have used MOSFETS and other things to try and substitute for s1, but the results are not good, the extra charge seems to get lost in them.

With mechanical switching I have noted o/u in some setups, but capacitive coupling between c1 and c2 can destroy the free energy effect and run down source capacitor c1.

Apparently Mr Bearden has applied for a patent of this circuit, or one like it. He calls it “step charging” the capacitor, as each “click” of the relay notches up the charge on c2 by an increment.

Bearden also notes that step charging has been mathematically shown in graduate physics texts to be able to charge a capacitor without entropy – that is, without the disordering (loss) of the original charging potential. This circuit has been mathematically shown to be capable of o/u, apparently, but no one yet has gotten around the mechanical switch and made it solid state. Even Mr Bearden has not been able to, but that is no grounds for discouragement.

In the circuit, one can see that there is never even a physical complete circuit between the plates of c1. How, then, will it ever discharge? The charge “spray” (as Bearden calls it) is rectified by the diodes without any electron flow. This charges c2 without discharging c1. I personally will testify to this, I have seen it myself. This effect only occurs when the capacitive coupling between c1 and c2 is broken.

If the output of c2 could feed a coil that spun a disc with magnets on it, and the magnets turned and flew by the reed switch so as to trigger it many times, maybe this circuit would be capable of self sustained o/u action. The reed switches are only good to about 500 volts though, and the output of this circuit is proportional to the square of c1’s voltage.

This entry was posted in Graham. Bookmark the permalink.