I recall everybody that the MEG I work with is the same as J.L.
Naudin 3.1 one:
All my mast year experimentations here:
There is only two differencies with Naudin'one: magnet and power
supply is not the same. In effect my magnet was a stack of 10 disc
rare earth element magnets, 5mm diameter. And my power supply is a
fix 30V output, 1A max. Specifications are those of Naudin.
All my tests were negative, power calculations were very poor, with
COP less than 0.5, no overunity. A big problem was that I had not
sinusoidal signals for currents and voltages at all. And when my
load was a MOV, there was a phase angle and calculations have proven
that power out was little.
I have bought a new magnet and made new experiments. My magnet has
been specifically designed to fit the MEG C core (very expensive
magnet), in rare earth elements. Density flux is 1.2 Tesla, less
than saturation core in theory.
I have taken in account the fact that as described in Bearden
papers, input coils must be feed with currents signals that produce
magnetic fields that contrary the one of magnet. In effect, one of
my input coil was not correctly oriented. I corrected this.
All experiments I made these last days are negative again.
I mesured input signals in the primary coils, I signals are not what
they should be. They should be square signals, 30V and 0V, when one
primary coil is fed, the other one must not be.
My control board is exactly the same thant JL Naudin's one, my
primary coils are wounded with the same number of turns, and the
same wire that he specified, but my input signals are not the same
I saw that input signals were excatly what I waited for when
changing operating frequency (frequencies beetween 1700Hz and
5000Hz); but completely deformed at JL Naudin's control board
operatingfrequencies (between 17 000 Hz and 50 000 Hz). I mesured
plate 30V and some pics to -80V at higher frequencies, not enough to
have a plate (and it's not 0V).
Induced back currents in primary coils (when FET basculates) is the
cause of these perturbations. I have tried too solutions.
First one, a diode in parallel with my primary coils (in the
opposite way). But too many currents was flowing through, and this
lead to an excess input consumption. I have added a resistor with my
diode, but when the resistor was some kOhm, it was too much in
comparison with primary coil impedance at operating frequency, so
induced back currents were not going through the diode but through
the coil, and a too small resistor led to a high power consumption.
The input signal was correct with a diode.
Second solution: a capacitor in parallel with my primary coil
instead of a diode; few nF (ranging from 10 to 30 nF). This
stabilised my negative pics at higher frequencies and my input
signal was then acceptable: plate 30V and half sin from 30V to -60V.
I used this montage.
With this correct input signals, my output was different: always
sinusoidal output, for a resistive load or a MOV load.
So the reason why last year I never mesured sinusoidal waves is that
control board is not properly designed for operating frequencies
from 17KHz to 50KHz; but Naudin says he measured his sin waves at
20KHz approx. with his control board. How has he done, it's not
With my correct input signals, my correct primary coil orientation
so magnetic field of the coil contrary magnet field, and my new
powerfull magnet all was set up to be correct but ... negative again.
All experiments (more than 30 experiments) lead to that conclusion:
1) With a resistive load, current and voltage are sinusoidal wave in
phase. Power output is a quarter power input on each secondary,
which gives output power half of the input. Power calculations are
easywith such phased waves.
Pout = Umax²/(2*R) = Ueff²/R
Pout= Umax*Imax*cos(phase)/2 = Ueff*Ieff*cos(phase) with cos(phase)=1
These formulaes give the same output power each times (with little
My resistive load is 120kohm(0,5Watt) and 12ohm (10Watt) series
resistors, the second is used to measure current.
chA and chB are probes on oscilloscope:
(very bad transformer)
Voltages are not too high: 600 to 800 V pic to pic.
2)With a MOV load (I tried with one 420V MOV, 2 420V MOV, 3 420V MOV
and 4 420V MOV in series, to that was the same as varying MOV from
420 to 1680V) all signals are sinusoidal, but there is a phase angle.
If I insert a 120 kohm resistor in series with my MOV (and my 12ohm
resistor used to measure currents), phase angle becomes very little
(which is logical if you draw a phases in a complex plane: a big
resistor is a big real impedance and phase angle comes from the
complex impedance of MOV, whoch is little). But with a 120kohm
resistor in series, measurements are the same than with no MOV and
Phase angle is constant when varying frequency.
When varying frequency from from 17KHz to 50KHz, power output was
reducing, and so power input. There is no current pic as claimed on
JL Naudins website, so no "tuning". The less the frequency is, the
more current and voltages I have.
And there is no tuning as claimed by Naudin " the working frequency
must be tuned so as to get a pure sine wave and the max amplitude at
the output (>1KV peak-to-peak loaded),".
Voltages vary from 800V to 1400V pic to pic. There is a resonance
effect with MOV (MOV have a capacitor value) but the operating
frequencies used does not reach pic resonance.
Currents are much more higher than with a pure resistive load, but
with phase angle, power output is:
Pout= Umax*Imax*cos(phase)/2 = Ueff*Ieff*cos(phase) with
and this give Power output varying from 1/10 to 1/4 of power input
(depending of frequency).
Again best COP = 0.5 approximately.
But ther is no power gain.
So, what I gain in current I lose in phase angle.
Naudin says "the voltage and current are in phase as shown in my
scope pictures above,
- a "conditionned" RLoad (100 Kohms, non inductive carbon, 5Watts)
or a MOV (Metal Oxide Varistor) is REQUIRED for getting the output
datas measured above,"
Again measurements are contradictory to Naudin, where output current
and voltage are in phase with his conditioned resistor, but with 9W
neon too. With a neon, I obtain phase angle again.
This is logical, these non linear components give birth to a phase
angle. Naudin never give the schema of his load measurements?. He
gives us oscilloscope pictures, but does not give us what he
measured on his channel of his oscilloscope. This is very important
as to understand what I did to have no phase angle with a neon!
Naudin says "The current has been measured with a 10 ohms ceramic
and non inductive resistor . the same resistor and the same method
of measurement has been used for input and also the output"
So, he measured currents with his 10ohm resistor (I used 12 ohm
resistor). But for the rest? Has he measured both secondary coils
linked together, only one secondary with a neon and his resistor in
I bought batteries to try wether power supply was a problem, but
this changed nothing.
All Naudin claims are contradicted with a replication of his
experiment. He is the only one I know claiming a MEG success with
published measurements; but measurements are not reproductible with
the same device than his. His control board does not fit his input
signals. His loads do not fit his output signals; nothong is right.
I waited for one year to have an answer from him, but i had nothing.
Do you know another successfull overunity MEG construction? The only
thing I can read here is questions, but I don't read any success.
Maybe I have made a mistake?
When reading Bearden's Patent, we can see that no specific load is
needed (conditionned Rload or MOV), and patent says that secondary
output coils can be wounded with less turns that input primary
coils, so to have a very little voltage, and this leads not to a COP
loss for the MEG, so claims of voltage that must build up have no
Bearden does not reveal his scheme for his MEG, so we can not say
wether it works or not, but for Naudin it is possible. Why does not
he answer direct simple questions? Maybe because this is a big LIE.
I don't say that MEG is a lie, but Naudin replications may be. If
not, why doesn't he give us clear answers about his measurements? I
cantoo take pictures of phased voltages and currents, with high
current values, and saying you: that's my MEG results!! There is
But if I give you an exact way to reproduce my experiment and you
don't find the same thing you can say that I am a lier. But if I
give you only some partial results, not everything... and a special
conditioned Rload, and a magic neon, then I can tell you that you
experiment is not the same than mine, so I am the only one to have
Simple question to make the points? Who, in this list has had a
overunity MEG success in his experiments?