The Work of Andrey Melnichenko

Author Topic: The Work of Andrey Melnichenko (Read 11033 times)

lfarrand	Re: The Work of Andrey Melnichenko « Reply #25 on: 2022-11-07, 10:28:49 »
Full Member Posts: 163	I wouldn't really call it a device at this stage. A collection of components arranged haphazardly would be a better description for what I have. Nevertheless, behold: The magnet wire spool (yellow circle) is L1 and the Litz wire winding (green circle) is L2. I'm in the process of tidying things up and removing the unecessary bits so I'll post a follow up photo later.

F6FLT	Re: The Work of Andrey Melnichenko « Reply #26 on: 2022-11-07, 10:57:42 »
Group: Experimentalist Hero Member Posts: 1909	@lfarrand Thanks for the picture. I was mistaken, I had seen quickly the diagram of the Don Smith device and I thought that it was what you were doing, so I was surprised by the high current intensity because the metglass core is closed, there can only be a very weak flux coming from the outside. But with Melnichenko, this is normal. I had modelled his setup with LTspice some time ago, without noticing anything abnormal, but well, it was a simulation so not surprising. Only experimentation can show something if his patent really works so I'll follow your results. --------------------------- "Open your mind, but not like a trash bin"

lfarrand	Re: The Work of Andrey Melnichenko « Reply #27 on: 2022-11-08, 21:41:42 »
Full Member Posts: 163	I built the circuit that I posted earlier. The load and my switching board are both powered by a 12V 7AH LiFePO 4 battery. The switching board is powered by a 15W AC/DC adapter, which is an inefficient DC-AC-DC conversion. I could probably go DC all the way from battery to board, but going through an AC inverter allows me to measure the power usage a little easier. I'm switching the coil on the left of the ferrite rod using my MOSFET switching board. Power usage of the switching board when idle is 4.7W. The fans consume 0.5W each. Switching at 100kHz takes power usage up to 7W, so 2.3W for switching at 100kHz doesn't sound too bad. I got this interesting waveform on the scope. The yellow trace is the left coil and the pink trace is the right coil. The left coil is connected to the power supply whereas the right coil isn't connected to a power source directly and is instead induced by the left coil. The negative spike on the yellow trace is clearly visible. I'm not sure what to make of the waveforms yet. I'm going to reflect on them and post some thoughts later. I also need to test DC power consumption with and without the right coil to see if Melnichenko was right in saying that the diode should prevent the right coil from consuming from the source. If he's correct then I'd expect power consumption to remain the same. If power consumption changes then either I've made a mistake somewhere he's wrong.

lfarrand	Re: The Work of Andrey Melnichenko « Reply #28 on: 2022-11-08, 21:54:31 »
Full Member Posts: 163	Another interesting effect I observed was when I placed a coil with windings perpendicular to the left exciter coil and then put a ferrite rod in the absolute centre protruding downwards about 2cm into the coil. The following waveform appeared on the scope. The frequency matched the PWM frequency of the switching board, but it's now appearing as a pure sine wave. If the ferrite rod was moved up, down, left or right away from the position shown (relative to the red coil) then the signal on the scope disappeared. Moving the red coil with the ferrite rod as one unit towards or away from the exciter caused the signal to increase or decrease in amplitude but the waveform shape remained the same.

lfarrand	Re: The Work of Andrey Melnichenko « Reply #29 on: 2022-11-09, 13:14:04 »
Full Member Posts: 163	Quote from: lfarrand on 2022-11-08, 21:41:42 I also need to test DC power consumption with and without the right coil to see if Melnichenko was right in saying that the diode should prevent the right coil from consuming from the source. If he's correct then I'd expect power consumption to remain the same. If power consumption changes then either I've made a mistake somewhere he's wrong. So I tested this theory out. I added a switch to the secondary coil (right hand side) so that I could open/close circuit it at will. I also tested with and without the diode + capacitor. What I found was that the secondary coil consumed no additional power from the battery. Without the diode + capacitor the secondary coil consumed additional power from the battery. It seems that Melnichenko was correct when he said that no power would be consumed from the power source if you introduced a diode to block the positive current (establishing the magnetic field) and only allow the negative current to flow on the secondary coil. Interesting...

lfarrand	Re: The Work of Andrey Melnichenko « Reply #30 on: 2022-11-09, 13:23:23 »
Full Member Posts: 163	My next steps are to test the following: Move the secondary coil to a separate ferrite core instead of sharing one with the primary Add more secondary coils Use 2000 strand Litz wire as a secondary coil During my testing I've found that the primary capacitor will charge to 200V and the secondary coil will charge to 100V when running 12V through the primary coil at 100kHz. If it is possible to add more secondaries without affecting the power consumption then that would be a big deal.

lfarrand

Re: The Work of Andrey Melnichenko « Reply #31 on: 2022-11-09, 13:48:30 »

Full Member

Posts: 163

My coils are quite rudimentary. They are 50 turns each wound in the same direction, 230uH inductance, 0.13ohm resistance and some low pF capacitance.

The capacitors are Panasonic 18uF 1100V EZPV1B186MTB.

I also tested different diodes for capturing the inductive kickback. Here's my relative comparison that some might find useful.

Model	Left coil voltage	Right coil voltage	Total
C4D02120A	214.5	100.3	314.8
UF5408	216.3	94.8	311.1
SF1600-TR	219	91.9	310.9
MUR8100EG	213	97	310
RHRP15120-F102	211	97.7	308.7
BY203-TAP	217.6	90.7	308.3
IDH085120AKSA1	189.5	104.2	293.7
VF25-12X	203	82.5	285.5
GB10SLT12-220	175.8	100.7	276.5
GC10MPS12-220	172.5	91.9	264.4

As a result of the test above I am using Wolfspeed C4D02120A SiC Schottky diodes in my circuit.