Sorry I don't have pics yet, but you can build it yourself and see!

Preparation Time: 1 hour

Requried Materials:
   2 Cardboard Tubes (from the insides of paper towel rolls)
   1 Roll (200 ft or more) of 22 Gauge Hookup Wire
   1 Wire Stripper
   1 Roll of Aluminum Foil
   1 Roll of Electrical Tape
   1 LED, any color
   1 Oscilloscope
   1 Function Generator (or a 555 timer)
   Connecting Cables (I used alligator-clip cables, with just a wire going into the function generator because I was too lazy to get the correct cables)
   
1. Roll up a cylinder of aluminum foil that is slightly smaller than the tube, and place one of these inside each cardboard tube. Use as much foil as will fit inside the tube.

2. Wire a coil of wire around each of the tubes. These coils should be as identical as possible. Secure the wire on both ends with electrical tape. Leave a few inches of wire going off each end of the coil, and strip the insulation off of the last half-inch.

3. Place the coils side by side, right next to each other.

4. Connect the positive (+) side of the function generator to one side of one of the coils. DO NOT CONNECT THE NEGATIVE (-) SIDE TO ANYTHING. If there is a negative lead, try to keep it as far away from the coils as possible. Better yet, disconnect it completely.

5. Set the function generator as follows: low voltage amplitude (like 1 volt max), square wave, a frequency around 1.7 MHz, and minimum duty cycle (you may have to set it to maximum duty cycle, then invert it). On some function generators you may have to increase the duty cycle a little to get to the higher frequencies.

6. Connect the scope probe to the coil that the function generator is connected to, but to the opposite end. Leave the scope's ground probe disconnected. You will have to adjust the oscilloscope until you see the waveform show up on the screen. It should look like a sine wave. But we are sending in square wave pulses, right? This is the effect of near-resonance.

7. Adjust the frequency and duty cycle until the sine wave reaches a maximum amplitude. If you need to, increase the duty cycle a little more so that you can reach higher frequencies. Just find the "sweet spot" where you get the maximum output.

8. Increase the input voltage amplitude until the output is at the maximum that your function generator can run at. WARNING! Do not increase the voltage too much or you will ruin your function generator! I would stop when you see around 30 V output.

9. Connect an LED to both ends of the other coil. It doesn't matter which end gets + or -, since we will be feeding it an AC signal. If you want to, connect the other scope probe or a voltmeter (set to AC volts) across this coil. WARNING: Never connect two scope ground connectors to two different places in the circuit!

You should see perhaps 1V AC across the second coil.

10. If it isn't working already, try pressing your hands flat against the surfaces of the cois. This will help boost the capacitance in the circuit, and drive the resonant frequency down a little further.

That's it! You are providing real energy to an LED to make it light up! Where is this energy coming from, you ask? From the function generator? I don't think so! Remember, we left the negative connection of the functiong generator disconnected. The primary circuit is an open circuit, there is no complete path for electrons to flow through! That means, input current is zero, so input power is zero. Output power is something more than zero!

-omnispace

...
The primary circuit is an open circuit, there is no complete path for electrons to flow through! That means, input current is zero, so input power is zero. Output power is something more than zero!

-omnispace

The circuits act as radio antennas in near field. It is easy for any ham radio operator to power such a circuit from a distant one. Generally a transmitting radio antenna is not a looped circuit and nevertheless current is consumed through the space impedance, or capacitively or inductively with other circuits in near field.
You can't say the input power is zero without careful measurements using particular HF means, especially when square signals are used, such as a selective HF power meter (to measure power at each harmonic component) and selective VSWR meter.
Can you provide details on the measurement protocol ?

Sorry but I'm not following... may be missing something here...

you're pumping 30V pulses into a coil at freqs over 1.7MHz,
and getting out 1V AC at what freq?
In any case, it sounds like you're getting 1V that pulses the LED... If you would connect 2 LEDs, one p-n and the other n-p, to the same
wire terminals of the secondary, you would rectify the AC and the LEDs would blink in turn, wouldn't they?
But still, if we're measuring 1V AC and that is powering the LEDs,
would you not be much better off pumping your 30V pulses directly into the LED?

Seems to me you're still using quite a lot of power to get out very little...

You see, I told you I wasn't following

Yeah weird eh, I also noticed that my university electrodynamics books said very little on RF and related stuff...
I was lucky enough to find an old 1950s textbook on radio waves and transmission circuits, which helped me out a lot back in the day.
But it would be nice if they'd teach this in high schools or college... seems to me you've missed something if you don't know at least some about it...

anyways, how much are you getting out extra compared to your input?
You said a little but how much is that?

I got 1 VAC out, with about a -5 mV DC bias.  And enough current to light the LED.  I haven't measure the input or output power, if that's what you're asking for.

I did try grounding the core when I was first trying to resonate the primary coil by itself.  That didn't seem to have any noticeable effect though.  I was also checking for any kind of significant voltage developed on the core (for TPU application), but was unable to get anything more than some weak noisy signal.

I don't think I mentioned this yet, but the reason I set this thing up was to try to recreate the MIT "WiTricity" effect.  It was only when I saw the huge gain on the primary coil that I thought "hmm...overunity?".

But for WiTricity, I'm fairly certain they were using sine waves (AC).  I'm hitting the transmitter with short, quick pulses (RF) and "tapping" it into resonance.  Isn't the pulse, the sharp gradient, considered the basis for all radiant energy?  I hope to far exceed that 40% efficiency someday.  My first goal is around 90% efficiency.  Then we'll see if it can be pushed "over the edge" and become overunity.

Dear omnispace,

It would be nice if could let us know what happened to your experiment. Sorry I am asking after 7 years.

--- Mr. Teslonian