Can we demonstrate over unity energy?

[D.R.Jackson]

OK, ....... etc,

Regards,
Pm

Partzman;

I had dismissed this circuit and put it away then something occurred to me,
so I figured I would try that and dismiss it and be done with it. 
Yet I am wondering now?

I set the simulation to skip the initial operation point solution
so that the capacitors would not be pre-charged, and offset the
start up point by 10ms to skip past a transient spike. 
And ran the simulation for 100s.   .tran 0 100 0.01 50u uic

Below is the result I got, and so I took snap shots of the waveform
averages and added them to the circuit snap shot.

I would not believe this unless you looked at it and ran it. 
Doing what you do with LTSpice.

As for the way it appears at the moment this is small in magnitude
on the power scale.

simulation time 99.99s

V1 = 199.89µW  19.987mJ

V2 = 106.15pW  10.614nJ

C1 = * 362.88µW  * 36.284mJ

So this is the result I got after the 100s run. 
I went on to 200s and it will run out that far without loosing steam.

I removed the DC current path from the circuit, and C4 is
being charged by the alternating current on the emitter.
When I remove C3 this circuit will not do this.

106.15pW + 199.89µW  = 199.8901062uW

362.88uW / 199.8901062uW = 1.8153975046514833459025897500874 = 1.8

199.8901062uW / 362.88uW = 0.55084354662698412698412698412698 = 0.55

The spice file is below.

Oh you can increase C3 on up to 20u and find that the output increases into the 1mW range. 

Not really sure how to interpret this at the moment however.

[partzman]

Partzman;

I had dismissed this circuit and put it away then something occurred to me,

[snip]

Not really sure how to interpret this at the moment however.

DR,

I am busy at the moment with a project so I don't have a lot of time to analyze your latest results but I have some suggestions.  I would recommend viewing only the last two cycles of your 100s scan by starting at 99.998s and ending at 100s.  This way, any circuit parameter you wish to plot takes much less time to display thus speeding up the analysis.  In this manner you can now analyze the source of the energy in your circuit by focusing around Q1 using KCL (Kirchhoff's current law) as this device is the nexus of all the circuit's potential energy sources. 

Also, I do not quite understand your "C1 power = V(vc2_out)*I(C2)".  Can you extract real power from C1 under these conditions?  Replace C1 with a resistor whose value is V(vc2_out)rms/I(C2)rms and then check the results against the input power.

Now, I'm not sure if you are familiar with T Bearden's discussion and analysis of the "Fogal transistor" but your circuit with C4 in the emitter of Q1 would seemingly model this device.

Regards,
Pm

[partzman]

OK, the most efficient solution would be to use PWM or pulse width modulation.  This is the technique that most if not all class D audio amplifiers use to achieve high power with high efficiency.  Basically, a high frequency square wave oscillator has it's duty cycle modulated by a lower frequency source and a low pass filter is then used on the output to attenuate the high frequency.  In your case, this low frequency would be a positive triangle wave but it could be any wave shape unipolar, bipolar, or complex as in music content. 

There are really low cost class D amplifier pcb modules available from China thru Ebay which would be cheaper and easier than trying to build your own.  You would then only need to provide the power supply and input signal in the wave shape of your choice and connect the load.  IIRC, I have purchased 100w class D stereo modules in the past for ~$15.00 and free shipping.

Hope this helps,
Pm

Forest,

Here is a relatively simple solution for generating linear currents needed for a Figuera device using split power supplies and gate drives phased 90 degrees from one another.  Nearly all the energy drawn from the bottom supply is returned to the top supply resulting in overall high efficiency.

Regards,
Pm

Edit:  There is a more complex circuit which would re-circulate the collapsing inductive energy to the source supply.  I will post if you are interested.

[forest]

partzman
Can you post ltspice asc file ? I don't know how did you created those "waveform" windows.

[partzman]

partzman
Can you post ltspice asc file ? I don't know how did you created those "waveform" windows.

Forest,

I first take the sim results, save it to the clipboard and then paste it into MS Paint.  I then create whatever plot waveform window in LtSpice I wish to use and with a snipping tool (Win10) transfer it to the clipboard.  I then paste that into the image in Paint and continue until all windows are placed.

I've attached the asc file below.

Regards,
Pm