Pulling energy from the ambient energy field using a coil capacitor

[itsu]

Oops, you are very right Gyula, i missed that.

So 68KOhm is the load,  thanks.

Itsu

[Jack Noskills]

Thanks for the new scope shots Itsu, very informative! Two different currents at coil ends (that are not connected to anything) confirm the presence of magnetic field potentials. The input current from FG seems to be a blurred sine wave with some spikes in it. The sine part is not coming from FG which is feeding square pulses only the spikes are created by the FG when capacitor is charged. This can be verified easily by reducing the duty cycle below 1%. To get rid of it connect the same ends if output coil pair together so that result is two parallel coils not connected to anything.
The blur is caused by high frequency oscillation which we already saw when blocking diodes were used. You should be able to tune to this frequency, make a sweep using 50% duty cycle.
So what has been confirmed in my mind so far:
-   [/font]Why a coil capacitor with turn offset is better
-   [/font]Capacitive pulsing is a working solution
-   [/font]Charge collecting capacitor is charged to a higher voltage than what is around it. Either serial or parallel connected capacitor can be charged. Ground connection will pull the charge out from the capacitor.
-   [/font]50% duty cycle and shorter 10% pulse can create the resonant rise effect.
-   [/font]Larger surface area in the energy collecting coil capacitor (copper tape) increases amperage.
-   [/font]Magnetic field potentials are present in the output side.
Putting next some formalized test cases (TC), maybe it helps to see some structure and what is left to complete phase one of this project.
Using the 50% turn offset system pulse it with 1kHz and 10% duty cycle pulse. Connect opposite endpoints of different coils with safety spark gaps. Measure electric field potential across one spark gap and current from either side of the other spark gap. Take voltage and current from the driver as well. Then do the following test cases:
TC0. Measure the capacitance of both coil pairs. Verify that coil pairs are separate from each other.
TC1. Air core without capacitor C2. This was already tested without current probes with 0% turn offset system. I think it is good to test it again with 50% turn offset and see the current also to get a baseline result.
TC2. Add C2, use variable capacitor if possible. As capacitance is increased the oscillations in the output side become stronger. At some point amplitude increase stops. What is the capacitance at this point ? Scope shot.
TC3. Remove C2 and insert ferrite core. Scope shot.
TC4. Keep ferrite and add C2. Again look for the point where the amplitude increase stops. Scope shot.
TC5. Decrease the duty cycle of the pulse so that it lasts one half cycle of the resonant frequency, take scope shot. Then reduce it until oscillations are still present. Use large enough C2 so that current is stopped while C2 is still being charged (current reversal method). What is the length of the pulse ? Scope shot from this also. Use system from either TC2 or TC4 in this test.
These tests will show what is the best pulsing method and system. The effect of using C2 is very important. If it works well then it can be used for power control and also in electric motors.
TC6. Add blocking diodes to coil ends for this test and connect first same ends together and then connect the two ends together using spark gap (upper circuit of figure 8 without C, C' and FWBR). Use C2 in the driver. Pulse it at resonant frequency using 50% duty cycle and short pulse. Measure voltage across the spark gap and current from either side of the spark gap. Take scope shots from both cases. Next add ferrite if you dare. Based on the result of TC4 you can decide if this is safe to do. Maybe lowering the voltage of the pulse is a good idea, or start from lower harmonic or just use a single pulse to get the ringdown. If ringdown has voltage over 100V then it means 10kV will be induced in the charge collecting capacitor if it is present.
TC7. Add series capacitor C and FWBR in the system used in TC6 but remove ferrite. Find the optimum value of C, procedure is described in more detail in 'Conversion to hot electricity'.
TC8. Remove C, add C' to FWBR and connect ground to negative terminal of the DC capacitor. Test both metal plate and earth ground. Put some resistive load and measure output power.
TC9. Build the oscillator system. One option is to use the circuit described in figure 9. Maybe a fast diode is needed from emitter to collector so that coil capacitor can be discharged. Energy collecting coil capacitor should be disabled during this testing. Connect the coil pairs together so that result is two coils connected in parallel. Then there will be no fields and system is safe to work with.
TC10. Connect a properly working oscillator to a coil capacitor system to replace the signal generator. Start pulsing from lower harmonics and measure the output power. More detailed testing procedure is explained in 'Solid state AEC reference design'.
[size=0pt]If tests up to TC8 can be completed and they give positive result then I am sure more experimenters will join and help to do the remaining two tests. Then we will get a simple reference system anyone can build at low cost.[/size]

[itsu]

Thanks for the new scope shots Itsu, very informative! Two different currents at coil ends (that are not connected to anything) confirm the presence of magnetic field potentials. The input current from FG seems to be a blurred sine wave with some spikes in it. The sine part is not coming from FG which is feeding square pulses only the spikes are created by the FG when capacitor is charged. This can be verified easily by reducing the duty cycle below 1%. To get rid of it connect the same ends if output coil pair together so that result is two parallel coils not connected to anything.
The blur is caused by high frequency oscillation which we already saw when blocking diodes were used. You should be able to tune to this frequency, make a sweep using 50% duty cycle.

Jack,

concerning the input current from FG being blurred, this probably is not caused by any hf oscillations, but just the raw signal from the current
probe.  In the next screenshots, i have set the scope to 4x averaging to reduce this blur.

The next 3 screenshots show:

again the input voltage (yellow), input current (green) and input power (red = yellow x green) at resonance (291Khz),
the input voltage (yellow), input current (green) and input power (red = yellow x green) NOT at resonance (191Khz)
the input voltage (yellow), input current (green) and input power (red = yellow x green) at resonance (291Khz) but at 0.9% duty cycle.

finally the 4th screenshot shows the same as 3 (resonance at 291Khz @ 0.9% dc) but now with the output coils paralleled.
This is causing more current spikes (green) and power spikes (red)

Itsu

[Jack Noskills]

Noticed interesting thing. When input pulse was 5 volt 10% duty cycle square wave the output ringdown was about 8 volts peak to peak, 4 volts from peak to zero. When input pulse was 10 volt 50% duty cycle the output ringdown was about 60 volts peak to peak, 30 volts from peak to zero. The one side of the wave was about 50 volts, about 25 volts from peak to zero. These two are not quite comparable due to duty cycle difference but anyway the increase in output ringdown exceeds 16 volts, 4*4. When pulse voltage is doubled current in the driver should be doubled also and q*q of the Coulomb's law becomes 2q*2q which means voltage between two charged bodies should be increased four times from 4 volts to 16 volts. But it seems that voltage increases 6 to 7 times instead of four. This has got something to do with capacitive pulsing, there are two separate coils so rate of change is more effective compared to plain solenoid. Another possible explanation is that FG is pushing more than double current when voltage is doubled. Just keep on eye on this one when changing voltage of the drive pulse. Using 20 volts pulse should produce easily over 100 volt ringdown. Maybe worth a quick test using air core system when doing TC6 for example.
Beware, that using capacitive pulsing with 1000 permeability ferrite core q*q could become 1000q*1000q which means million times greater voltage oscillations! This is the reason for those safety spark gaps so look out and begin using small piece of ferrite and low input currents.
I forgot to put FG driven MOSFET in the test case list. It can be done before starting to build oscillator circuits. Easier to test and it also shows if the used MOSFET is a good option for the oscillator. I will write this in form of a test case:
-----
TC8.1 Using the system of TC6, connect FG to a MOSFET (or transistor) to drive the primary coil capacitor. Ensure that the MOSFET is fast and that it can switch at least one ampere currents at the required speed. Verify that source can deliver at least 1A of current. Start below 100mA current pulses and 1.5V, 10% duty cycle at 1kHz, use the C2. Observe the ringdown. Next gradually increase the amperage in the primary keeping the voltage of the pulse fixed. C2 will charge faster so larger value might be needed for it. If the input current goes over one ampere then it can be said that ohmic resistance of the coil capacitor (which is 2-3 ohms in this case) has no effect to current. What is the maximum current (Imax) that 1.5V pulse can push through (keeping in mind the MOSFET limitation) ? Scope shot from this.
In case there is some sort of limitation with the current then use 5V or 10V pulse and repeat test.
Change the pulse from 10% to a short pulse that is less than half the cycle length of the resonant frequency. Use large enough C2 so that pulse goes off while current is still present and current reversal occurs. Use aircore and Imax current in the pulse. Observe the ringdown. Put a fast diode across MOSFET from drain to source. Does the ringdown increase ? Scope shot without diode and with diode if diode improved the result. Next insert ferrite core, set the input current below 0.05 mA and then gradually increase it. What current gives equal ringdown compared to aircore ? Scope shot. If absolutely safe then get the ringdown also when using Imax current pulse.
Remove ferrite core. Use C2 whose value was found above and short pulses at resonant frequency using Imax current to get the resonant rise. If the diode improved the ringdown then leave it in place. Measure the electric field potential across the spark gap and current on both sides between blocking diode and the spark gap. This is to test if the current difference is still there. Take scope shot. It will be interesting to see if voltage and current oscillations are now in phase similar to woopy's copper tape system that was also using 50% turn offset. Then measure the current between blocking diode and coil end and compare it to current measured between spark gap and blocking diode. Use the larger output current side if current difference existed. This test is to study if the blocking diode is attenuating the current. In my tests it had no effect but I could only watch for bulb brightness.
Add FWBR, C' and ground and measure the output power. Start using lower amperage short current pulses and gradually increase drive current to Imax. Additionally use ferrite core and start below 0.05mA current pulse. This power can be computed from earlier results using single pulse so this test is optional. Do this only if it is absolutely safe. No need to burn those expensive probes.

--------
This test case is kind of large but completing this will make completing TC9 and TC10 a walk in the park.
  There seems to some new scope shots, will download them and check them out.[/font]

[Jack Noskills]

Looking at the third scope shot of #218, with short 0.9% duty cycle pulse the oscillations are still present but amplitude is much lower. Obviously they do not come from the FG. Interesting that having the primary coil capacitor charged caused so big difference in the oscillation amplitudes. Over 40 time increase in current amplitude, from 5mA peak to peak to 200mA peak to peak. This amount of current cannot go through 2nf capacitor at 10V with 291 kHz frequency. So these oscillations are created by oscillating electric field. The lower amplitude oscillation was 2.39V and the higher amplitude was about 15 volts. Difference is 15/2.39 = 6.27 and this squared is 39.35. Pretty close to 40 times increase in current amplitude we see. I am sure this will change when blocking diodes are added so I don't think it is necessary to dig this further at this stage.

According to third scope shot the power used by the input pulse is about 50mW. Energy used is zero as everything is returned back to source which can be seen as negative power spike giving the second pulse. Coooool! Very nice to see this getting confirmed. I am sure that when the pulse it cut off while current is still present then the negative spike will be larger than positive spike and more charge will be returned back to source.

What could be done next is to take scope shot from the output coils similar to #214 using the short pulse (dual current measurement). Then the ringdowns created by using 50% and 0.9% duty cycle pulses can be compared. We could also see if this has any effect to phase difference between output current and voltage. There is a test case for this but it uses blocking diodes and 50% turn offset coil in addition. You could also decrease/increase the 0.9% duty cycle pulse length to see what is the minimum pulse length to get the maximum ringdown. Then putting FWBR+DC cap in the output could be tested to see what voltage is charged in the DC capacitor. I think that in case current and voltage of the ringdowns are out of phase then it must affect to this voltage. To get the U3*I3 connect ground to DC capacitor and put some resistive load across cap and measure power. Not sure if shorting the capacitor is a good idea when ground is connected. Maybe using low resistance load and then compute I*I*R is better option.

When doing experiments with C2 capacitor it would be good to see close up from the drive pulse, both current and voltage. Use as fine accuracy as possible, let's put that 1.5gHz sampling rate in to good use. When current lasts longer the current pulse should become more square like instead of spike. I forgot to put this detail in the test cases.[/font]