I checked the voltage sag of the 615pF and it is still there, just a bit less

only L4 connected, L3 still close coupled but open ended.
yellow=L2
orange is L4 inside rim
green is L4 outside rim/grounded capacitor

again, green shows the ripple of L4

I am wondering of L3 should present a load, so the low impedance inflow of energy only flows between L2 and L4 (in the gap).

I haven't even loaded this...
just playing around now

I will give L4 a parallel cap. use L3 for output

This is L3 series with L4
L4 outside rim to cap.
L4 inside rim to outside rim L3 and parallel cap (green)
L3 inside rim to parallel cap (orange)
L2 is yellow

now L3 inside rim has high voltage (so coil cap L2 L4 is less charged)
L3 is riding on top of L4 so it only appears to be high voltage while in reality it is compounded

power 2x 15v

This is L3L4 inside rims tied together to the parallel cap. (green)
outside rim L3 is grounded by cap.
outside rim L4 is free (orange)
L2 yellow
low power setting 2x 10.6V

outside rim L4 has low voltage
inside rims do look nice.

but I dont agree with an open ended L4. makes no sense.
so, ground outside rim  L4 through cap, connect inside rim L4 to inside rim L3. this needs L3 to flip over.
L3 havinf the parallel cap.
I dont like flipping coils over...
So instead connect inside rim L4 to outsid rim L3, and load L3
hmm yes I just did that...

Then... L4 needs to be open ended.

or... stop playing, and put a load on it

L4 can be open ended, if L4 serves as a current provider for L3.
So L3 outside rim grounded, inside rim of L3 to L4 outside rim,
L3 has parallel tuning cap, and L4 inside rim is open ended, (and has high voltage, charging the L2 L4 coil cap)

So as load... since I use dc offset, I cant use ground.
instead I can rectify both ends of L3/C3

no... makes no sense. I cannot rectify the outside rim of L3 as there is nothing to rectify. It is not resonant, but connected to the grounded cap. still it can provide as a floating ground. as long as the dc power cap isnt grounded, The cap needs a polarity so ... yeah, I'll use it as ground, no need for rectifying outside rim of C3

ok, this is with L3 loaded, with a 42W halogen bulb, with first rectified L3 inside rim onto dc caps.

orange is L4 inside rim
green is c3L3 inside rim L4 outside rim (power source)

lamp doesnt glow. no resonance. no power. no high voltage L4.
hmmm....
ok

another though I had before, is not connect L3 to L4 at all. just capacitive and inductive coupled to L4.
this way I can ground L3, and use ground for my load...
but... L4 has  a high ring. I could use a parallel cap again. so, L4 becomes L3 again.
Although a cap can tune the frequency down, and turn the voltage down, it can still be high when the cap is small enough.

yes ok, Load L3 C3. and use L3 to couple to L2. ground C3 outside rim via cap.
 L4 close coupled to L3 for output .
can be directly loaded, or could even be made resonant. But L4 is not connected to L3
tired

forgot to look at the diode. voltage drop seems to be the same. must be taking longer to close after these high currents, thats what the data sheet says.
MUR15120L also looks interesting.

only L4 connected, L3 still close coupled but open ended.
yellow=L2
orange is L4 inside rim
green is L4 outside rim/grounded capacitor

 green shows the ripple of L4
I will put a parallel capacitor over L4, and tune it down to a odd sub harmonic.

This was at 116.2 kc/s giving, 7 rings.  So Fr L4 is around 813kc/s
3rd sub=271.13
5th sub= 162.68 kc/s
7th sub= 116.2 (duh)
9th=90.38

If the parallel cap is kept small, the voltage of L3 will be much higher negative. So I'll aim to the fifth
sub harmonic around 163 kc/s

Why the odd harmonics? I learned from the half bridge circuit, that when I use double impulse polarities I need odd to lign up with the right polarity of the wave. one going up the other going down.
Maybe I'm wrong...
6th harmonic=135.57kc/s
8th sub= 101.68

a more precise measurement of L4 ringing (orange), with L3 unconnected close coupled.
at 114.6 kc/s  gave the best result, another at 147.6kc/s.
Measured 826.446 kc/s for L4 a bit higher.
so odd  (even dont work) sub frequencies are:
5th=165.29 kc/s
7th=118.06kc/s

These aren't presice but a good aiming point,

Tuned with C3=5nF parallel to L4 (4x1nF+10x100pF)

I should again, call this L3. will make a new drawing if this turns out fruitfull

I used math to show the voltage difference between L2 and L3 (primary and secondary).

This was at 168.7 kc/s

input power was low, 2x10.0V at 0.63A = 12.6W

I did not yet look at the output coil L4, (open ended and close coupled)

so now the stack is again:
L1 L2 - - - L3 L4
L3 having a parallel tuning cap c3=5nf
and a grounding cap of 615pF which is dc charged and discharged.

L4 is close coupled to L3. but not connected.

L4 can be grounded, if so. There forms another coil capacitor. between L3 and L4.

and since L3 is dc charged and discharged, with L4 grounded, it is a negative charge on and off. between the coils.

L4 could ring, but... is to be loaded.
so what would that do?

How much would L4 load down L3?

one way to find out.
load it

OK, I grounded the L4 coil (on top of the stack close coupled to L3) and measured the voltage (green)
It shows a copy of the L3 sine wave, except for the dc charge and discharge component.
L4 is unloaded.
power from psu: 2x 12,8V at 0.77A at 168.3 kc/s

Now I need to look at the currents

edit.
So the grounded coil plate doesn't show the current impulses.

I should do this again, and measure the ground connection. because this really should work as a capacitor. if there are ground current, then its interesting to say the least.
A diode between coil and ground could then serve to trap the negative voltage in the coil plate.
It could feed the resonance of the coil?
gaining energy from ground.

forget about those frequency calculations.
147.6kc/s worked, so tune to that one.
and use the current probe to see what happens

Measured current of L3 (green)
psu: 14.6Vx2  @ 0.75A
L4 voltage green is out of phase with current, so standard resonant behaviour
F=147.3kc/s
with C3=6.6nF (6x1nf+6x100pF)

I'm feeling a bit lost again...
Wondering is this the way to go?
Shouldn't I discharge the cap into the coil?

Still... I dont expect ffcts with these low settings. I really need to crank up the power, to get some proper current impulses.
until now I play with 300V which is nothing.
L2 and L3 have a 16mm acrylic plate between it. It should have thousands of volts before anything starts happening...

So... then why not use the capacity of L3 and L4, for the current impulse, use that as a coil capacitor, as it has close coupling.
much larger ....ahhhhrrrgg...
back to something else...

so the current impulses are pumping energy in and out at twice the frequency.
which is the condition for parametric resonance.
but I am not really changing the capacity or inductance, only the voltage.
so... should I short a coil out (together with a series cap)?

scr421 shows curren in green of L3 inside rim, L4 not present
L3 is parallel tuned with 11.6nF to 115.9kc.s
outside rim of C3 has cap to ground.

current probe (green) suggests there are two negative displacement currents
which also explains why I see this working at odd and even harmonics.

But the voltage difference between L2 and L3 (magenta math of previous test scr413 https://open-source-energy.org/?action=dlattach;topic=3660.0;attach=31076;image ) shows a positive voltage jump, but that is on the L2 coil plate

cant seem to get a cumulative wave.
lets try discharging the cap through the L3 coil.
this means I should connect the cap to the inside rim of L3? but then... cap is grounded, and outside rim of L3 is to mosfet switch, which has a diode, to ground?

a cumulative wave, or trumpet wave, could be based on negative feedback, were the secondary is fed back to the primary?

or using negative resistance :
https://www.circuitstoday.com/negative-resistance-oscillators
half way the page is this attache picture.
the circuit on the right interest me, as the capacitors are in series parallel with the coil.

a reaction from Romero UK:
To obtain that shape I had to construct a special coil. I do not remember all details now but it was created by winding CW and CCW with capacitors connected at every point where i changed direction. I have also managed to do it using a split core towards one end.
from this video:

https://www.youtube.com/watch?v=KpShM3yjoSI
following that comment, was another reply from elisagroup:
this trumpet waveform is simply showing a stray electric field (scope probe being capacitively coupled) of the charging capacitor from a HV HF source. The spike at the end shows a discharge of the capacitor. It's easy to see this waveform when you charge a cap from solid state NST, for example, and leave the scope probe nearby

So I call L2 the Primary and L3 the secondary.
But with these current impulses on L3,
doesn't in become a primary for L2 which then becomes the secondary?

So both coils are primary AND secondary.
one feeding into the other,
while the other feeds back into the one.

L2 feeding into L3
while L3 feeds back into L2.

that would be a negative feedback.
which can be contructive.

but there are always losses from resistance.
so... hmm

its like the electric motor driving the generator driving the electric motor.

but now with dielectric feedback

measured ground current on the ground side of the L4 diode
I close coupled L4 to L3,
L3 has a parallel tuneing cap, and a series tuning cap to ground, which is charged by the impulse, and discharged by the sw2 parallel mosfets

So L3 and L4 are capacitor plates.
L3 is charged and discharged (DC)

L4 has a diode to ground.
since L4 will follow the voltage of L3 it will need to draw voltage/current out of the earth,
which passes through the diode.

current indeed is measured.
probe was at the more sensitive 200mA per volt setting, scope at 200mA per division.

scr426 shows the current of the ground side of the diode in yellow
orange is L3 voltage DC probed
green is L4 voltage dc probed, notice the negative dc offset, due to the diode.

I detached the current probe, but kept it in the same position.
This showed the same pattern (yellow)
scr427 is probe disconnected
scr426 is connected (previous post).

So not a valid conclusive test, although there does appear to be a stronger negative current at the discharged from sw2.
And this is what I expected to see.

what If I only used L1 and L2 to create voltage impulses?

tuned outside of resonance I can create strong impulses with barely any power?
like I showed in april 2019

use these to charge a small cap, to create the current impulses.

and use the current impulses as the true primary coil.

only to make l4 resonant and tap into that

and... maybe...  use the dc offsets again, but not on L2, L3
but on  L3 L4
one positive one negative dc,
and combined with the current impulses discharging and charging the field

I should measure again, with a reverse diode to ground, making L4 positive, from the quick discharge. measure current agian.

current probe isnt fast enough....

I want to test if this works for creating positive and negative dc offset from L1 voltage impulse.

if so,
these could be used to give L3 and L4 positive and negative dc offsets.
since they are tight coupled, capacity is much higher.
L3 L4 are then only coupled but not connected.
D2 should be able to withstand the voltages,
a 1000V impulse would need a 2kV diode (2 in series)

the negative c3 then is discharged through a sw2 parallel cap, to see the effects.

edit:
I did the test, rmoved dc offset from L2, still probed with yellow
DC offsets indeed are created, but... with only half of the expected voltages positive and negative.
which is to be expected , as the diode now is not referrecnce to ground, but to the negative dc offset.
So it works, bu the other method gives twice the voltage of the impulse, while this only gives the impulse voltage.

green is the negative offset voltage,
orange is the positive dc offset, its shows the impulse, as it is not connected to a grounded cap.