I placed the 4x 330pf back in (instead of 4x 1nF)
So C1 again is 1,9nF
PSU=0.9A 2x19.1V=34.38V (still plenty of headroom to increase voltage)

C2 is now 2nF (1+10x0.1) and C3=7nF (6+10x0.1)
f=97.66 kc/s

green again is C1
yellow= L2 inside rim
orange=L3 inside rim.

look at that voltage of L3. as predicted it becomes much higher than expected with 7nF in parallel.
it is as if it is matching the L2 voltage.
L2=1656V pp including the charge and discharge of the DC offset.
L3= 1575V pp
so only 81V difference, While L2 has 2nF in parallel and L3 has 7nF in parallel...
pretty amazing how high the voltage in L3 becomes.
Can this be due to the displacement current?

Time for a current probe measurement.

also... ringing is present but not to bad.
a 1 ohm resistor directly connected to the mosfet source, before the c1 L1 junction, could help in bringing the ringing down.

charge duration of c1 is now around 700ns  (scr502)

scr503 shows the ringing of C1, around 10Mhz. 98ns duration for a period.
discharge speed of C1 is around 60ns still very fast, more then 10x faster than the charge time.

The ringing is from the c1 and the discharge leads, that have extreme displacement currents.
it only rings when C1 discharges, as this is a different path.
charge is from L1 to C1,
while discharge is from C1 to mosfet source, and drain to ground.
I tried keeping the leeds thick and short, to reduce resistance and inductance,
but I might need to increase resististance a bit.
But for now... its enough.
it works.

R1  1 Ohm, might fix the ringing...
 BUT Why do I have these expensive 21mOhm mosfets ???
on the other hand... if this works... I can buy cheaper mosfets

green is the current of L3 inside rim, 8.8A pp  from 7nF
ornage is the L3 inside rim voltage, and yellow is L2,
same power as before

I hoped I would have seen non linear current flow, instead of a perfect sine,
which would have indicated amplification from the displacement current.

Still... it is a lot of current

I wonder, what happens if I increase C3 further.
The freqeuncy should drop further, but also the L3 voltage, while current should increase.

scr506 is with 1 nF added to C3 making it 8nF
I had to change the amp scale to 2A per div horizontal.
psu 0.92A 2x19.2V
f=93.06
L3 12.5A

scr507.
PSU 0.97A (same voltage)
f=88.96
L3= 12.94A
C3=9nF (+2)

scr508
PSU 1.00A
F=85.15 kc/s
L3=13.1A
C3=10nF (+3)

scr509: detuned above fr. to 87.35kc/s
look at the wobbly L2. that is the sub harmonic of L2.

scr510 zoomed into L2, which measures around 480kc/s, 2us
So... I should tune L3 to an odd sub harmonic of L2, to keep it Max, from perfect sine, or else it wobbles.

So the c3=7nF setting looks perfect, because it is the 5th subharmonic of L2/c2.

so if around 100 is the 5th sub harmonic.
then around 70 should be the 7th subharmonic.
That would give a lot of current.
and that was the aim. high currents.
but the voltage would decrease, but will they?
as I have already seen it increase...
I wonder how big C3 needs to get to hit that 7th sub harmonic of L2.

I don't expect much at this stage. which is very good. look and learn.

I did another test to compare the currents in L2/C2 VS L3/C3
It amazes me, to see the large current in L3 while L2 barely has any current!
While both coils share the same space in time, as they are close coupled!
scr512 in green shows the L2 current at its inside rim 2.62App barely a sine wave. only current is from displacement current.
scr511 in green shows the L3 current at its inside rim 6.74App

this was with 0,9A @ 2x19.0V PSU @97.36kc/s
yellow= L2 orange= L3 inside rims voltages
C2=2nF C3=7nF

so... now tune it lower, by increasing C3

out of curiosity  I measured the current of the EARTH GROUNDED outside rim of L3 (where C3 is also grounded)
I measured voltage in orange, which was as expected a flatline zero volts.
In green, the current. That is NOT what I expected!!! it should be ZERO. but it is 6.92A pp

I checked if I could measure current, while probing mid air near the conductor , and I did not measure any current there, Like I did a year ago.

And this time, I do see the expected deformation of current, at the time of the C1 discharge.

VERY COOL :cool:

zero volts... 6,92 Amp sine wave.... WHAT?

I increased C3 to 16nF (10+5+1) and tuned it to
f= 69.57kc/s
PSU=1.25A 2x20,1Vdc =50.25W
L3=1200Vpp and 9.58A
C2=2nf C1=1,9nF
DC charge is around 700V
DC discharge is around 500V not bad
delta V L2 L3 at discharge is 1488Vpp.
so the coil cap discharges from 1488 to around 988V

L2 doesn't look as perfect as it could be, but the current do increase, while still having a fairly high voltage for L3.

So, I should measure the single geound cable and see if there is current flowing.

And I should rectify the negative period of L3, as there the current isn't pumped.

That could even influence the charge cycle positively.

And  I could still create a feedback loop into L1.

if I place two speaker wire pancake coils om top of eachother.
then I have 4 stacked windings.
of which 2x2 are naturally close coupled by being speaker wire.

so If I use the two inside windings as one bifilar (L3) coil,
and I would use the outside windings as one coil (L2) or, just 2 coils

then coupling would be perfect.

can you explain the green trace?

Quote from securesupplies on June 11th, 2023, 09:41 AM

can you explain the green trace?

it's the current measurement of the current probe. Inside rim of L3, is the resonant current.

but... the outside rim of L3 is the displacement current in and out of the earth. (also green)

scr519 yellow=L2 inside rim voltage and orange is the L3 inside rim voltage.

ok, I now measured the current ( in green) at the grounded outside rim of L3,
but also, on the earth ground cable connected to L3/C3
This confuses me a bit.
As there is no voltage on the outside rim of L3/C3 ground side.
But, indeed current is flowing there?!

scr520
now the earth ground cable itself, so, outside of the L3 and C3 connection,
still shows the displacement current, that is flowing into the system (negative=generation)
so that is the real energy that is being pumped into the system, by the negative voltage capacitor discharge of C1, which is extended to the L2 L3 coil capacitor (L2 plate)

I hard wired,
eh soldered the caps to the PCB of the parallel MOSFET switch.
3x 330pF wima fkp1 1600V in parallel per MOSFET.

This removes a lot of inductance, causing it less to ring.
I have an old PCB design, that had room for caps. Might want to revise it and update it.

Sadly the DC to DC converters have been postponed for delivery again.
It was May... now its august...

This electronic war between the US and China is messed up

charge and discharge times look good again, and with little ringing, so far.
This was just with L1, and no L2 L3 coil cap.

I also have a SiC diode that was gifted to me, which I will use in series with L1.

Charge is again around 700ns and discharge 40ns

I was gifted 2 1200V SiC mosfets I will use one for L1 to block the C1 discharge.
specs are vague. so I want to see how it performs.
It has 3 legs (two positive anodes) but I've decided to only use two legs.
since diodes dont share current when in parallel.

I also am going to tune L1c1  to L2 C2.
Which means that C2 will probably be smaller than 2nF.
the quarter wave charge is now around 700ns,
which means the L2C2 full wave would need to be around 4x 700=2.8us (full period)
that translates to 357kc/s
I might need to tune with the small 68pF ceramic disk capacitors.

L3 would again be at a sub harmonic, with a much bigger capacitor

I will rectify L3 with only one diode, into a DC buffer capacitor. It will give negative voltage relative to earth ground.
for a dc load I will use 2x 20W 230V in series to keep a higher negative voltage, in L3.

I then can see how input vs output power behaves, and how they related at different power levels.

I expect the output power to become more intense at higher voltage levels of the c1 discharge

I used a SiC diode, C2D10120 (wolfspeed) for L1.
just used one positive anode leg.
And it has no more overshoot.
Beautiful.
https://www.mouser.com/datasheet/2/90/C2D10120-31091.pdf

out of stock, so a replacement could be:
https://nl.mouser.com/ProductDetail/?qs=D3QG1iK7Fqh1Sgqzz9i2wQ%3D%3D
C4D20120A  (might be a proper over kill over design)
https://urldefense.proofpoint.com/v2/url?u=https-3A__cms.wolfspeed.com_app_uploads_2020_12_C4D20120A.pdf&d=DwIGAg&c=vY6N1du11amUSmqXT2U6Yqw6I-si6nWulSErsUHFzRg&r=Vsd0wO-QhHuOn_uef9GFTpTKbojymn6C_HPCMlXjhwY&m=bJHv0avYVX6LRL-XnsBN0VM1ePY1YEm6VwOeUL7FL74&s=hEejehA2KTxMfP4_qyYSAW0MKMEBQfpc9ZDe1YSoy4g&e=

This was with only L1, no L2 L3 coil cap.
PSU: 2x8.6V @ 0.54 A =9.3W
f=69.57kc/s

I also decided to switch back to a single channel of the power supply, as I do not need to go that high in voltage with the L1 having less inductance (unifilar)

I placed the new epoxy cast coil capacitor L2 L3 parallel to C1, with out the the C2 and C3 caps.
I tuned to the 5th sub harmonic of L2, While L3 was grounded on the outside rim.
green= source
yellow= L2 inside rim
orange= L3 inside rim (out of phase)

L2 without C2, has a nice high voltage.
Challenge now is to Tune L3 with C3, to the 5th sub harmonic, of 53.2kc/s at which this scope shot was taken.
This will make the L2 voltage very high, as it should produce a single sine wave, instead of 5, per period.

voltage in C1, now doesn't sag, I hope this will stay stable, when L3 impedance increases.

PSU=0.6A @ 17.6V

I now gave C3, 29nF and this tuned it to 53.2kc/s
But, as can be seen, the L2 starts ringing at a much higher frequency,
So I do think I need to give L2 a bit more impedance to ride this wave with a stable sine.
C1 shows a a 1/3 sag of voltage from -300 to -200V

PSU=0.61A @ 17.6V

but first... lets take a look at the current of L3 and the current on the earth ground,
as 544Vpp for a c3 of 29nF is actually already very high.

L3 amps in green, voltage in orange, with same settings.
4.17A pp @ 528Vpp   
while the power input from psu is only 10.7W

indigo= math, current* voltage.
only the negative cycle can be used, when C1 charges up.
that is still -300W,
negative watts, because it is generated.
so, 300 quarter wave, is 75W, RMS=53 W, 1:1 output impedance matched=26.5W output,
that is COP 2.6
IN THEORY and ideally

I also measured ground current (green), which now looks like a sine wave of 740mA pp including the spikes, of c1 charge and discharge.
Thats not what I wanted to see.

psu 20.2V  0.67A
f=55.0kc/s
yellow= L2 inside rim voltage (c2=2nF)
orange= L3 inside rim voltage (C3= 25.4nF)
Green is EARTH ground, at the wall socket, 200mA/div, 700mA pp

I tried loading the negative half of L3, with a diode and 1.4uF cap, with halogen lamp over it, but, then L3 flat lined. L2 also
so no more resonance.

weird to see the L2 voltage, is in phase with the ground currents.

I tried to get L2 into a perfect  sine wave (chopped by charge and discharge of c1).
but no go. keeps wobbly.
Maybe I should get C1 and C2 perfectly equal. So use 6x330pF for C2

I changed the C2 from the tuning board, to 6x330pF in parallel, same as I used for the C1. Wima fkp1 1600V
I gave it full power of one channel, 1.09A x 32.1V =35W
C3= 25.3nF (2x10+5x1+3x0.1)
which tuned it to 55.2 kc/s this was where L2/c2 had the third subharmonic, with C3 removed.
it is still wobbly, so I might need to fine tune C3, and the frequency to get a prefect sine wave.

L3 had 961.35 Vpp
and the delta V between L2 and L3 is 1191V at the time of discharge. Discharge is 380V, to 811V.
So now I could calculate the energy in that discharge using the coil capacity which I measured to be 580pF
charged=    0,41135949mJ
discharged=0,19073909mJ
total energy discharged=0.2206204mJ and, that happens in 60ns
and it happens 55200 times per second yeah, ok whatever

and here are the ground current at the wall socket, scr526
 and the ground at inside rim of L3 scr527

Now, what I wonder about is if the earth ground is flowing in and out.
then, what about the L4 connected to L3 inside rim?
Will that also push in and out the current?
And could I tune L4 again, with a parallel cap, while it is close coupled to L1, to see if I can produce feedback?

I guess I would first need to see what happens to the current in L4, without it being coupled.

L3 current is 9,53A pp and curiously again it is not a prefect sine wave, instead it looks wobbly like the L2 voltage.
weird...
is this because it is still a displacement current, that is a dielectric current, which is different from the magnetic current?

also... the earth ground current, is again in phase with the L2 voltage... What about that?

I should calibrate for DC and measure current again. see if negative is bigger. What is that positive bump doing there?

L2 voltage is in phase with the displacement current at the wall socket.
SO I feel this is not a magnetic current, since L3 also doesn't produce power, which it should.
In my heart I feel this is a pure displacement current, which is still measured as amps.

So, we have now 3 field components:
The magnetic current (where the power is)
The Voltage (the potential, as the dielectric field)
The displacement current (Dielectric energy flow).

I wonder if this dielectric energy flow, needs to be combined with the magnetic field, to give it its power.

Since L3 produces the ground current, I assume, it will also produce the same dielectric displacement current into L4.
L4 can than be coupled to L1 which has a magnetic field.
Then the displacement current is mixed with the magnetic current, to create real power