so this is the way the DC offsets work. Giving twice the impulse voltage difference between positive and negative.

the circuit caps are around 1uF.

The coils are then grounded through small capacitors, which are charged with the positive or negative dc offset voltages.
this then charges up the 2 coil capacitor plates.

What would be very cool is to make a spark gap or high current switch between those two coil plates, to make them series connected, making it a bifilar pancake tesla coil. the discharge would make it ring madly

but the main idea stays, to discharge the negative dc offset to zero, and see if I can make them ring

This is a first test with the L3 having a positve dc offset (orange)
and the L4 having a negative dc offset (green) which is also discharged.

It does show ringing on both coils, so that is interesting.

Now it needs to be refined.
positive ring should be 180 phase reversed, so the voltage difference between the coils is bigger.

Now the outsides of the coil plates are resonant, they should be made facing each other.

The voltage impulse on L2 (yellow) should also be tuned to match the ringing frequency,

What is interesting to see, is that the ringing is distorted after the negative dc discharge.
As if the frequency is suddenly higher.

Duty cycle of parallel mosfet switch2 is still only 5%... maybe I should make it 50%

which coil should I reverse to flip phase?

L2 series tuning cap C2 is still 62nF
C3 is removed... well... not fully, it is connected with one plate... maybe thats the cause of ringing strangely. I should remove C3 completely

I wanted to see If I could get the 2 plate coils (L3 L4) to ring, only by the discharge of the negative capacitor plate. And yes that works.

Still ringing down.

I completely removed the L2 primary.
I placed the L1 away, in a position so it would not induce L3 or L4.

L3 L4 are close coupled equal size coils. I connected it so that the facing windings are resonant.
This to increase the voltages on the plates. Which works.
One coil is flipped, so the phase is 180 degrees out of phase, so one rings positive max while the other is ringing negative max during the grounded capacitor discharge.
the positive cap is large 0.68uF
negative cap is small 615pF (and is discahrged through parallel mosfet sw2

PSU: 0.56A @ 2x9.8V =11W
F= 52.3kc/s

So L1 on;y produces the voltage impulse (yellow)
L3 orange has a positve dc offset,
and L4 green has a negative dc offset which is discharged

note how the orange is pushed up in positive during the negative L4 discharge. This should be prevented.
Which could happen if the impedance was increased, by adding a parallel tuning capacitor

now, can I reduce the power draw, by adding L2 close coupled to L1, and make the same impulse, from less power input by havinf the magnetic fields feeding back?
I think not, but I forgot...

although I seem to remember, when L2 is out of resonance, the impulse is almost free

lets check my video of april 2019 again, I showed it there
edit yes around 16:20

https://youtu.be/1Flj1i0zQ-8?t=1001

ok this is L1 L2 close coupled, L3 L4 disconnected
The DC caps are still charged, and the negative DC is still discharged (So there is some energy wasted).
yellow shows impulse between C2 and L2 ,
C2 = 62nF
L2 and C2 are series resonant around 72 kc/s
but I tune way above resonance, to get less current, only the impulse generation.
I tuned this to 120.0kc/s.
200v/div shows an impulse of around -500V
PSU is providing 2x15.7V @ 0.51A =16 W
So this is much more efficient then only using L1 without L2, to produce voltage impulse, which are needed to produce the current impulses.

damn so tired...

I got this idea in my head not sure how it will work out.

a bifilar coil has 2 winding, that are series connected.

what If I use a switch to connect them.
turn the switch off, then I have 2 windings that are capacitor plates. basically what I tested with. so give them a dc offset.
one positive one negative dc.

then... turn the switch on.
this changes everything. suddenly the 2 plates are a single bifilar coil.
Basically the terms for parametric resonance.
so then the discharge would create a high current, and make the bifilar ring.
until the switch is opened again.

assuming it is still ringing. the capacity and inductance again changes. making larger voltage and currents...

The reverse could maybe also be done.
close the switch, charge the bifilar coil with dc, then open the switch up. both will ring.
if the switch would be a mosfet then 2 voltage mpulses would be created simultaneously. of opposite polarity. if diodes and caps would be placed. maybe... need to work that out.

so... instead of flipping one winding, I could reverse connect one.
meaning, now the cap acts as ground both on the outside rims.
But I need to connect one dc offset cap to the inside rim of one winding.
that would also reverse the phase.
the open ends could be used for discharging through connection.
mosfet diode would naturally only block the dc. not the ac.

OK, same test, but now not with 2 bifilar coil.
I now only used 1 bifilar coil.
but, I opened up the series connection, so now I have 2 separate windings, which are close coupled.
I connected one dc offset cap to the outside rim,
the other dc offset cap to the inside rim.
loose coupled 2,5mm to L2

very high frequency 201.7kc/s
PSU: 2x24.1V @ 0.57A dc

yellow=L2, which is tuned way above resonance (Fr=72kc/s)
green is the positive DC offset winding, probed on the inside rim (cap on outside rim)
orange is the negative dc offset winding( facing L2) probed on the outside rim (cap on inside rim).

again, green shows the to be expected jump in positive voltage, when the negative dc is discharged.

Notice how the ringing is more intense after the negative discharge. the negative charge up barely seems to effect the ringing.
So this would still demand a parallel cap on the positve dc winding, to increase impedance.

But the Idea is now changed, Instead of discharging the DC negative, via the mosfets to ground.
I want to place the mosfet sw2 between the 2 windings. to series connect them.

regarding my health problems, related to the coils,
I looked up neuron impulses and found this graph from this site:
http://www2.tulane.edu/~howard/BrLg/Neuron.html
it suggests the duration of the impulse is around 3ms.
So a full wave would be 6ms.
which converts to a frequency of 1/0.006=167 c/s which is extremely much lower that the speeds I use.

I used 2 coils again, placing L3 on top of the 4 coil stack.
L3 is inside rim grounded through a positive DC offset cap, and has a 11,5nF parallel tuning capacitor
L4 is close coupled to L3, but not connected.
L4 has a negative dc offset that is discharged through sw2 mosfets

psu=2x 10,2V @ 0.46A =9.4W
f=99.7kc/s  (=Fr L3)
 so tuned above L2 resonance (which is around 72kc/s)
C2 is 62nF

Voltage impulse is around -390V (safe levels)
L3 is around 549.9V pp
I haven't measured L3 current but should be fairly high with 11,5nF
scr439:
yellow= L2
orange= L3
green= L4

Interesting to see L3 (green) follow the voltage of L3.

L3 is now inside rim grounded, which is done to get the opposite voltage polarity of L4, but ?I could also flip L3 over again.
L3 now is resonant on the outside rim. So better, to flip it over.
And... measure current.

additional:  the increased impedance from parallel cap c3 over L3 indeed keeps it from jumping up in voltage. So the L4 discharge is now singular

same test as before, but now yellow is the current probe, set at 0.2A/V
When the impulse charges the negative dc cap, a clear dip in current is visible.

When the negative dc cap is discharged, there should be a amplification of negative current, but it doesnt show up.

The orange is the voltage of the same coil, but shows a perfect sine. No trace of the current dip.

Going to reverse connect the L3 coil again, so ground it on the outside rim, so inside rim is resonant.
And flip it over, so the voltages are opposing again.

Note how the phase between L3 current and voltage (orange yellow) is not 90 degrees shifted...

reversed connected L3 and flipped it over. So now its 180 degrees out of phase with L4 to increase the voltage difference.
yellow= L3 current
orange= L3 voltage
green is L4 voltage

F=98.5kc/s tuned to the current maximum (almost the same frequency).

current (yellow) in now a perfect sine wave.
It appears to be amplified, but only slightly due to the low voltages.

the L4 green ring, doesn't look perfect any more.
The negative discharge is very erratic, this could be due to the current of L3, when the current opposes the impulse current it works as a diode, blocking the flow.

It might also happen, due to the not perfect tuning, whereby the negative voltage discharge opposes the ringing of L4.
As it does seem to stabilize after some time.

L3 already has a nice voltage and current for this low power setting
V=562Vpp and I=3.64App   
PSU: 2*9.6V @ 0.46A =8.8W

I just tested the impulse power requirements with and without L2. And I was wrong. With L2 it indeed needs more power. I guess I made an assumption which wasn't correct.

I tested at the same frequency of 105kc/s
L1 only made a -800V impulse of 418ns duration for only 1.3W
While with L2 and C2 (62nF) which are resonant at 75kc/s , but tuned to 105kc/s for the impulse,
It produced a -800V 388ns (slightly faster, so less power) which needed 5.2W to produce.
So with L2 it needs 4x more energy to produce voltage impulses,
Which makes sense.

So Why would I use L2?
I could make the C2 cap small and tune L2 to a sub harmonic, which will give me an impulse every 3,4 or 5 periods of L2.
If L2 is then matched to L3, they would ring at the same speed.
but because of the loose coupling, I dont see why this would help.
As L4 would be tuned to that sub harmonic...
hmm...

So... lets first continue without L2. and learn from the current impulses.

input power 1.04Ax2x8.1V=17W
no L2 used.

just pure dielectric induction. at 102.5kc/s
C=11.5nF parallel over measured coil

current is green
orange is voltage of the same coil
magenta is the product of measured voltage and current.

current probe has a slight deviation, due to phase shift. but its clear how much power is present in this resonant coil.

I put a 42W halogen lamp (load) directly over L3 (which has a 11.5nF parallel tuning cap.
It only slightly detuned it to F=106.2 kc/s (L3 max voltage in orange)

PSU: 0.75A 2x13.1V dc=19.7W input power.

green is the voltage of L4 which has the series 615pF cap, that is charged and discharged each half period
green in the L1 coil at the source

L2 is not used and removed, I still want to test including L2, and wonder If I can match it with L4.
This means L2 will be in TEM mode, as it will be driven by a sub harmonic. So, that makes it best to flip L4 over, instead of L3.

There are still distortions on L4 after the discharge, very high frequency, so I guess it is not properly tuned yet.

The lamp glows, I would guess it is weaker, then when a direct 19.7W was fed into it.

it again is impressive to see L3 (orange) produce a 333V pp sine wave under load. it does not look clipped like a square wave.
current must still be present (not measured).

The 42W 230V halogen lamp, is probably a mismatched load, impedance wise.

Still, very interesting to see it glow, and to have the sine wave.
I expect the current to be shifted into phase with the voltage again underload, as I have seen this before.

Hmm... strange.

I now run the same experiment as below, but now with series resonant L2 close coupled to L1 again.

Sw2 which discharges the negative voltage of the 615pF cap, has a temperature sensor on it.
It started at 19C before the test ran.

While testing, the temp dropped to 15C

after the test it rose back up to 18C again.

EHHH WHAT!?
It should have heated up from the high discharge currents... not cool down... :shocked:

ok, my negative DC capacitor, of 615pF basically is parallel over L1.
it has a diode in between, so it only changes the first half of the voltage impulse.

So it makes sense to also tune this capacitor, so the negative dc plate coil, resonates at the same frequency.
This way it will create higher resonant voltage, adding to the energy of the plate capacitor

If my AC output, has a positive DC offset,
I can only rectify the negative dc part of the wave.
else I would need to rectify into dc, without earth ground reference. the ground reference would then be the postive dc capacitor.
The caps would then need to be converted to AC, with a isolation transformer. very tricky

Looks like your doing some great research ivo! Keep up the good work.

My load impedance should match the impedance of the resonant power.
So, Let's design in from the load.
My electric radiator, has a dc wire resistance of 111 Ohms.
So... I need a resonant impedance of 111 ohms.

That will probably mean, I need a much larger capacitor over L3,
as the capacitor will deliver the amperage that will heat the wire in the oil radiator.

So I need to do some impedance calculations.

A larger capacitor will lower the resonant frequency, an will lower the resonant voltage.
which means the characteristic impedance will shift more to the amperage side. less voltage.
This is what I did with the primary L2 coil. So I might need to reverse that.
Get a high voltage primary L2, with a small capacitor,
And a high amp secondary L3, with a large capacitor.

It's the amps that heat. not the volts!

Also with a small cap on L2, it will drain less current from the supply, meaning, it will draw less power.
While it still will raise up the frequency high enough, to compensate for the L3.

to be clear, this is all based on my older research, and not related to the current impulse experiments I have been doing latetely.
So L2 is the series resonant primary, which now will have a small tuning capacitor, which also will get rid of the ripples again.
L3 is the parallel resonant secondary which will have a large tuning capacitor, with high currents and medium level voltages.

Whereby the impedance of the L3 C3 parallel resonance should match the load of 111 Ohms.
And I am talking about the capacitive reactance here. the capacitve impedance. as that is what power the load.
I should do some calacutlations here to get a better grip on this.

Electricity one-seven
by Harry Mileaf
https://archive.org/details/electricityonese00mile
login, and read an hour for free. can read another hour again and again.

Basic Electronics
by Bernard Grob
download link:
https://archive.org/download/Basic_Electronics_Bernard_Grob/Basic_Electronics_Bernard_Grob.pdf

So current and amperes are not the same.

Current is more related to the displacement current, from a current impulse. from a sudden change in high voltage.
which makes the dielectric field available for producing amperage. this is the longitudinal dielectric flow. which does not produce a magnetic flow.
current does not produce power, it does not give heat

amperage is the continuous transverse flow of the dielectric field, which produces the magnetic field, which can do work for us.

so the current impulse provides the field energy, to produce amperage.

So if I have a coil  capacitor, with one plate havimg a parallel capacitor.

then a current impulse from the primary,
provides the dielectric field energy for the secondary.
the secondary is then sets this dielectric field energy in  motion, producing amperes.
using the large capacitor, to slow the amps down, and produce high amps. for high power.

hmm not totally clear yet but it is starting to click

So a 240V ac 42W halogen bulb (assuming it is fully resistive) has a resistance of:

first, turn ac to DC, 240/1.4142=169.7V dc
P=I*I*R
P=U*I, I=42/169.7=0.2475A
R=P/I*I=42/0.06125=685.7 Ohms

To get a capacitive reactance 685.7 Ohms at 100kc/s, I need a capacitor of
Xc=1/(2pi f C)
685.7=1/2pi 100000 C
C=1/2pi 100000=1.59 uF

but wait... eh... its parallel resonant. so, the Xl inductive reactance doesn't counter balance it,
But... is it really parallel resonant? since one side is grounded, of both coil and cap.
So I should be able to trreat it like series reosnance...
Lets just try it.
I never had a cap that big on L3. should produce a lot of amps.
and low voltage. not near the 230V ac the lamp would need.
hmmm..

Lets just play with the concept, of small C2 large C3.

so, a small C2 in series with L2, and make it fully resonant, wherby the impulse of L1 is fully used to charge the c2 cap, and produce a perfect high voltage low current sine wave of series resonance L2 C2

Then L3 parallel with a large C3 lets aim for that 1.6uF
And see how that works out