tested again.
setup only works when the bifilar coils are NOT bucking. this means series connected via outside rime to inside rim of the other coil if two coils are used.

thats a good thing. this indicates the coils work with the pulses of the bucking coils.
thats pretty funny since they are in phase...

another thing is, the resonant coil needa to be grounded, or else only the pulses show up.
another funny thing... the resonant coil needs to be grounded on the outside rim to get the biggest voltage rise.

next thing is connecting the center coil to the bucking coils.

center coil of the 3 stacked bifilar cd coil is now pulsed by the bucking toroid (with center tap of toroid grounded.

top and bottom coil series connected as before outside rim to inside rim.

probed on the inside rim grounded on the outside rim.

resonant voltage rise at 91khz (lower) with 7nF in parallel with series coils.

voltage rise is 225V

A lower res frequency with a higher rise.

the pulses at the bucking coils look sharp. around 400V dc.

now lets add some more voltage and current. because the IGBT stays relatively cool.

I also think there might be a better design to the bucking toroid, as one side shows a little bit of a ringing and a lower pulse. this could mean only the voltage differnce of the 2 pulses accounts for the voltage rise...
and when the pulses are matched... all is gone.

hmm what if i remove the ground from the center of the bucking coil and move it to one end?

nope center needs to be grounded.

here are 2 shots of the pulses one on each side of the bucking toroid coil. the one with the sine is the coil without the primary.
If i press my finger on it (add capacitance) the pulse drops slightly in voltage (more ringing)

I'm tempted to wind another meter of primary on the other winding of the bucking coil. since its purely the b emf i use...
I could wind the whole toroid.

is this what nelson is doing?
one single coil pulsed.
back emf fed into a bucking coil?
bucking coil into ...

inductance of the primary is 0.15mH still the IGBT stays cool (91khz). it should get hot I think. the bucking coil does a good job in extracting the back emf out of the toroid into the bifilar and ground

duty cycle used is still 84 %

The bucking coil should be made better.
I now have only one side covered by the primary coil. the uncovered side gives a lower resonant frequency resulting in a lower voltage pulse with energy left in the resonant sine.

I have 2 options,
A is to overlap the primary on both sides of the bucking coil.
B is, using 4 equal parts of the coil, for the primary and the bucking coils, so the primary and bucking coils do not overlap.

A might result in equal pulses, with a higher resonant frequency.
B is based on where the bloch wall is of the primary (in/under its windings) and the bloch wall of the bucking coils (outside of the windings)
of the magnetic field collapses into a dielectric back emf, the place of the bloch wall might be relavant to where the energy moves inside the toroid.

Again... my mind cant grasp this concept, because I dont really know how this all works. So...
lets experiment again... so far what I know is a covered bucking coil gives a higher voltage pulse. so version A. That is also easy because I only have to add 1 winding to the one I already have.

 

added 105cm(total 2/3) to the primary, wound over the other part of the bucking coil.

the inductance of the primary now is 0.44mH

no succes.
without load the pulse is time stretched becoming a triangle with a lower voltage.
the shot is at 91khz giving a 170V peak

plan B
wound the primary more on one place to leave the bucking coils open.

excact same results and same inductance.

maybe the extra added length was a bad idea

or the symmetry should be avoided

I shorted out one of the coils
now the spikes look better. equal on both sides 225V

I removed one half of the primary. So back to 1/3 lenght.
the coil is between the bucking coils endings.

pulses look good and equal, but with a slight ringing.  380V pulses at 93khz

now the resonant rise is very low.
conclusion is that the equal spikes now dont do very much, or perhaps the primary should be placed asymmetric

I think the voltage difference of the spike might be responceable for the resonance...

So... did Nelson do it different?
I think so. I still think he used the back emf of a single coil. maybe in combination with the back mmf of a capacitor

when I connect a bifilar coil to one of the outputs of the bucking coil, i also add capacitance and this creates a ringing resulting in an even lower frequency where the pulses stay high in voltage.

I made the bucking coil into a normal coil by reversimg the connections on one of the 2 coils.
then I added a fast uf4007 diode in series and a 47uF 450v capacitor in parallel to be charged by the b emf.

i pulsed at 75khz and the voltage in a few seconds was several hundred volts.

If i would use a smaller capacitance cap it would load much faster.
then I could discharge the capacitor into a bifilar coil. how fast depends on the capacitor size and the b emf.

because I have only 1 igbt I could use a joulethief to create the back emf into the capacitor. then I could use the igbt to discharge the cap into the bifilar coil.