Instead of tweaking my PI filter, for reducing the PSU noise,
I'm going to use a 12V battery for my delayed pulse circuit and see how that works.

noise is not from the psu.
I hooked up a 12V battery to the delayed pulse circuit.
clean signal with system turned off.

loads of noise when turned on...

I need a faraday cage.

https://www.youtube.com/shorts/KlGQMKDE0ek

Time is a (fuel) field that induces transformation.
voltage and capacitance is related to time,
as a capacitor which has capacitance can store a dielectric field.
a static voltage in time.

 no transformation is present if we assume resistance is 0.
but when the voltage changes then a transformation starts.
the time starts flowing.
the rate of transformation which is time,
is activated and manifests itself as the displacement current field in space.

so in the dimensions of time and space,
time converts into a displacement field,
(the spatial field) by fluctuation of the voltage.

by fluctuation of the voltage the time field is activated so that's why we need high voltages which are present around the coils and can interact with the time field around the coils.

once this is established and we create a current impulse by a negative voltage rapidly (in no time) changed to a 0 neutral voltage.
(so a single Direction change of voltage)

than a single Direction displacement field current is created.

it's not magnetic yet because it's a displacement current but it's around the coils in the space around the coils and it is activated due to the change in the time field.

so the time field (fuel) around the coil activates the spatial field around the coil to create a displacement current which than is picked up by the coils and is transformed into a magnetic field of spacial conductor bound current

I made an extended series pi filter for the 12V,
and put the circuit inside a grounded faraday cage.

now the signal is clean. 5v out from the circuit is a clean pulse.
and at the gate of one of the B mosfets the signal looks good.

Still at low power,11V input. but already workable...

break time

scr291 in the previous post shows a lot of challenges.
the impulse opens up the body diodes,
the dc does not charge up slowly.
dc doesn't discharge properly
sine wave is erratic.

I think to solve these challenges, I need to charge the negative dc offset of L2 on the C2 side, and discharge c6 during the impulse.

since c6 is much larger in capacity the discharge currents are much higher.
So I might probably need more parallel mosfets.

I also feel C6 doesn't need to completely discharge to zero, as the lower voltages have less current en are slower in voltage change.

the beauty of discharging c6, is that on the other side of L2, c2 is already discharged by the L1 impulse.

so the shock wave that L2 creates is over its complete surface area.

then C2 quickly bounces back, while c6 is slowly recharged via L2
this creates an interesting energy flow.

I have 2 outputs on the circuit so I can easily double them to 4 Mosfets.

L2 is the primary coil and is bifilar.
it has caps on both ends (c2 c6) that provide a dc offset charge.

I can discharge them both,
if I connect my B mosfets, in the MIDDLE of L2.

probably causing a lot of HF ringing. due to the resonance of the separate two windings of L2 with their different size caps

discharge would be slowed down by the additional L2 resistance. probably an bad idea.
still fun to explore

I stacked the B mosfets. now I have more room, and I can shorten the high current wires and not have them tangled

attached is the new circuit. DC offset now enters via C2, and C6 is to be discharged by the mosfets

edit:
wont work

doesn't work.
no more positive impulse or negative dc offset.
probably due to the body diodes of the B MOSFETs SW2

the second half of resonant L1, (after the impulse which is the first half) passes through the body diodes of sw1 A mosfets.
It then is normally blocked by the mur440 diode (so it van flow into V-) and energy is stored into C4 (negative voltage).

But... Now the body diodes of SW2 B MOSFETS is opening up towards the negative dc offset, which is produced by the positive impulse.

So... the positive impulse is not able to charge the dc offset, because it is instanlty discharged again by the second half wave negative impulse...

BUMMERRRRRRRRRR

back to dreamland...

it was a high risk anyway.
Discharging a cap via a MOSFET is asking for trouble. much to high currents.

Still the idea is very good. to creat a DC offset on a coil, and discharge it.

I probably need  to use another type of switch, without a body diode.
Thyristor... in series.

the only solution is to put a diode D4 between source of SW1 A MOSFET and C6

there is no discharge through the diode so it should work

C4 can still be charged from the dc offset

no this won't work, as v positve from C6 is blocked ti v negative.

but there is another way.
whereby the diode isn't in the c6 discharge path.

let's see if this new circuit will work.
I changed the position of C6, moved it away from the source, and placed a diode in between.

Another diode and capacitor is placed at V-.
This is to be able to safely discharge C6 without a diode in its discharge path.
The C7 cap, is another buffer for during the discharge of C6.

The SW2 B mosfets are parallel to be able to handle the high discharge currents. Up to 4 in parallel are possible, and probably needed.
but I will first test with low voltages and 2 mosfets.

Sw2 is timed during the L1 impulse. This way both ends of L2 are simultaneously changing in voltage, creating a displacement wave from L3 into L2. It flows into, due to the polarity of the discharge. negative to positive voltage. Creates an INFLOW.
I'm making a short video to explain this.

lets build and test this.

I added a mur8100eg cap to the negative power supply in series with the mur460 which connects to source of the series mosfets.

I changed the 4uf to 1 uf on the source to positive cap c4.

the 10uf is now only 4uf 100V
a bit low, but enough for first tests.

the series mosfets can now become a single mosfet, as I wont dare to go very high 8n voltages. Since the mosfets are limited in their discharge capacity.  all is setup for testing.

orange is the c6 cap, so it starts to do what i want. quick discharge slow charge.
but yellow, L2 C2 doesn't show the impulse, so there also is no real dc offset.
green is the signal at the coax of the gate drive input of sw2. some noise there, but might be from the ungrounded probe.
so... still no impulse... let's work this out.

edit: yes green is clean, when I ground the probe

If I disconnect the coax of the sw b mosfets they stay off.
then the impulse is there, and dc offset charges up (as it is not discharged). so no problems with the body diodes.

Could it be I am discharging to quickly, before the dc offset can be charged full enough, to show an impulse?

as the impulse at the start would be clipped.
L1 only has a few hundred pico farad, while c5 c6 and c2 together are above 2uF.
so I dont get enough punch, to charge those large caps.

option is make the delay longer, set it to 25
And make the dc caps smaller, small enough for the impulse to charge them up.
C5 is only a buffer to change the polarity.
C6 is the ground cap for L2. if it is to small it will also start oscillating making L2 in a half wave resonant mode. 

first test the easiest. longer delay.

this test was with a bit more input, 13Vdc
once every 25 impulses, sw2 discharges.

but... still no decent impulse, and dc doesnt continue charging
hmm :-/

I dont get why sw2 when connected to the coax, destroys the impulse.
Sw1 doesnt have that problem, but the positive impulse is at the drain.
sw is connected by source to L2.
Which would open the sw2 body diode up for the +impulse.
BUT I never have seen an impulse at C6. The impulse is only visible at the C2 side of L2...
so.... why oh why is there no impulse when sw2 is connected to the coax.
because even when sw2 is off, there still is no impulse...

I still fail to understand why my impulses have disappeared. And only with the coax of sw2 connected.
so, I will take the circuit apart and check the parallel mosfets.
I stacked them, maybe something is wrong there.

But how can I have 30V dc offset?

the problem lies in the impulse. it flows away somewhere.
and I need to find where it goes wrong.

My dc offset caps are definitely to large.
if you look at the bottom of the yellow trace, there is a minimal change in dc offset towards more negative.
I ordered some caps, 68 and 150nF but might even need to go smaller.
will first try with the caps I have.
this was with 9,6V dc input, which gives around 200V dc max

since I cant solve my challenge yet, I will first focus on something else.
tuning the L1 impulse to the L4 extra coil.

I have made new tuning caps for l2 and l3/l4
10nf and 60nf.

the 10nf is capable of around 3500V AC
made of 4x 10nF series parallel.

Now I am not limited anymore by the 1800V AC of the switchable tuning board.

I made measurements, of the impulse duration, and the L4.

Surprisingly, the L4 frequency is perfectly visible, on the outside rim of L3.
L4 is grounded and L3 is series connected as before, inside rim to inside rim.

DC offset is put away for now

scr297 clearly shows the L4 ripple, measured on the L3 outside rim. The impulse is clearly faster.
scr298 shows the impulse halfwave duration of 380ns
scr299 shows the L4 full wave ripple duration of 1114ns, so a half wave is 557ns
scr300 show L3/L4 inside rim, max voltage at max power, of 2747V pp (64V dc input, 157kc/s)
scr301 shows current of L3 in green 15A pp, and 90 degrees out of phase with voltage (orange)
scr302 shows current of L4 in green 7A pp but IN phase with voltage.

now the impulse can be slowed down (adding parallel caps), or the L4 can be sped up (removing windings)

L4 is now 897.666 kc/s
while L1= 1316 kc/s
rather unique, as the difference is size and windings would suggest differently.
This again shows the uniqueness of the L4 extra coil. Which has voltage and current in phase, and is clearly much faster then expected. Again proof for the longitudinal modus it vibrates in.

its wise to first tune to a subharmonic of L4.
897.666
5th=179.53 kc/s
6 th=149.611 kc/s
7th=128.24 kc/s

I did these tests at 157kc/s far from Ideal...
going lower would mean more caps parallel to L3.
let's add 5nf and see if I can get near 149 kc/s

I doubled c3 to 20nF and this dropped the frequency to 115kc/s from 157,
which is near the 8th sub harmonic of 112 kc/s
I expect the system to speed up, when properly tuned, so 115 is fine for now.

I made a current and voltage measurement on the outside rim of L3 and compared it to the inside rim.
Unexpected is to see the current is much higher.
inside rim 9.78App and 1163Vpp (scr304)
outside rim 13.64App and 498Vpp (only ripple) (scr303)

I wonder, what makes the current so high on the outside rim of L3?

this was with 1/2 the PSU supply, at 2x15.1V=30.2V dc  @ 0,93A dc so I could pump the amps up much much higher!

I measured the impulse and L4 ripple again.
note that the impulse now wasa only 700V due to the ripples on L2, which eat away, so again not properly tuned.
impulse duration is now 392.8ns  just a bit slower (scr306)
L4 is 1118ns  894 kc/s  (scr305) just 4ns slower then before.

So no real change in speed.

I am tempted, to blast the currents with the full power of the 64V PSU.
but... I wont. I know these coils can run hot from the current.
instead I wonder If I should remove 2 series  caps from c3 to make it 15nF.
resulting in higher frequency and voltage.
Hmm...  I dont know yet.

It suprises me that the L4 is so slow. I had it matched before. maybe I should tape it down to L3 a bit bettrer,
so the mutual capacity is higher, and the frequency will go back up

I couldn't resist.
I pushed 1,89A @ 2x 32.2V DC from the PSU, at 115.1kc/s
and got (scr307) 24.255A peak to peak on the OUTside rim of L3 (green)
orange also on the outside rim L3, still showing 1030Vpp from the ripple of L4.
note how the current is some how leading the voltage! arrow of current probe was pointing in to the coil, away from c3 cap (as always)
What does that say about the transformation process?

Also included is scr308 measuring voltage and current on the inside rim of L3
current is lower, 18.914A pp so 5,3A less
While voltage is up to 2494Vpp

tuning to this lower sub octave is better for the current measurements. its more precise. but in the end the dc output power will be much more interesting.

Also note, I tuned for max current. not for max impulse. impulse now is 1700V but there are ripples eating away at it.