My energy has been restored, i feel vibrant again.
I felt like an old man the past year.
now I am alive and kicking again!
lets get this thing working!

today I'll flip L3 and L4

OK, So now L4 is flipped over, and L3 is not.

Now the discharge of the grounded cap, makes the resonant voltage of L3 jump up at V+ max, and down at V- max(not visible yet due to poor impulse).

So if I compare this to the other coil flip modes, Now L3 is flipped in phase, which is good.
But the current isn't in phase any more.

orange= L3 inside rim voltage
green is L3 inside rim current
yellow is L2.

I still need to remove some of the series diodes, and reduce wire length.

Also, from the PCB design GOD I learned, I need to have a send return path for the energy flows.
I now use single wires, to charge the caps. But these wires should be accompanied  by their return path (ground).
If done correctly this will prevent ringing.
I think this might help Bogirish also, with his ringing problems.
So bottom line: each line should be accompanied with its counter balance line, one line for voltage change, the other line for current change.
its not fully clear yet... but Altium design youtube channel explains. I posted it a few weeks back

scr388 ingreen shows the charge and discharge of the cap, which corresponds with the L3 voltage jumps

ok... Do I want the voltage of L3 to jump?
Or do I want voltage and current of L3 to be in phase?

I would say, the last... I think, I've logged enough data for now.

time for other stuff, and to enjoy my health!

Then later, rebuild my circuit, clean up the mess, provide return paths,
and look into ordering parts for the new isolated gate driver PCBs

I'm trying to get some rest, away from the project, but my brain keeps working...

So I have my positive primary plate coil L2
and my negative plate coil L3/L4

normally 2 plates act as one capacitor which can be both charged and both discharged.

But now, I only discharge 1 plate! the other stays constant.
only the negative L4 is grounded. L2 stays positive.

What happens to the field energy between the plates, if only one plate changes voltage?

Tesla said it creates a longitudinal wave.
and I feel it's a displacement current wave.
(if the discharge is quick enough).

This current should amplify the resonant fields around the plate coils.

So what I tried, with the voltage change on L3 isn't really essential. I thought it should increase. but not directly from the change in voltage, but from the induced displacement current.

So the L3 flipped, L4 non flipped, might show the wanted result. as the current and voltage of L3 inside rim are in phase.

while the resonant voltage appears to be deminished, it is only the dc offset. not the ac component.

then the cap to ground is an extension of the L4 plate coil
just like c2 is an extension of the L2 plate coil.
so the cap is uneven.
the L2 plate is much larger then the L4 plate.

Energy flows in the lowest impedance.
which is series resonance.
So getting L4 series resonant (with c3) seems essential.

the energy then is stored and build up in the L3 C3 parallel resonance which is high impedance?

that confuses me.

pumping energy in and out of the L2 L4 plate coil capacitor is key.

capacity is low, so voltage should be high.
and its about the displacement current. not the charge perse.

as charge is mostly produced by the L2 current inducing L3 voltage.

bla bla
I don't  know
back to rest

I didnt give this enough attention.
only L3 is used (L4 still on top but not connected, C3 also disconnected)

outside rim L3 has the cap (green) in series to ground, inside rim is probed (orange)
L2 is yellow

scr392 shows L3 has a ringing rrsonant freqeuncy of 788,644 kc/s 1268ns duration
so impulse, should match at 634ns duration

L3 sub harmonic can be
5th=157.73kc/s
6th=131.44kc/s (which this was indeed already tuned to during testing, which gave max resonant amplitude)
7th=112.66kc/s
8th=98.58kc/s

now, place C3 parallel over L3 again, and tune to the 6th sub harmonic

Note that the impulse and the cap discharge have different speeds, and therfore are not placed on the same wave phase.
tuning to the sub harmonic should solve this

psu: 2*13.3V @ 0.47A DC
@131.4kc/s

with C3 parallel over L3 tuned to 131/4kc/s with 7x1nf + 100pF (8nF)
it looks like this.
Tuning is very precise (high Q)

I still need to rebuild the setup, for the lower voltages
Still this low power setting 0.63A 2x13.3V already gives the slow impulse a near -300V offset on L3

The L3 coil, is actually the previous L4 coil, so it is flipped over. This means that when in is not flipped over, the negative voltage is much higher negative, when discharged. which is good.
This is interesting!

I wonder what the current does

OK I flipped L3 back, so its the same as L1 L2. I now completely removed L4 from the stack, which makes L3 a bit faster.
L4 is to be placed back if it works, and loaded,

L3 rang again
PSU is now 2x 12.3V @ 0.47A DC
F=134.6 kc/s

L3 rings at 1084.6kc/s 922ns
sub harmonics at:
10th 108.46
9th 120.51
8th 135.58 kc/s this is near where I tuned it for the max.

Now lets plug C3 back in and tune it

So, now this is with C3 parallel over L3 again. 7x1nf +100pF=7.1pF (not accurate measured)
tuned to 135.6 kc/s

Very good to see, the L3 negative voltage maximum, is indeed where the capacitor discharges and creates a current impulse.

Still no L4 extra coil.

current (green) of L3 inside rim (voltage is orange) is not a sine wave. large distortions at the negative DC discharge of the grounded cap. while only small, at the time of the impulse.
Fascinating :yoda:

I replaced the cap, with 9x68pF parallel 3kV ceramic disks, total around 612pF
and only 1 diode for the negative dc, so now restricted to - 1000V dc

also provided the source to diode line where the negative voltage impulse from L1 travels, with a ground line so it can travel better.

let's see how this behaves

looks a lot cleaner, less HF ringing.
and fast!
PSU: 0.49A 2x10.1V DC
F=134.8 kc/s
yellow= L2 orange= L3 inside rim green is L3 outside rim(cap to ground)
scr403 shows L3 inside rim current in green

allready -300V at 1/3 max power.

Ok, so Now I can use L4 again, in series with L3, to step the voltage way up.
making the negative much more negative,
Increasing the the voltage difference between L2 and L4.

and when maximum negative, I can discharge the cap that holds the dc charge.
so the negative voltage only discharges a relatively small amount.

but because L4 has such high voltage, which also will have the change in voltage, the electro static effect will be bigger.

Thanks Linus Tech Tips, for the inspiration.
just saw a video with him and elctroboom.
they don't even understand why that wire was moving...

https://youtu.be/1ugJ1BJx0HE

OK, I attached L4 in series with L3 again. no coils are flipped.
outside rim of L3 to outside rim of L4, as the resonance of inside rim L4 needs to be as high as possible.
Not yet tuned to a harmonic.
7.1 nF parallel to L3 (c3)
615pF from C3 outside rim to ground
L2 has 57nF in series.

F=123.8kc/s
PSU=2x13.1V @ 0.54A

yellow= L2 shows negative voltage impulse of -440V around 600ns in duration (still unstable)
green shows L3 inside rim with dc charge and discharge, 865.8Vpp
Orange is voltage of L4 inside rim, 1256Vpp also with dc discharge.

So now, L2 is positive charged to around 450V
while L4 is has a dc offset of around -400V but... reaches -750V when discharging.
This is still untuned, and can become a lot higher.
Total voltage difference between L2 L4 is now, 450+750= 1200V with a negative discharge of -400 (from 1200 to 800V)
And this is still untuned at low voltage settings.

Cool! now... again, measure ripple of L4, and see if I can get the impulse decently shaped.
I need to rebuild the Positive dc offset caps and diodes, as they are now build for 4kV while I only need 1kV

OK, yellow is L2  orange=L4 voltage (200V/div)
MAth in indigo light blue is, the voltage difference between L2 and L4. in DC (500V/div)

So a build up of negative voltage from zero, to maximum negative, a quick discharge from switch 2, then a slow descend back to zero. Voltage difference never becomes positive, Which is also interesting.

With better tuning, the DC voltage build up becomes higher, and the discharge also bigger.

Very cool! very happy with this!

today I had caffeine ( chocolate) which I normally dont use.
Normally it gives my nervous system stimulation as the drug is supposed to do.

today it did not. I was extremely tired and slept on the couch. this was after testing. I was very weird in my head and body.

I also had a allergic reaction from spring, probably a tree that had an orgasm.

this could explain the drowsiness.

if not, I feel its time to ground my shielding.
it was al in place.

feeling good again.
I want to see what the currents are doing in this new setting, and make a ground connection to the shields

Since I do not need series mosfets,
I made a new single driver, from the last parts I have. I hope mouser or digikey will soon stock parts again, so I can share the circuit and sell a few for those interested.

much more compact setup.
since I now am limited to 1000V dc
I have scaled down to single mur8100eg diodes with 1600V dc Wima mkp10 0.68uFcaps for dc offset of the L2 primary.

High voltage is obtained with the L4

I had some left over acrylic plates, 2mm thick.
I cut them into 11 coil size plates and stacked them.

cool it that the material already is super electrostatic

new setup with some explaining about the new setup
lol the thumbnail is from the old setup

https://www.youtube.com/watch?v=TvGipRRh5OQ

Did my first test with the new setup. It works.
around 2x27V dc input, I get a -1000V so that's now my limit.
voltages and signals look good.
temperature of the parallel mosfets rose to around 46C which is acceptable.
1000V discharge in 40ns, while charge was around 400ns
I recorded all, and need to do a quick edit
the negative dc blocking diode, also gets warm, the heatsink can be touched for a few seconds.
Might need a fan on it also.

added a cooing fan to the heatsink of the diode that passes the negative 1000V impulse to the cap.
hooked it up to one of the 2 parallel mosfets 12V

During testing I noticed my temperature said 30C on the mosfet. But when I turned the system off it said 43C!
Must be from the strong fields. So I kept it running for a while, then turned it off and it measured around 47C which is fine.
Will, need to let it run for a longer time, and look at it with my IR camera to make sure they dont run off and break.
very impressed with the current capabilities so far.

123kc/s 1000V with 615pF within a 40ns discharge :shocked:

replacing the 9x 68pF cap for a single 68pF cap while also replacing the mur8100eg diode for further testing with 1000V square wave

I only replaced the cap, not the diode.
so now 68pF instead of 615pF

It tested fine.
Still sag after the negative charge (green) so now I need to replace the diode, and see if its better.

Also, when not tuned, the green negative voltage stayed negative, but when tuned the negative voltage dropped, corresponding with the L4 resonant wave.
This is fascinating, as there appears to be some suction effect?

So the cap is to small to stay satble, lets get the 9x 68pF back in, and change the diode. see what that does.

I recorded it all