Like this.

choke only passes voltage. its reactance limits voltage.

dc positive is a large capacitor.

HC coil is parallel resonant.

so 50% of the time the center of the high current coil is grounded to neutral.
 switched at zero reactance point of the high current coil and high voltage coil

looks good.
No current sine wave at all. (without HV coil powered) green
in orange the gate signal, 49% duty
And in blue, the voltage between the choke and the HC coil, at the drain of the mosfet.
Very abrupt voltage changes.
somehow the voltage is higher than expected. my need to check the probe again, to see if it shows the correct voltage.
PSU is 0.3A at 3,9V so the choke really needs an upgrade.
Also... I reverse connected the choke. I might need to correct that back, or else it might become resonant with the coil capacitor.

This is basically the same as what I did in my last video of summer 2023.
But instead of charging and discharging a very small capacitor using expensive parallel mosfets,
I now use a single cheap mosfet. while using a choke to limit the magnetic current, and a very large capacitor to stay DC charged.

I added the power to the HV coil (yellow)
some current is now indeed (green) present in the HC coil.
While blue has still a relative low voltage.
orange shows the gate switching, pumping the voltage of the coil capacitor.

It appears as if the choke coil has to much current, which turns into a positive back emf spike. seen as the blue jump in voltage.
choke is still not reversed connected on the second winding, but it already looks good.

I need to fix the sharp points, as the smell of ozone is dominant. cant have that.

good to see at least some current present. still not much, but it is there.

phase is not perfectly tuned yet. But now I have 2 mosfets to choose from, which are out of phase, so I can easily tune the square wave

PSU:
current coil: 7.9V 0.1A
voltage coil: 6.5V 2.38A

I will refine this setup.
more inductance on the choke.
fix the HV leaks (ozone)
more voltage to the current coil.
fix the DC caps of the current coil (I have 2x25V rated in series that hum, better put them in parallel or replace them)

I will now remove the choke and mosfet from the center series connection of the 2 windings of the bifilar HC coil.

and will connect the choke and mosfet to one of the capacitor plates, so on the coil ending.

I will probably have to retune the parallel capacitor again

my choke was 2x 2,9mH
Now I found a choke with 2x 12,8mH which I can increase in inductance even further.

4.4 times more inductance gives 19.5 times more reactance. lets see how that behaves.
I also replaced the buzzing 1000uF cap of the dc supply for the current coil

I also see I connected the choke coil wrong.
The center tap should connect to the positive voltage supply.
then one winding to the hc coil center tap,
and the other windings to the negative passing diode to ground

I added more windings to the choke. and this time, I series connected the 2 windings into one big 103mH choke (no diode to neutral). the current is somewhat less, but still very easy to drive to 2.5A not good.

Tuned again. placed the ground on the load side where the High voltage coil ends.
seems to make no real difference.

When I increase the power to the high voltage coil, I do see the psu current of the HC coil drop. which is good. as if the displacement current is assisting the magetic current (as they are in phase now).
Increasing the power to the HC coil does not degrease the psu power of the HV coil.

green current is now probed on the other side (where the load is not grounded)
orange is the gate of the other mosfet, so should be looked at inverse (low=on, then HC coil at cap is grounded to neutral).

PSU:
1.85A 5.0V for HC coil
2.37A 3.2V for HV coil

So the big problem still is, the high total duty cycle, drains way to much current, it is 1,85A at 5.0 V
While I would like to see 0,1A at 15V for the HC coil.

Also. the current coil is not properly tuned. the high voltage coil has dropped to 14.66 kc/s so I should add more capacity to the parallel capacitor of the HC coil. Then the switching will also align with the current maximum again.

So. should I increase the choke inductance to even higher ridiculous amounts?

Or... should I connect the choke to one end of the cap, and retune the capacity. I think that last is my best bet.
because then the 50% duty cycle makes more sense.

I also still wonder if the load ground has a logic, on one side or the other.
also I still need to fix the ozone problems.

If I push enough current into the HC coil then I can light the lamp. But When I increase the HC then nothing really happens. the light even dims a little bit.

Priority is reducing the current through the choke.

2.46A 3.2V for HV coil
0.5A 15.5V for HC coil.

I connected as said in previous post. this does work better.
but still the choke (103mH) needs to be increased in size.
I fail to understand why. but the HC coil does produce current, from the HV coil induction. so my guess is that this current is fed back into the PSU.

the choke now has 9487ohms inductive reactance at 14,66kc/s so the 15,5 volts should never be able to get this high in current (0.5A) .
A blocking diode comes to mind, to prevent the feedback.

When the HC current coil PSU is turned off, then the HV coil power is inducing already some impressive current in the coil.
it also appears to be better tuned now. It switches at the maximum.

So a series diode to prevent reverse current flow back into the PSU is needed.

orange lead is now probing the right gate so high is mosfet on, coil switched to neutral, orange low=15 volt

1.176 A from only the HV psu.
So the amps are increased to much higher levels. I am starting to see what I want.
Lets decrease the current of the choke to lower levels, by inserting a diode

I can now increase the dead time also, since I only use one mosfet

diode is in place.

current appears to be decreased, but is rising (thin wire of the choke?) this was 0.43A and 15.5v on the HC psu.
the current looks weird when the HV psu is off.
is this due to the HC coil not being properly tuned?

when only the HV psu is on, there is no voltage or current reading on the HC psu, with or without the diode in series with the choke.

when the mosfet is turned off (orange = low) there is a weird jump.


break time

If I now have the HC psu on, it has 0.25A and 15.5 volt.
the current is fairly low 3A peak to peak.
the lamp is off.

If I now increase power to the HV coil, the HC current is amplified, but the HC PSU doesnt rise!!!
That is what I wanted to see.

ALSO
the lamp starts glowing when the high voltage is present, this is due to the amplified magnetic current of the HC coil.

this is good. the high voltage coil displacement current is now amplifying the magnetic current, as they are in phase.

Now. WHERE is that current of the HC PSU flowing though?
something is leaking

I thought I saw smoke at the gate driver/mosfet. but felt no heat.
But the current does seem to rise indicating that something is running hot.
IR camera should point out where that current is flowing hot

maybe my scrubber circuits are the cause of the currents.

I took a look at my 2 snubber circuits
which are identical.

one has a strong color change on the resistance indicating heat.

It does not explain the extra current, as the diode checks out being perfectly fine.

the positive spike at the drain passes through a mur460 diode into a 100nF cap which has 2x 100 ohm in parallel. so this 50 ohms are driven hot.

I took out the snubber and
placed the diode of the choke to the positive supply.

Now things really have changed.

current is absolutely zero on the HC PSU with 16.6V
so the choke now works.

no parallel resonance with only the  HC coil powered.

If both the HC and HV coils are separately powered, I do see current, but nothing as expected.

the yellow voltage shows a dip without a cause.
at the same time the current shows a negative dip.

Right now the ac coil is earth grounded on one side, I should ground it on neutral.

the anomaly happens when the HV push pull switches.

I removed the earth ground from AC transformer, and grounded on neutral.
now large spikes are rising, so snubber is indeed needed.
voltage can be observed rising from zero. no negative side, which must have been passed through the body diode.

I gave a low voltage at the hv psu, 0.3V
yellow shows the change in voltage of the hv coil
the hc coil is given 16.6V dc

orange is measured at the drain. the positive voltage is around 75V and is discharged by the mosfet.
it does ring.

the whole system now is floating, except for the output coil.


I am glad to see no current on the HC coil psu. but now other things arrise. And WHY was the snubber a problem?

I connected the other mosfet to the circuit, that still has the snubber circuit.
and yes, now current does flow, and the 50 ohm snubber resistors do get hot.

so. case closed, its the snubber.
lets redesign it properly as this version clearly is bad

I made a more traditional snubber. a 10k ohm resistor in series with a 2.2nF capacitor,
calculated from the 25mH choke and 20uH choke plus coil, which should allow around 7.2mA from 16.6V
The Q of the total inductance is around 0.653 uJ
so this q is the same for the cap when the cap is 230 pF  but that is only the coil.
in reallity there is alse a parallel capacitor that provides even more energy when resonant.
 so I used 2.2nF its still a guess I guess :D

The drain of the mosfet rings like a bell, and gives a large spike when the voltage coil is provided with power.
so.... I guess I need an even bigger capacitor, which also must be fast enough.

green shows the current, induced by the high voltage (yellow)
orange is the drain, with a large spike,
I provided 16,7V to the hc coil (0,007A)
and 3.8V to the HV coil (2.95A)

note the very slow decline of the orange voltage when the mosfet is turned on.
This voltage should have been instant.... ASAARGGAASVAVD SGADFV BZ SFBDLM:: JOLKSAEBG JO:IALGE PJOK:ARGEHAEG B:RJKLO

so the snubber that is now parallel to the mosfet, prevents the fast discharge of the dc voltage offset?
While I still have a spike...
this is bad
will this turd never end?

this makes me so mad.
the solutions of general physics always SUCK

I just want a fast change in voltage. how hard can it be?

so, back to designing the snubber.
best is to ground the snubber on the positive supply, so that the mosfet will not be affected.

because when the snubber is in parallel, then the the parallel cap is discharged through the mosfet, but the resistor limits the discharge....

ahhhhhrgrgrgr

Why must I do all this alone?

I have come to the conclusion that a snubber at the mosfet makes no sense.
As the HC coil already has a parallel capacitor that basically acts as a snubber like spike eater.

I actually have 2 coils.
and the choke seems to provide the biggest problems. So lets place the snubber there.
lets choke the choke spike.
that also is a good reason why the mosfet is on the other side of the coil.

Let that coil be resonant at the mosfet side.
and snubber that choke on the other side.

(2psu)
this is without snubber, without a second choke coil. without a diode.
just one diode between the positive psu and circuit of the hc coil

voltage on the hc psu is 17.0V with 0.00A (must be around 7mA)

orange is the voltage at the drain which is the same at the choke side of the hc coil.
not bad just 100V.

no current is flowing, the coil is clearly not resonant now. so it needs to be tuned.
also higher voltage are needed.

just some 1000V pp high voltage, which is around 3k5V in reality

(nohvpsu)
When I turn off the hv coil psu, (1A 1.1V) then the orange positive bump changes in voltage.
the hv psu at 0V shows 0.1A, so some currrent is inducing the voltage which can be seen as the yellow is not a straight line.
lets see if I can get this tuned without any snubbers

I made some tea, saw that some of the HV wires were close to other non HV wires, so I bent them away (it was the output coil)
and put the power back on. Changed nothing else.
but it paint a differnet picture on the scope...!!!????

hc psu 0,00A 17.0V
HV psu ?A 2.7V

same frequency 14,5kc/s

voltage at the drain is now in the hundreds, much higher than before. Why is that? now I do need a protection circuit
for this I want a diode parallel over the choke, so the positive can flow back into the capacitor bank for reusse of the energy.
and what is happening with the current? nothing is switched at that point where the current spikes negative!
only the mosfet body diode starts conducting the negative voltage.

blue is still the signal at the gate

I put a diode over the choke. and turned it on with 17V still on the PSU.
The lamp lit brightly, and my audio produced noise.
I saw a large voltage on my scope (dont remember what is was).
I quickly turned it off again, reduced the voltage to zero, and turned the system on again.
I slowly raised the voltage to 2.7V 0,29A on the HC psu. HV psu still turned off.

Luckily the system still works, and also my scope.
pfew.

so with the diode parallel to the choke, current flows, and the coil becomes resonant with high amps.
weird.

This works differently that expected.
orange (probed between hc coil and choke) shows some ringing, and a spike of 46V.
This will be a problem when the high voltage coil is powered, as it then is amplified.)
dc voltage does look good. but the ringing at turn off doesnt look good.
I want a single voltage change, no ringing.

resonance also looks like a weird sine wave.

This all must be related to the capacitive coupled HV coil, which is grounded on neutral

the HV yellow is a flat line, so I left it out (when I power the HV coil also, all remains the same, the orange spike grows just a little bit)

I can replace the parallel diode, with a diode and a resistor in series, to reduce the current.
but first I like to understand how this can draw current. That diode is good. I checked. it blocks the positive psu.
what dark magic is this powering the coil
 :emperor:

this might be a solution.but the current limiting 2.2k ohm might allow the spike to still be high