I went to the sauna today and got a massage.
I was totally at peace.

then I realized.
the cold current induced by the M mosfet discharge,
flows in the series resonant HV coil.

So there is high voltage WITH high displacement (cold) current in that high voltage (extra) coil.

this is what starts the parallel resonance.

the push pull transformer secondary endings are 180 degrees out of phase,
so the sine wave gets chopped by M and this can have a duration of 25% duty. after which the body diode of M conducts to ground the negative cycle. then when the positive cycle starts again the voltage sine builds slowly up. to its max positive again, where it then is fasted chopped to ground by M again.
that voltage is from the push pull.

Meanwhile the other side of the push pull is 180 degrees oit of phase, so the high voltage is maximum negative voltage at the M. chop.

that is when the displacement current sine wave starts. or... is it?

any ways.
the conditions are perfect to get the cold current 180 degrees out of phase with the magnetic current.
so counter flowing.
just like the water vortex and the central air column.

so... two high currents counter flowing.
one in the HV coil the other im the HC coil.
both acting as primaries for the output coil.

real power from the combination of these 2 opposing currents.

another quarter, nearly there.
cant wait to test all these insights.
soon

I ordered some IRFP260N mosfets to replace the irfp460 in the push pull. They should run a  lot cooler, since the internal resistance should be lower.
also got some c3m0260090D mosfets, which have a very fast turn off and turn on, for the M mosfet.

I also will remove the transistors from the PCB, and re adjust the VCC of the delay IC, to get an output pulse below 3.5V

I also have 2x 750V dc generators from AliExpress.
I can power them with my isolated DC DC converter.

This way I can create 1500V dc from them, and use it in a coil capacitor test setup,
Whereby one of the coil plates is quickly changed in voltage

cool! it works. I now have a perfect pulse. no transistors needed.
I just lowered the voltage supply of the delay IC to 3.4V to get a pulse of 3.34V
8% duty for now, at 60.9kc/s

now lets see if this can work with the next IC (tl494)
nope. strange when I insert the u6 IC, the pulse shape becomes a perfect square, which is good, but only at the output of the delay.
Then there are 2 resistors. one to ground of 1k2 ohm, which appears to be perfectly fine.
but then in series to the next IC, there is another 1k2 ohm after which the signal is gone completely.
So that resistor will go.

strange the input measures 4M5 ohm. that should have worked... hmm

hmm...
The voltage drops to 1.05V when no resisitor in series is present, only 1k2 to ground.

not enough power to drive that input of 4M5 hmmm
that is weird.

another challenge.

But it did work before, with the transistor. So... I do nneed a transistor to buffer the signal and give it more current.
still weird, but it is what it is.

ah wait... max 3V on input not 3.5v
I need to lower it further

I made R7 ground to emitter 100ohm
and now I need to scale R9, from out to the base.
R24 is 1k2

I had a short somewhere and spent hours finding it.
Then it suddenly vanished without a cause, very disturbing.
So time for a break. At least I had a signal for just a short while.

But not square enough. the input is from 2.5 to 3.0 V so I can't have a sagging voltage there.

u3 output is max 25mA
and the transistor is max 250mW

Voltage from the puck converter is now around 3.5V
Which directly feeds the transistor at the collector.

so output goes to 1k2 to ground, then should also  go though a current limiting resistor to base, through 100 ohm to ground

I could use the inverting output, and use a pull up resistor.

also instead of an emittor follower
I could use an inverting buffer, with the output on the collector.

if I use the inverting output then the duty cycle should be 92% instead of 8%
and the 1k2 is then tied to the 3.5v supply

when the signal is high then the base of the transistor will turn on, and only the 100 ohm to ground from the emitter will then limit the current. so still a base resistance is needed,
to further limit the base to emitter current.

its one of these days...
no music over the speakers, because some how the output was muted.
psu needs to be set to 1.1A 16.6V for it to turn on, or else it behaves like a shortcut.
probably due to the large inrush current of the large caps.
maybe a choke to limit that inrush surge. or smaller caps...

I used pin for which is the inverting output.
still used the 1k2 resisitors.
one to base the other to V+ which was set to 3.5V
measured along the signal path.
at the base the signal was already far below where it should have been.
so, now to scale the resistors.
definatly need to make some changes in the relations.
last voltage was at the emittor 100 ohms to the input (IC was inserted)
so... I should not use 92% duty, but 8%, just need to set it differnetly. but that is fine, as the pot was inverted, and now it should turn in the right direction. nice.

enough for today.

hmmm.... what if I just skipped the tl494 and went straight to the gate driver?
if it has a ttl or cmos input it would work.
my gate drivers have that. hmmm

If I want only one mosfet, I should tap into the output of the first tl494 as the clock is double the frequency.
It would not really matter, as the body diode of the mosfet, would already conduct the transformer negative half to ground.

The tl494 has an output of 200mA. it also needs to drive the ir2110, so lets say, I'll use 10 mA that is enough,
then use a voltage divider to get the right voltage for the delay IC HC123.
15V signal, 10mA is 1500 ohm total for the divider.

if I want 5V then the ratio is 500 ohm plus 1000 ohm. or 470 plus 820 ohm

I would love to use a voltage divider for the delay IC output, but it has max 25mA and it seems the tl494 input low in impedance altough I measured 4.5M ohm dc resistance. weird

https://www.onsemi.com/pdf/datasheet/tl494-d.pdf
page 10
slave Rt pin 6 needs to be connected to Vref! pin 14

I have it tied to ground!!!

OMG it works!
pfew. very glad I decided to look at the onsemi datasheet, which gave the correct information.
it was the pin6 tied to ground that messed up the signal.

so now I have the inverting output with 2x 1k2 in parallel (600ohm) to base (no resistor to ground)
emitter has 1k2 to ground.
annd it works

2 stable signals.
for testing I only need one.
But I realise, I need to tune higher, as this is half the clock so 30kc/s
the trim pot is easy for that.
I can stable tune to very low duty cycle.
So sweet to finally have this working!!
I already updated most of the PCB so if testing is successful I can order it again with all the mistakes polished out.

since the clock is now double the frequency it sits at 123 my delay is lower in range.

I measured the voltage of the HC coil nnot the hv coil, so that might give a 180 phase shift again, I should thus check that before.
I now use the inverting output of the delay, so I could easily use the non inverting output again, reset the duty cycle and tune the pulse to th emax negative voltage of the HV coil.

I have everything setup so I can also record

yes phase is 180 reversed between HC and HV coil, so I should indeed use the non inverting output.

that was easy.
I now have it tuned as should be. pulse is 7% duty, at negative Vmax of the HV coil.
HV is now around 850Vpp  which is fairly low, but safe enough to connect the mosfet to the HC coil.
power right now is 5.3V 0.4A so plenty of head room I should be easily going to 15V, and get much higher voltage.
But first I need to confirm that the HC coil really starts being parallel resonant, by measuring the current.

Because this will make the HC side of the PP transformer  high impedance,
While the HV coil is series resonant and low impedance.
I hope that the voltage of the resonant secondary of the pp transformer, thus will mainly shift its high voltage to one side, the side of lows impedance, where the HV coil is connected.

now it becomes interesting!!!

oh! another thing.
since I only use one current coil,
I only need one pulse.
But that means I can switch the second Tl494 from push pull into pull pull, and use the 2 mosfets in parallel mode.
Vgthres should be equal,
and pin 13 (output control) should then be connected to ground instead of the Vref
if I do that, F out will be the same as the F clock

first test appears to be success full
in blue the current of the HC coil, measured on the high voltage side, pointing arrow towards coil.
around 400mA pp on the HC coil
and 400Vpp on the HC coil.

I noticed large oscillations on the gate, but that might be from not being grounded.
I left it running for a moment without problems, maybe a minute.

The current of the HV coil looks to be 180 degrees out of phase with the HC coil which is good.

I only used 4.2V 0.22A for this test. 0.92W

I need to check the voltages of the HC coil, and see if the tuning is still good.
I expect the voltage on the HC to be lower than the HV coil due to the high impedance parallel resonance of the hc coil.

To soon to celebrate, but, it does look like the HC coil produces current and thus is resonant

I'll have to do more measurements.
check if the voltage at the drain of the mosfet is low enough to raise the psu

check of there is a voltage difference on both sides of the parallel capacitor, probably small due to to low inductance and high capacity.

then maybe install a bypass diode over the body diode, but it would need to be faster and V thres should be lower.

Or even better, add a series diode from source to ground, but that would need a pcb modification.
if done, then a parallel capacitor to the mosfet comes to mind, like I did before. but that was very different as it would be grounded on the possitive psu. Can that still work? as the coils are now floating, with the only exception when the mosfet is on, it connects to neutral.
hmm. then with cap it would connect to psu positive. that would be ok. but that cap would never discharge again.

ok this needs to be drawn out.

yellow is M mosfet gate

Green is HV coil
orange is M mosfet drain (PP secondary to HC)
blue is the current on HC (on resonant end)

First scr14 was without the HV probe connected (green)

tuning is essential, as can be seen by the change in current (blue) when the capacity of the probe tunes the HV coil.
HV has indeed higher voltage as predicted, but not as high as I hoped for. maybe I need to tune better. with a loose coupled HV probe.

But! already 2.5A pp so the parallel resonance definitely is working!
I did not tune the freqeuncy used the same as yesterday 63kc/s

Strangely the pulse has again phase shifted to the positive maximum of V+ as it did once before.
I can solve this phase glitch, by setting the clock to half speed (tapping into one of the outputs of the first tl494 and using a resistor divider to feed into the delay chip. while setting pin 13 of the second tl494 to a pull pull.

What really suprises me, as that the HC voltage is not rectified at all. is is as iff the mosfet isnt even there. would it already have died on me?But if so, then how is that parallel resonance being created?
HMM
PSU was 7.4V 0.38A

also:
the yellow trace is the output of the second tl494 the voltage is good around 14V but it rings. why? it has a 1.2k to neutral.
everthing floating isn't funny... I need ground. or... an isolation transformer

now lets see what current is available at the HV coil

I put the probe on the HV coil, so blue shows that coil, less current but, it is there. 500mA pp look at the phase... arrow pointing towards the HV coil. but that is actually a capacitor plate. weird phase angle. (is this displacement/cold non magnetic current?)

suddenly just after a few seconds of turn on, the MOSFET started working.
The yellow trace from the tl494 now shows very high noise levels.

The orange HC coil, MOSFET drain, shows indeed the rectified negative half.
And the positive voltage is indeed chopped to neutral.

it looks like the pulse is to early, the green HV max, is after the discharge.
green now is only 698Vpp while orange sits at 220V peak, so I could raise voltage further.
PSU still at 7.4V but now only 0.12A

If I would change the pulse to the right, to -Vmax of HV coil, that would also shift the current on the HV coil, and bring it in phase with the HV coil voltage. But then the relationship to the current of the HC coil???