measuring the coils.

scr731 High frequency coil is now resonant at 556kc/s (loose coupled probe, driven from 20V squarewave gen at 43.9kc/s)
11th sub harmonic= 50.54kc/s
10th sub harmonic= 55,6kc/s
9th sub harmonic= 61.78kc/s

scr733 high current coil resonant at 57.6 kc/s with 3uF parallel to it

added an extra 1uF which makes it 4uF.
high current resonance now at 49.9kc/s (scr735)
measured by driving the high voltage coil with the square wave, while grounding one side of the parallel resonant coil.

measured the high voltage coil again (after being moved) being 548 kc/s.
this maximum was driven at 53.3 kc/s on the current coil. (the 10th sub harmonic)

I want to keep using the same size capacity, so the current is even spread over the caps

ok. the high voltage coil definitely rings to high.
I need to. bring it down.
So I will treat it as an extra coil.

I will make a tuning coil (in series) that is made to resonate at the frequency of the current coil.

it needs high voltage, so low capacity. this means tuning with pure inductance.

I have iron ring cores. but they are small, to get to 50kc/s will be tough.

So I will use 4, and stack them into a single core. And then I'll wind them with my machine like a normal helix coil.

Ill first do a fixed set of windings amd measure the frequency, then I'll try to get it to ring at the right frequency. once done correctly,
I could even add a primary (few turns) and drive it with my pcb half bridge.

the extra coil will ride on top of the voltage of the tuning coil (harmonic coupling).

I tink that will work best

I stacked 4 ringtoroid on top of eachother, and taped them together.
I'll use my coil winder, to wind it.
The goal is to get it to be resonant at 50kcps with pretty high voltages.

I could drive it with another series coil which I already have, gradually stepping up the voltage.

but... since these toroids give a center hole, I can wind a primary over it (at 90 degrees angle) just a few turns te get a high voltage. driven by the push pull.  no problem.

but with the high voltage, I will need to have separation, between the layers of windings. so I think it will be best to tape it down with each layer.

problem still is how much windings do I need?
I could make several taps, and mark them, like layer 5,7 and 9. to get a feel for how much inductance there is, and what frequency they ring.

In the end I can easily cut the wire, measure the ring, and resolder the wire and continue winding.

best would be to do it after a winding is complete.

I put 4 layers on the cores, with 0.4mm wire, and measured it. to my surprise it works very well!
26.1mH 8.3 Ohm and a ringing frequency of 82.9 kc/s Very good!
I will put down several more layers, and give each end a tap, so I can choose the frequency.

I put electric tape on each layer, to reduce capacity. As I need high inductance / low capacity to get high voltages.

I also adjusted the code on my coil winder, as it had a 1000ms delay built in which was much to high. I set it to 100 and now it is very responsive. Nice!

I now have 8 layers of windings, and it rings at 56kc/s

I'll add 2 more layers, to bring it below the 50kc/s
If the frequency is to high, I will need higher PSU voltages to get it working.

Done! the coil is now 10 layers of windings, separated by tape.
it rings at 48.7kc/s
it measures 104.4mH with a Q of 131 (measured at 10kc/s)

great. I have debugged the pull pull. one of them didnt work because u6 pin 5 signal input was to low. it needed to be around 2.5 V
this was caused by the transistors being connected wrong.
after playing a bit, I found out, I can use the 4V supply and leave the top transistor out.
for now it works. so I can proceed.

going to connect the push pull to a primary wound around the resonant coil. lets see how that works out. I need high voltage. so It might be a small problem there

I placed the irfp460a mosfets in the push pull section and the gate signals look fine.
I made a primary of 8 windings of 1.5mm2 enamaled copper wire. but first several layers tape.
This coil probably will need to be remade, as it should give out high voltage. 30kV

The gate signal for the duty cycle of the pull pull section also looks good. I now have it tuned to 48 kc/s and a duty of 8%

I probably will need to add some extra capacity to the gate drivers, to get them to switch really fast.

it all is going in the right direction. :D

time to hookup the push pull coil and tune it into its resonant freqeuncy. and measure the voltage.
Then hook it up to the coils, and see how the frequency changes with the added inductance and capacitance.

moving forwards feels good!
:)

very proud of what I have learned so far, and proud of being able to design and get this circuit to work.

I connected my high voltage coil to the push pull and it works.
pretty decent sine waves but I needed to tune much higher then expected, 105,8 kc/s
Primary side was 7,6A pp
from a DC supply of around 1A and 2.8V
secondary was 723Vpp
so a step up of roughly 1:258
If I want 30kV I would need 116V? no way...

For now I am happy that it works. I tuned for a proper current sine on the primary.
but The secondary still produces high voltage when detuned. so I need to see how the secondary behaves, and if this also is the highest output voltage point.

still I am aiming for 48kc this is much higher at 106. on the other hand, I also havent connected the high voltage and high current coils, which will add capacity, and bring the frequency down.

If I need to rewind the high voltage resonant driver coil (which I now tested, yes lets call it the driver coil) then I will make a tube, where in the ring toroids can slide, so I can tune the inductance (and frequency)

I tuned it to the max voltage
1.24A 2.8V dc PSU
f=78 kc/s
primary current 9.1A pp
secondary voltage 1.18 kV
step up ratio roughly 1:422
so 30kV would need 71V on the input.  I can give 64 so ... not bad
frequency is a bit high, but would be able to work, If I make the parallel caps smaller.

but that current... is high! A ZVS comes to mind. the irfp460A can handle 20A continuously. but it definitely will need more cooling than the heatsinks I now have.

I haven't checked for the spikes yet, and heat on the snubber parts.
still. Very happy with these results
 

pin5 , which is where the coil is connected to the source and drain of the MOSFETS shows a bit of positive spike, but not excessive.
2.5V positive, with a 2.8V dc supply.

I didn't use the mur460 diodes, as the holes where to small.
I do wonder why the positive spike is that much larger. the design is symmetric. is it the ground? because in the end I will need to remove the ground

I attached the high voltage coil, and increased the power supply to 3 A 10V 30W
the mosfets do get hot I need a fan.

Q factor is pretty wide band which is good. I dont have to fine tune for high voltage.
but the voltage is to low. while the amps of the psu are to high.

2.7kV at 67 kc/s
while I aim for 10x higher 30kV

I also measured the current coil voltage, and the voltage  it is clearly present there as well. so the capacitive coupling is good enough.

So.. to drive this I need to make another coil. to step the voltage up higher.
this time, a tunable coil, with variable inductance. again with low capacity.
but now only aiming for 600V. with more windings on the primary to keep the current low.
600V secondary with 20V primary 1:30

I will use this coil, and add much more primary windings, as it is already tuned.
it will give a lower voltage output, and draw less current.

then I'll make the high voltage step up coil to work half wave resonant at the same freqeuncy (tunable with variable core)
then I'll add it in series between the 2 coils

I added more windings to the primary of the push pull coil
when it now is resonant by itself at 89kcps (scr746), it consumes 0.87A at 5V from the power supply. it then has 1420Vpp

but I can tune it down to 48kcps (scr749) and then the output  voltage is still high enough and sinusoidal.
power then drops, to 0.06A at 16.6V PSU, while the secondary voltage is 1433V

So the Idea is to make a tunable coil with a variable iron core. I'll use 5 iron rings for it. wrap it in paper, and wrap another paper tube around it so it can move. then wind a coil on that.

ah now I see.
the HV coil and the HC coil together form a capacitor with which the tuning coil is series resonant.

That is why the push pull coil needs an ac output, it is to drive the series resonance.

and that will create the high voltage.

so to tune, I need the high current coil connected to neutral and the output coil in place and earth grounded, as that gives the largest capacity

core is removable, so the inductance can be tuned.
will get 10 layers at least of 0.4 mm wire

I have wondered about how to isolate the windings, as it should produce high voltage.

it will not be half wave resonant so I need less windings then I thought before.

ok, since I need high Q for high voltage, I have decided to use 1.5mm copper wire for the tuning coil.
Since the output coil and the high current coil, form a series capacitor with the high voltage coil, it makes sense that it will bring the freqeuncy down.

So I will need less inductance.  still the voltage will be very high in the tuning coil, so I will need to separate the windings with isolation layers of tape.

First I will wind the output coil, and measure the capacity with the high voltage coil. I would love to see it close to 500pF but my guess is it will be lower. the current coil already was low by itself 68pf is what I remember.

let's see

output coil has only 4 windings. I will connect them in series, through the center of the coil.
it is 200C silicon wire. not the best. as the output voltage will stay low, but the current will be very high (and hot)

I could have best used thinner wire, with more windings to get a higher output voltage. but it is what it is.
I will probably need to find an appropriate load, with the right resistance to get the maximum power out

ok, the output coil has been installed. now its time to measure the capacity.
I will connect the output- and current coil, together as they will behave as one capacitor plate.

and measure the capacity between the high voltage coil and the output+current coil.

147 pF is what I measure between the HV coil and the current plus output coil.
Less then I wished for. but it will have to do.
The coils fit tightly together with the silicone, so that is good. I will try several more measurements with small changes in the coil position to get the maximum out.

I could have made the ouput coil with a single wire of around 11.5 meters.
just swap it through the center hole.

the coils do not fit tightly enough, so I might surround it with paper tape, or teflon tape

I wound 2 layers of 1.5mm diameter copper on top of the flexible core. then my spool was out of copper...
I isolated with 2 layers of tape between the 2 winding layers.

with the iron stacked toroid core fully inserted the inductance measures 544.6 uH with a Q of 24.5 (100kc/s measurement)
with the core fully removed, the inductance measures 136.9 uH and a Q of 25.4 (also 100kc/s measurement)

Together with the coil capacitor I will now make it resonant and measure the frequency while the current and ouput coils are both grounded. this will give a good indication of the frequency range, and If I need more windings.

I'll also give the current and output coil a tape layer to make it a better fit, with less air.

I still have several left overs of the 1.5mm diameter copper so I can add more layers if needed

I wound the current+ output coils with one layer of teflon tape, and 3 layers of paper tape, to make a more tight airless fit.
and it works. a measurement at 100kc/s gave an capacity of 175pF

ok first test done.
I drove it with the first coil I made which is now resonant at 67kc/s, the voltage is consistend through the 3 coils, and the PSU has max amps at that freqeuncy. so this is not the right series resonance.

I tuned up and found a freqeuncy much higher around 450 kc/s which might be the right frequency.
I am not sure. I assume I need a lot more inductance to get the freqeuncy down.

But I also need another push pull coil that is not resonant at 67 kc/s (but much higher)

I could simply drive the series resonant system from the push pull driver, without the first driver coil. then I know the freqeuncy of the system.
 
that might be better

I remembered I had a large spool of wire. it is 1.4mm thick. good enough!
so I soldered it to the coil, and started winding again.
I have now 6 layers of windings. of which the first 2 are made with with 1.5 mm diameter wire.
My hands need rest. I am not sure if I need to continue winding.

I do feel I could have gone with more stacked toroid cores. I now only have 5, but I easily could have used 7.
That would mean less layers.
I didnt use my machine as the wire is stiff and needs constant tightening.

It starts to feel like a nice handgrenade.