I guess what you're saying so on the PCB may arise overvoltage and damage the transistor
2.5 kv peaks need to isolate legs see my old film:
https://www.youtube.com/watch?v=wKKweX0k4JA
for long cables without shielding bring a lot Operating Noise! you need a ground GND metal housing in other types of potentiometers avoid unpleasant surprises.
view photo:
driver Ir2113 it also has the disadvantage (under load at 10K gates. resistor to close the transistor remember)
Another thing is the distortion of the input signal.

on the PCB of Ruslan waveform is different I suppose that it is not inverting ?
Itsu worked on another system
https://www.youtube.com/watch?v=hDHbVJRSHKU

changes in the type of driver tc4420 does the opposite ucc27322
They have the same pins and act differently and the same sea to be with Tl 494 one works up to 1 MHz and the other to 3MHz.
At the moment you have to try different combinations to get a similar type of operation.

     In this Akula video, he is showing how to tune the different circuits, of his second self runner video device, using a function generator and a scope.
  Although I can't understand it, I do get some idea.
  On that particular schematic (which was made by him), hopefully nothing is hidden.
   NOTE: He is giving the kacher's transistor rectified high voltage, from the 28 turns yoke coil circuit.

   https://www.youtube.com/watch?v=-rALEpvBQig

Hi ,
Tesla talk in a conference about the other engineers of their time have some problems to reproduce some of their circuits of high frequency because  this replications have trend to concentrate all the power in output and not in a balanced way to make the "swing" happens.

I see that the most replications of this type of circuits, the main focus is producing the max output of HV and seems not the correct path  , because higher voltages , more work to step down . We could high streams of high voltage in output of kacher driver but will not enough. People  forget the feedback , and this feedback it only happens if the circuit work in resonance.
 
My opinion about grenade coil is that is only a diversion , in fact it make some job if are correct tune , and act like a capacitor in bifilar portion of coil so is so trick to obtain the right values of capacitance,  but if don't properly tuned , it will be a waste of loss.
Any configuration of coil it will work , any have their advantages and disadvantages but if not correct tuned it will not work too.
Member Geofusion show a good video in their youtube channel,  of process  of tuning resonance at output with different caps and it show how could be trick get a output work in resonance.
about push pull configuration i have my doubts about their efficiency in that type of circuit .
I design some circuits to use in my studies and when properly tuned can output some 2kv with perfect sinus  with low current and low voltage input . I left a video to show one of this circuit working and the output scope waveform .
https://www.youtube.com/watch?v=3S4lCUvzn0E.



 Nelson I’ve seen many of your device videos on youtube you assume a lot, would it be humanly possible to explain this circuit to us ?



 

Lol . is only a moment of relaxation with friends and my poor sense of humor 
This circuit is not yet finished and is not a full resonance , because i have sure the main transformer will fry because is designed for low frequency and the frequency are tuned for reasonable work without risk of damage of transformer .
But my path is a bit different of work that is present in this subject.
I only give a opinion about the subject of how drive to provide hv nothing more .
I dont say that the work of others like Akula or ruslan are wrong , but we take differing paths to achieve some similar effects and my suggestion is to people start think by own head , some much divergent opinions can do mess in people when is time to make good choices.
Good luck for your work