1- I will advice you to use a smaller L2 coil . I mean with a low ratio of turns and thicker wire in relation to L3.
2- about the frequency:
Most of the tests i did, i use frequencies around 80 at 100khz , but this frequencies rise until 500khz when have a load attacked to L3, when load is removed the freq lower again .
3- About the Radiation:
When i conducted my tests i could reach extreme high levels of radiation , much more higher than the values you mentioned , because the air around the coils could reach a level where the air around become ionized . https://youtu.be/gGdKqdpztuY?t=1
Take some minutes and play this video . I kow the circuit in the video is much complex :) but observe how the field behaves polarizing all the metals in the area of the coil including the water in the glass. It happens because in some way the air and surround air becomes ionized and i think is this ionization that create the radiation. This type of radiation could cross metals like aluminum steel , water glass etc.
I don't know exactly what type of radiation is ( alfa particles maybe ? ) but seems important we take care about that .
4- The resistor in L3 .:
I advice you in my last emails to use a resistor in L3 exactly to you tune at max point of current measured in resistor after that , you should remove them of course , this type of circuit dont like resistors :( only inductance and capacity otherwise you will have a source of friction on L3 when you don't have load make the idle current raise and we don't want that right ? when you tune at max current point you should put the capacitor in parallel in the coil to adjust the voltage output . :) YES it will regulate the voltage output, the capacitor in parallel with the coil .
I know that circuit is not easy to tune because myself take some months develop and study it until able to tune it fast and without any problems .
I have a American friend that find the same problem's that you have at moment during the tune of circuit . We take almost two years until we succeed,but I had to send him, a replica of the circuit he had to replicate :) .
About the mosfet normally i use IRFP460 20A 500V Power MOSFET N-Channel Transistor TO-247.
He emailed before about L3 needing more windings than L2, but still having the same mass so a thin wire for L3 or a thick wire for L2.
I might take some other speaker wire I have, and measure the weight, without the pvc coating, so I could make an equal mass coil from it, with less windigns. OR... a l3 coil without center hole, from 10 m of the wire i used for l2, which should give more windings.
The goal is to have more windings in L3 than l2, but equal mass L2 and L3
tuning with the resistor for max current, then remove resistor, and tune without load with a parallel capacitor.
than attching a load, will increase the frequency. So thats why I could not tune the circuit. I had a load attached.
That radiation, seems gamma to me, penetrating everything. only Lead would protect... better not waist to much time tuning, and keep the device off, and at a distance... damn dangerous. This also comfirst my radiation measurements where wrong, due to the HV EMP.
Mosfet he uses: IRFP460http://www.irf.com/product-info/datasheets/data/irfp460a.pdf
gate capacitance around 3nF (dependend on voltage etc...)
going to order some of these. maybe even rebuild the circuit without pulse transformer.
I also recieved some information from J D stating a 10uF 200V cap charged to 200V discharges better faster snapier than a 10uF 400V capacitor charged to 200V.
The energy is the same, but I guess it is about the density of tension build up between the plates of the capacitor. So picking the right kind of capacitor also plays a role I guess. Also looking at internal series resistance, to be kept low for fast charge and discharge.
I have 1nF 1000V wima caps and 10nF 600V wima caps. Might buy some 400 or 500 volt rating caps. cheaper also :)
Again, solid new Info to work with! Very happy with this and inspiring to continue