Best primary coil for SSTC/DRSSTC?

[Laci]

After experimenting a bit with my coil,I wanted to to change the primary,because right at the beginning Mads said that it has too low primary to secondary coupling for SSTC.So now I bought 3 meters of 0.5cm litz copper wire and made 6 tight turns on a plastic cylinder at the bottom of the secondary.The distance between the secondary and primary is 1.5cm.
I tested this setup on a full bridge of mosfets at around 150v,interrupted.It was interesting to see,that the arcs are not longer with this primary than with the "too low coupling" coil.Before this 6 turn primary I also tried more turns of around 0.2cm wire.
I tried different capacitors with the primary.With the 6 turn primary the best output is with 2x220nf caps in series,but not as good as the low coupling coil tapped at around 5 turns,and 3x220nF caps in series,which produced up to 35cm arcs at 150v with high on times!

In conclusion the new primary coil is worse as SSTC and also as DRSSTC setup,while it should be really better at least as SSTC due to lower inductive reactance.What do you recommend to improve on this primary?Should I go back to the lower coupling primary?

[Mads Barnkob]

That was a bit of a surprise, could you please try to run the numbers on your coil with conical and new tight primary coil through JavaTC and post the results here?

[Uspring]

Metoo, i.e. being surprised. More coupling should improve performance. Here's why:

The primary coupled to the secondary can be modelled by a replacement circuit consisting of an effective primary inductance Leff in series with an effective resistance Reff. The importance of Reff is, that the primary current flows through it and dissipates energy. Actually it is the power consumed by the arc that is being transferred to the secondary then then dissipated by the arc. So the coils interesting power is Ipri^2 * Reff.

For a SSTC driven at resonance, we have the two equations:

Leff = Lpri * (1-k^2)

and

Reff = k^2 * Rarc * Lpri / Lsec

k is the coupling, Lpri and Lsec the primary and secondary inductances and Rarc the resistance of the arc.

Leff and Lpri are quite close to each other with k being around 0.2. Reff is proportional to k^2, so for a given primary current, the power delivered to the arc is also proportional to k^2.

A larger primary inductance Lpri will also increase Reff but at the same time will reduce primary current for a given input voltage due to its higher impedance. So it is best to use the lowest primary inductance as possible. That is limited by the current capabilities of your FETs.

[Laci]

Cheers for the replies.Today I removed the primary from its plastic support to get a bit higher coupling and tested it with my SSTC driver.It works well,but I still can't see any difference with this primary,expect that there are no sparks from secondary to primary at rectified mains.Maybe a bit higher coupling is possible.

[Mads Barnkob]

Could you please also do the same JavaTC attachment for the conical coil

[Laci]

Sorry for the late reply,but I am busy with school things recently... :/
Anyways here are the conical primary informations.Note that the primary has more turns(14) with a total height of 27.5cm,I just ran JavaTC with 6 turns,because the tap point is usually there.
As you can see,the coupling coefficient is higher in both cases then the recommended.

[Mads Barnkob]

There is virtually no difference between your coils, they have identical inductance and coupling to the secondary coil. There must be something else affecting performance if these simulations reflect the actual setups.

I can not pin point from these why it would perform worse.