[TCML] 20KV primary drive voltage tips ?

[Chris Swinson]

Gary,

>Yeah, if you have only a very small number of turns, they the spacing needs 
>to be comparably wider.  But remember that gap losses increase 
>significantly as primary inductance goes down.  Most report >much better 
>performance as Lpri increases.

Thats an epic in itself too! If you treat the primary as a current limiter , 
something like 10 turns would be needed to get the current down to 100amps 
peak or so. Lower number of turns would need a much better DVDT/DVDI rating 
in the 1,000's of amps range, but pushing more amps across the cap in one go 
results in quench problems, so larger number of turns will always work 
better to a point if you look at it from that point of view.  I am going for 
higher voltage rather than higher current, so hopefully I will be able to 
gain better efficiency with fewer turns if built right. I did plug all the 
figures a while ago, was something like 3 primary turns would be as low as 
you could go from a practical point of view.


>I couldn't comment on the strange humps, but the simulation makes no 
>attempt to predict racing sparks, which WILL destroy your secondary long 
>before you get close to 0.3

The humps only appear visually from 0.38K. Its probably they start to happen 
a little less maybe 0.30.  I do wonder though if larger dia wire and higher 
frequency will effect the racing arc problem.  My coils are taller than 
normal so maybe that in itself will help things. I never saw racing arcs on 
my coils though I only ever ran up to 500W anyway. The coils I was building 
could take a lot more before being stressed so hopefully the coupling will 
not be a problem for them.



> >A vertical (helical or solenoid) primary is typically shunned in favor of 
> >a
> >flat spiral primary, for the very reason of having TOO high a coupling.
>
> Understood, but if I do go for higher coupling then it could work out 
> better
> for me in that respect.

>Again, let racing sparks be your guide to coupling rather than a waveform 
>simulator.

Could probably be best to build a cone primary, then if racing arcs happen, 
just lift the secondary coil up a inch or 3....  Simulation can't show 
racing arcs, but it does show that >0.38K cause frequency splitting which I 
presume is the casue for racing arcs.  As a Coil is not a perfect device its 
probably other factors could lower the realworld results to below 0.30K... 
I've not done enough work myself on coupling so I don't know what works in 
the real world. Though simulation wise, I would say up to 0.30K would be the 
best max figure...


Chris