new idea for ferrite output xfmr

I have a whole box here of salvaged ferrite cores, and I think I've 
found a use for them.
I have come to believe that the major disadvantage of using an output 
transformer of the type that I outline in my solid state driver is 
that the large number of secondary turns makes for a less than perfect 
transformer.  I don't think its the leakage inductance (which becomes 
part of the coil series inductance at resonance), as much as the self 
capacitance of the transformer windings. 

I say that since I notice that the driver is under much more stress 
without any load applied, than with the coil applied (cap is absorbed 
at resonance).

[This leads to a driver secret that I forgot to include in my circuit 
description (stolen from the ham radio old-timer): take a 250 watt 
light bulb and put in series with the output transformer.  If anything 
goes wrong, it provides a nice destruction dissipating load, and also 
a nice indicator of coil tuning.] 

Thus much higher performance, especially with higher frequency coils 
can be expected if the number of turns in the transformer is kept to a 

Since Bmax is proportional to 1/Cross Sectional Area of Core; instead 
of having 10 primary turns, one could get away with 1 primary turn 
with 10 cores in use.  Then you could have only 20 secondary turns, 
but would have to figure out some different way of insulating turns.
(I refer people to the Chryssis book or others for discussion of 
saturation of magnetic cores)

This type of transformer would look very similar to the ones in the 
Motorola RF Databooks (by Granburg, I believe) where there is one turn 
of silver plated copper pipe thru the cores and the secondary is 
looped thru the inside of the pipe.  (in the RF application, this 
makes a fractional ohm to 50 ohm output transformer).

Whem I get my mosfets, I will try this out with the lightbulb ballast.  
Gee, maybe I'd better call Digikey and order more fets.  Rob.