Sweet spots, break rates, etc.

 More observations and comments:
 1.  A corregated toroid may need to be as much as double the
 size of a smooth toroid to insure just one breakout, and to obtain
 the same spark length.  Large toroids may increase the spark length
 mainly by reducing the number of simultaneous streamers.  A bump 
 placed onto a smaller toroid may result in the same spark length, but
 spark may be a little dimmer. 
 2.  Spark length seems to correlate very strongly with output
 power rather than voltage, at least within some range.  This makes
 precise toroid sizing and secondary geometry much less important
 than otherwise.  A tall narrow secondary may be just about as good
 as a short wide one.
 3.  High break rate sparks may have more of a tendency to arc
 downward and hit the primary for a given toroid size and spark length.
 4.  Question:  Do high break rates become more efficient at higher
 power levels?  I cannot test this, but I could build a smaller TC 
 (smaller than my present one), and see if high break-rates become 
 even less efficient at even lower power levels, which might give 
 some clues to the question.  It would be interesting if the efficiency
 of high break-rates is strongly power-dependent.  I've never heard
 anyone suggest this before.   Comments?
 John Freau