<div class="moz-text-flowed" style="font-family: -moz-fixed">Dex and Ken,
The physics of spark propagation is markedly different for positive
versus negative sparks (in a divergent E-field, such as around a TC
topload). All other things being the same, positive sparks propagate
more "efficiently" in air. Once initial breakout occurs, a
positive-going high voltage pulse will travel further than a similar
negative-going pulse. In a diverging E-field, a positive spark will
bridge a gap at a lower voltage than a negative spark. This is still
true, even though negative corona will "break out" at a lower voltage
than positive corona. These "polarity effects" are well known by
professional high voltage workers and engineers.
Ken is indeed correct - there is an "optimal" voltage risetime that
leads to maximum propagation "efficiency". One noted researcher, Yuri
P. Raizer, has developed a relationship for the optimal voltage
risetime for a positive spark to travel a distance of L meters ("Gas
Discharge Physics", page 362):
T(optimal risetime) = 50*L (in microseconds)
Unfortunately, although the above relationship appears to work quite
well for monopolar impulses from Marx Generators, it's not at all
clear how (or even if) the above relationship can be adopted to the
complex waveforms of Tesla Coils. Using either the RF waveform or
envelope leads to relatively low operating frequencies for typical
coupling coefficients.
We also know that the longest TC sparks are not obtained during single
single events (bangs), but instead via bang-to-bang growth. Newer
sparks build on the heated channels of their predecessors when the
break rate is sufficiently high (>70-80 BPS). This suggests that we
might try combining polarity effect and bang-to-bang growth by
polarizing the system so that the highest voltage peak after ring-up
is always of positive polarity. The positive peaks will provide the
longest "reach" during propagation. This should be simple to implement
through suitable coupling coefficient and phasing for SSTC, DRSSTC, or
DC-resonant SGTC systems, and should cause optimal spark propagation
for a given input power, frequency, and break rate.
BTW, an excellent book (also by Raizer), "Spark Discharge", 1991, CRC
Press, ISBN 0849328683 can currently be obtained for around $38 or so
on Amazon and other large book sellers. This book was originally in
the $130 range. It is technical, but quite readable considering the
complexity of the subject. Any serious spark researcher should have
this title in their library.
Bert