Spark Growth and Dead Time

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "by way of Terry Fritz <twftesla-at-qwest-dot-net>" <FutureT-at-aol-dot-com>

All,

I was thinking some more about the whole spark length/
spark growth issue.  A subject which I haven't seen discussed
much is the dead time between bangs or bursts.  This dead
time may be very important for producing long sparks, and
for keeping the input power low for a given spark length.

In a spark gap TC, the bang is a short duration event, taking
only a few hundred microseconds or less.  In truth, the
duration is longer than that as the resonator rings down, but 
the energy quickly falls to a very low level, which I doubt contributes 
much to the spark behavior.  If the break rate is 120 bps, then
the dead time is about 20 times longer than the bang duration,
if we ignore the weak part of the ringdown.  This long duration
dead time may be very important for letting some streamer
(leader) ions or heat dissipate before the next bang.  At high
break rates, the dead time is shorter, and this may let the
streamer channels get too hot, making the sparks too bright
and fat, hurting the spark growth "efficiency".  The sparks 
develop a low impedance, which loads the coil and prevents
the voltage from rising enough.  Conversely, if the break rate is 
too low, the spark channels cool off too much, and the sparks 
cannot grow very much.  By controlling the dead time, we are 
controlling the impedance, brightness, and thickness of the
spark, and the ability of the spark to grow efficiently.  We are
also controlling the amount of power needed to produce a given
spark length.  Above, I more or less dismissed the weak part of
the ringdown as being unimportant, however it may actually 
serve a vital role in keeping the arc channel somewhat active.
If so, this may change the story somewhat.

By using a large toroid on a coil, the spark breakout is delayed
some.  This delay increases the dead time a little.  Maybe this
is one of the benefits of using a toroid.  It may be that the time
difference is insignificant however.

The bang or burst duration also affects the spark behavior also.
Long burst durations, such as in CW coils, also make the sparks
bright and fat.

In a typical CW coil the burst duration is about equal in length
to the dead time.  This gives the CW sparks their typical bright
bushy appearance.  As the dead time is reduced, the sparks
become more and more bushy, short, and plasma-like.  
Eventually, when dead time is eliminated, as in a filtered DC
CW coil, the spark is very short, and forms a full cone of
plasma.  In a recent experiment, I powered a VTTC from a
1/2 wave rectified unfiltered supply to produce 120 PPS operation.
The sparks were shorter, and bushier, despite a greater input
power.  The dead time had been greatly reduced, and the 
spark length suffered accordingly.

In reality, it may not be necessary for the dead time to be
20 times longer than the burst time.  It may only be when 
the dead time becomes shorter than the burst time, that
the spark length suffers most.  Still, a short burst time is
needed to keep the input power reasonable.  This is one of 
the reasons that CW coils are so "inefficient".  The burst
times are so long, that most of the power is wasted in making
the sparks bright and fat, instead of long.  Some folks may
want a bright, fat, short spark, and that's fine of course too.

The best durations for the burst and the dead time are a
compromise......a compromise or tradeoff, which gives adequate
brightness, reasonable input power, and long sparks.  

Regarding CW coils, reducing the burst time, and increasing
the dead time may not make the sparks longer, but it will
reduce the input power.  This saved power can then be used
to increase the bang size (burst energy).  Burst energy is
the real key to long sparks.  The other factors just help to
optimize things.  A "trick" to get the longest sparks from a
CW coil, despite a weak burst energy, may be to run at a
rather high pulse rate, of around 400 PPS, along with a short
pulse duration.  This may not be too efficient, but may be
the easiest way to obtain longer sparks within the limitations
of low cost semiconductor devices.  A better alternative would
be to arrange multiple semiconductors to permit higher voltages
or currents to be utilized.  

I think Gary Johnson was able to vary his pulse rates, and
durations in his large SSTC?  I haven't looked over his work
recently, and I don't remember the details.  It's possible
he addressed some of the issues I mention here.

I will soon be testing my 120 PPS VTTC with a rotary sync
switch which will permit me to reduce the burst duration.  This will
also reduce the input power somewhat, and increase the dead
time.  I'll report on the results.

Comments welcomed.

Cheers,
John