```Hi all,
Firstly, I don't even begin to pretend to know all of what is
going on here. I'm as interested as anyone in finding out more. I
simply offer a few thoughts for what it's worth.

Drawing on the previous two posts, I'd like to offer a couple of
observations. The first is about the Q of the secondary (from Skip's
post). The Q is high ONLY when no load is placed on the coil.
Measurements made by others show that it drops by around half when
corona ONLY is produced. I also have noticed this drop but have not
quantified it. Prolonged rings still result but not as prolonged as
when a minimal amount of energy is lost. But an attached spark is an
entirely different matter. I have scoped the rings as dropping so far
in number that the multiple beats simply disappear. Since Q is by
definition the inverse of losses (internal to the coil or external to
it), you can see that it is not going to ring for very long at all
when energy is sucked out. Under these conditions, output current is
maximized, and in the extreme, for each half cycle of primary ring-
down, the energy that transfers to the secondary is lost in the arc.
I'd just like to illustrate what I mean. A typical system has
a coupling constant around 0.1 or so. The condition for critical
coupling (max power transfer) is given by : kc = 1/SQRT(Qp.Qs).
Taking the primary Q with gap of around 10, and an unloaded secondary
Q of around 100 (err on the worst side), this gives an unloaded value
for kc of 0.031. By definition, critical coupling occurs when k = kc.
k is set, kc is load/loss/Q dependent. Assuming Qp is constant (spark
gap), the secondary Q has to drop to 10 to meet this condition. The
system can do no better than lose energy at the rate at which value of
k allows it to be transferred from primary to secondary. From this
you can see where the oomph of a magnifier comes from - a much
higher value of k (greater rate of energy transfer) with a consequent
reduction in gap conductions and hence losses. Also, a far greater
rate of charging the toroid to max potential. I'd suggest that this
improvement in performance shows just how bad the gap really is.
The multiple beat envelope one gets shows that the system is
vastly overcoupled before spark production.

To something slightly less tomey (phew) - I think using the
camera is a must if we are to make progress in analysing spark
behaviour further. The eyes lie - there's no doubt in my mind about
that. When my coil strikes an attached arc, to the eye the discharge
goes from a purplish streamer to a bright thick blue white channel
that slowly rises. The camera tells quite a different story. It shows
the Richard's banjo effect in detail, and also shows that the
individual streamers are not not as bright as the eye would suggest.
This brilliance may be a cumulative time-lapse trick by the brain.

Ready for the bricks to arrive,
Malcolm
```