Chris,
Yes, long drawn sparks are more loss, but if the spark jumps "ahead" of
the
electrode presentations as it will do anyway, then there's loss anyway in
a typical system. In any case the system did not reduce my spark length
but reduced the input current and as you said, that's a good thing. It's
possible that without the offset, the gap was occasionally "refiring"
(firing more than once per gap presentation), and that may have been
causing the extra current draw. I suggest this because not only was
the current draw higher without the offset, it was also less steady.
"Refiring" is a bad thing. It causes inefficiency. There was a good
writeup
on this subject by Robert Jameson in the past.
I would also say (as I said before) that generally speaking, mechanical
dwell time does not affect spark lengths. There is no reason why it
should.
Of course a very long dwell time can cause refiring (this is affected by
various factors), and this can hurt coil performance.
Consider though an accidental experiment performed by Robert
Stephenson of Canada. He was setting up his large twin coil and
he forgot to turn on the rotary gap. The electrodes just happened to
be at a distance (based on the position of the rotor) where the arc
was able to jump the gap. He ran the coil and it ran perfectly. He
didn't even realize the rotary wasn't spinning. The rotary gap was
behaving as a static gap! Consider the dwell time of a static gap.
It is infinitely long since the electrodes don't move. This shows that
a gap quenches when the system runs out of energy, not when
the electrodes try to tell the gap to quench.
John