Yes, the 74J i see is just the peak energy stored in the primary capacitance
in Finn's case. Indeed the real spark energy (time integral of the spark
power during 1 spark event) is certainly higher than 74J (unless of course
he is recycling nearly all of his primary energy, i doubt it).
There is something unusual about Finn's system, at least compared to mine
(which is a bit smaller, but similar spark length). My primary current is
only 1300A when using a 1.25uF tank capacitance at 35khz. I calculate (just
for comparison) that the peak primary energy is only 14J in this case (not
the true bang energy!). The best sparks were about 12 feet in length but
surely not at a low 67bps! I think i see peak performance at about 250bps.
The DC bus power consumption is about 6.5kW in this case or about 26J/bang
(so nearly 2X the peak primary energy). When you compare this to other
topologies you can see the benefit in reduced peak current/voltage capacity
of your primary components, with of course the trade-off that the slightly
lower current/voltage persists for a longer period of time.
Steve