I just tried substituting a 220 ohm resistor in series with a
carbon-carbon spark gap as the base resistor. Indeed I get some
spark, although not very big.

However, the intriguing part of this is that at one point my battery
bank voltage jumped from 12.82 to 12.88 within about 15 minutes. In
all my tests thus far I have never seen this large of a jump in 3
batteries at this stage of charging in such a short period of time.
Total gain of .18 V in 15 minutes.

For the spark gap, I used two pieces of graphite "lead" from a
machanical pencil, barely touching so that it created tiny sparks.
One problem was keeping the sparks going probably because the
composistion of the pencil has many impurities that burn off and pit
the graphite, changing the gap width and breaking the connetion in a
very short time. My best results seemed to come when the ends of the
graphite seemed to be barely touching so it would stay in place.

I will see if I can get a pure carbon stick from an art supply store
to see if I can get more reliable results.

Still too early to tell if this is the missing link, but the pieces
look like they could fit. It would make sense that John Bedini would
tell us to build the SG EXACTLY as drawn, using a single 680 ohm
resistor that may degrade over time to create this effect.

Also, YoTango, it is very possible that Marcus did indeed use a spark
gap to make this work -- a spark gap in his half-fried potentiometer
between the wiper and the carbon.

Thanks to Marcus & Stephan for opening this door...

Jim