I will keep the fence active. Maybe the dog is smart enough to know what
the flashing lights are for.
I did notice all the suggestions about freestanding lamps. It's a very
interesting proposition, but I can't understand no clear reference to
ground. Would you have to respect the combined load and limit the number
used? Heck - you could put an infinite amount on a charged up wire and
get no current lighting.
The dog can probably make the connection between the lights and shocks
relatively quickly. Animals are a lot smarter than we credit them -
they just don't have the same priorities or abilities that we do.
Then you just need the lights?
There's no magic with operating gas discharge lamps with only one
connection. High frequency (20 KHZ and up) high voltage electricity
has a relatively low X sub C (capacitive reactance - expressed in ohms
but applied to AC).
AC will pass through a capacitor even though there is no DC path
between the leads. The higher the frequency, the lower the resistance
to AC, or/and the higher the capacity the lower the resistance to AC.
Remember seeing "plasma globes?" Low pressure gas breaks down at a
lower potential than high pressure gas - that and the capacitive
reactance makes the arc follow a hand brought near the globe.
Neon tubes and lamps have a partial pressure of gas in them and ionize
fairly easily. (Too low and it won't work - too high and it won't
work).
Getting back to fences and neon lamps . . . your fence charger almost
certainly develops a high voltage low current pulse by using the
collapsing magnetic field of a transformer to generate the voltage.
The faster the collapse the higher the voltage. The more turns of
wire the higher the voltage also.
With that type of generator, you have a fast rise time that is "rich
in harmonics." The faster the rise time the more/higher high
frequency present. That and high voltage may be enough to break down
the gas in the neon bulb and ionize it. The capacitive reactance to
ground and nearby grounded objects, completes the return path for the
electricity.
There is no free lunch - but if it works at all there should be no
limit to the number of neon bulbs that will flash when the spike comes
along. The bulbs are just turning some of the energy that would be
lost to capacitive reactance anyway (actually it may take a smidgen
more energy because the surface area is greater - make a capacitor's
plates larger and you increase the capacity and lower the capacitive
reactance).
It won't be as bright as with a hard wire connection - but you really
have to try it and see what happens.
Tesla coils can light fluorescent lamps from twenty feet away without
wires using similar principles.
Touching an ordinary incandescent lamp to a high voltage terminal on a
working spark coil will light the lamp with a faint blue glow (partial
pressure of nitrogen in the bulbs - so they work like plasma globes).
Some lamps light green when subjected to high voltage high frequency -
vacuum filled bulbs with 100,000 volts on them produce X-rays, and
flint glass fluoresces green with X-rays.