On Jun 6, 2007, at 5:01 AM, Michel Jullian wrote:
Hi Horace,
Sorry for the empty reply, my finger slipped.
There will be no such delay, that was my point, except of course
the subnanosecond speed of light delay for Coulomb forces to act
across a few tens of cm, even if it takes 50 milliseconds for the
"whatever" to cross the gap so that one might expect 50
milliseconds or more would elapse before current comes out the
bottom of the pan.
But that was *my* point. If threads are in place, there is a
conductor across the gap, thus the signal should travel fast across
the gap, not having to wait for insulated drops to carry the signal.
This is not the point I was making. The point I was making was on
the contrary that even if the current is carried by slow drops or
ions or whatever which take ages (milliseconds) to cross the gap,
the signal will still cross at the speed of light (subnanosecond).
Let me know if what I wrote previously makes sense in this light,
otherwise I can explain.
I think your statement makes sense in this context if you assume the
air gap plus drops, acting as a capacitor, conducts a signal similar
in orders of magnitude to what we would expect from a filament.
My underlying assumption is if there are no filaments there will be
no signal - not at the 10 nanoamp level measured by Bill Beaty
anyway, no signal until a filament contact is "made". The fact the
air gap is a capacitor seems to me irrelevant because the capacitance
is miniscule, so it appears we have differing underlying assumptions
in that respect that prevent us from mutually making sense of this.
It appears we both agree and assume the signal should travel fast
across a filament, i.e. not at the filament molecule speed. It was
about this I said: "But that was *my* point." We also both agree the
signal would travel fast across an air gap capacitor, but apparently
we have highly differing assumptions about the ability of the air gap
capacitor to conduct a signal, or at least the comparative magnitude
of strength of signal from an air gap vs a filament.
I suspect we are not communicating at all and have not made sense
about the nature of the signal onset though. If a continuous signal
actually is detectable and carried by the capacitance of the drops
plus air in the gap, then its onset, the rise in its magnitude,
should be slow. The onset of a (continuous) signal carried by a
filament should be very fast - at the moment of "make", and
disconnect fast, at the moment of "break".
If there are filaments making a connection then the signal travels
according to the transmission line characteristics of the filament,
which may well not be light speed, but will still be fairly fast, sub-
nanosecond I would expect. I expect there may be some surprises in
this regard, at differing frequencies and rise times, related to the
mass of the probable charge carrier in a filament, the proton.
A signal carried only by the differing potentials of independent
drops would take millisecond delays, dependent on the drop velocity,
and I think we agree on that, have made sense of that. A single
filament's signal onset when a "make" occurs would be very fast, not
dependent on the stream velocity. If in fact the signal were carried
capacitively through independent drops in the gap, where the gap plus
drops act as a capacitor, then the signal onset (of a continuous
signal) would be comparatively slow, because the strength of signal
depends on the capacitance of the gap, thus the geometry of the gap
and the speed of filling the gap with droplets.
The above may seem trivial, but I think the distinction is critical.
For example, given two adjacent target plates, and slowly moving a
source needle from above one to above the other, we would see a
(superimposed continuous AC) signal jump from one plate to the other
very fast if a filament were involved, and gradually switch between
the two plates if capacitive transmission were involved. If lots of
filaments from the source electrode were involved, we should still
see the change in signal amplitude occur in jumps, while if
capacitive conduction is occurring then the transition of signal
strengths would be continuous.
Regards,
Horace Heffner