Ron Thompson wrote:
>   ...
> receiving antenna is moved out of dead centre between the two radiating
> plates (where the two fields from the two radiators should produce a
> null point) it will capacitively couple to the closest plate and have a
> voltage on it that is in phase with the closest plate...
 
This approach sounds fine.  One variation that might be interesting is
to mount the radiating panel on the boom, but make that panel out of
double sided PC board.  Drive it with a center-tapped RF transformer. 
That way your 180 degree phase is on opposite sides of that panel and
you can resonate the capacitance of the panel easily, which makes
driving it easy.  Now the pick-up plates will be stationary and close to
the detector circuit and easily shielded with short lead lengths and you
can null their output with simple summing which is then fed to a
synchronous detector (an RF mixer) in order to recover the direction of
displacement.  Since the voltage on the radiating plate can be in the
volts range and mixers can have microvolt noise levels, the potential
exists for 120 dB dynamic range - now you've met many of your
requirements.  Everything mentioned above can be fairly hi-Q - read low
loss - so power consumption can be restrained and the additional side
benefit is hi-Q also results in good S/N levels which is necessary for
low noise floors.

I have to make additional comments on the area of VCOs.

First, in response to my comment about noise in the power to your VCOs
you mentioned using a 78L08 regulator.  Unit IC regulators have many
microvolt noise figures.  It can be a substantial problem in low-noise
instrument designs that require low noise power such as for references. 
The problem stems from the use of buried zeners or the even noisier
band-gap references internally.  The standard solution is to use lots of
RC filtering but this becomes difficult at really low frequency -
frequencies such as in seismic records.  One solution was to go to
mercury batteries as a reference - very low noise figures.

Second, for low-noise in a VCO, it is typically locked to a low-noise
crystal oscillator, which effectively gives it hi-Q.  Your application
by definition has to be un-locked.  This is a recipe for noise -- lots
of it.  Hi-Q LC oscillators are possible, but the detector scheme would
have to change.  An example I can think of is to use a high frequency in
the variable LC oscillator (not a VCO).  Implement the oscillator with
the boom causing variable capacitance, which can be done without
requiring any wires, so only a plate or tube is on the end of the boom. 
Take the oscillator output and use a D-flop mixer using a crystal as a
reference to bring the frequency down to the audio range.  Now you can
go a couple of ways.  A CMOS VCO such as a 4046 locked to the audio note
with the control voltage used as the output measure of displacement.  Or
feed the the audio frequency into a freq-to-voltage IC.  (National used
to have one and I believe there are others.)
This can be a low current scheme and low noise that has no sensitivity
to power line frequencies.

Regards,
Charles R. Patton
__________________________________________________________

Public Seismic Network Mailing List (PSN-L)