It appears I haven't been clear (or emphatic) enough in stating the profound differences between a vertical seismometer and a
pendulum operating as (i) typical horizontal seismometer and/or (ii) tiltmeter.  A simple pendulum does not require feedback;
moreover, when monitored with a position rather than velocity sensor, there is no falloff in sensitivity whatsoever when the
frequency is below its natural frequency (unless one stupidly  insists on the use of a velocity sensor, which for periods
greater than a few thousand seconds WILL NEVER WORK!).  Why so many are OBSESSED with velocity sensing for everything labeled a
seismic instrument is beyond my ability to comprehend!
    The acceleration that begins to become the most important component of what drives a pendulum at low frequencies (below 1000
s period) is that component of the earth's field ( magnitude g = 9.8 m/s^2) that is perpendicular to the case of the instrument;
i.e., tilt.  To see that there is no loss of sensitivity as frequency goes toward zero, simply look at my latest paper titled
"Study of Tides with a Pendulum" at http://physics.mercer.edu/hpage/tidal.html
It's not as though I'm talking about some recent discovery.  The great Lord Kelvin, with his understudy George Darwin (son of
the famous Charles) used a pendulum with optical displacement sensing more than a century ago to try and understand the
influence of the Moon on our planet.  (There is an animated gif (halfway down the page) at http://en.wikipedia.org/wiki/Tide
which might help you to understand how the Moon could influence by more than just the conventional sense of the tidal force.
The offset of the center of the earth from the barycenter by about 0.8 times the radius--causes the earth to be 'whipped around'
with a sidereal period of 27.3 days--bound to be important but mostly unstudied).
In similar manner, with Jim Shirley of JPL I tried with a pendulum (predecessor to the VolksMeter) to understand what happens to
the earth because of this, over periods measured in months.  What we observed in a time interval measuring about 1.5 years  is
very DIFFICULT to understand.  Kelvin couldn't figure it out, nor could we.  And it has nothing to do with electronics
limitations imposed on the instrument.  What the pendulum was measuring [without frequency limitations in the realm of interest
(months to years)] was the exceedingly complex physical changes to its shape  that occur in the earth.
    None of the greats of physics past would have fallen into the `rut'  has snared the world of seismology ('hung up' on a
particular type of detector to monitor the motion).  I once heard `rut' defined as a 'coffin with the ends kicked out'.
(applicable to those who refuse to consider anything other than a detector that responds to the time rate of change of position
of the inertial mass of a seismometer--otherwise called 'velocity sensor').
    So why then am I interested in feedback, which is necessary for a vertical, but not for a pendulum?  Because I (like so many
of the rest of you) would like to see earthquakes from all over the world, as small as technology will allow.  The simple
pendulum would be a GREAT candidate for so doing if I had a facility to hold one whose length were 10 m (or even longer).  As
compared to the VolksMeter, it would be 100 times more sensitive in every frequency range.  (For tide studies it doesn't need
that sensitivity; but for 20 s period teleseisms, it does.)
    Unlike the vertical that has become standard (influenced greatly by LaCoste), an ordinary pendulum is not capable of simple
mechanical 'period lengthening' by means of structural rearrangement.  But what makes any pendulum superior to any vertical
seismometer--is its ability to look at REALLY low frequencies in a way that will ALWAYS be impossible for a vertical.  The
bottom line is that we need to finally understand that different frequency regimes call for different instruments!
There never will be a single instrument labeled the SEISMIC-DO-ALL; since the physics refuses to cooperate.
 Randall