After following recent discussions concerned with the tradeoffs=
that regulate seismometer performance, I decided to offer the following. =
Just as Chris has years of significant experience concerned with one of the=
two primarily critical factors (electronics), so my physics research for m=
ore than two decades has been focused on the other one of these two factors=
(mechanical properties, especially 'imperfections'). It is the property o=
f real (structural) materials (imperfections associated with creep) that mo=
tivated the development of the professional standard of force feedback (rel=
atively complicated electronics).
Why are the mechanical challenges so important? The answer has=
to do with instrument sensitivity. For a 'pure' pendulum, the sensitivity=
is proportional to the square of its natural period. If one could readily=
make an ideal simple pendulum with a period of 10 s (very long indeed), it=
s ability to detect earthquakes would be a hundred times that of a pendulum=
having a period of 1 s. It must be understood, however, that period lengt=
hening (though necessary) is not a sufficient condition for improved (horiz=
ontal) acceleration sensitivity. It accomplishes nothing (for conventional=
seismology purposes) when the means for period lengthening involves moving=
the center of mass close to the axis of rotation. Such a device can be us=
ed for measuring rotational effects of an earthquake, but that is not yet a=
part of routine amateur studies. To appreciate what I've just said, take =
a uniform stick (such as a meter sick) and hold it vertical with forefinger=
and thumb just above its center (functioning as a physical pendulum). Rot=
ating the stick away from vertical by pushing with the other hand at the to=
p (or bottom), you will see that it can oscillate with a longer period than=
a string/bob (simple pendulum) of the same length. Now laterally (horizon=
tally) accelerate your hand holding the stick and you will see that there i=
s no rotation of the stick whatsoever as your hold-point approaches the sti=
ck-center. If creep were not an issue, this system would have zero (conven=
tional) earthquake sensitivity even as its period approaches infinity! For=
those who want a math-treatment of what I've just described, I can give yo=
u a copy of the invited tutorial that I wrote for the BSSA Special issue co=
ncerned with rotation, titled "Tutorial on gravitational pendulum theory ap=
plied to seismic sensing of translation and rotation". http://www.bssaonlin=
e.org/cgi/content/abstract/99/2B/1050
So what should be the primary concern of amateur seismologists.=
I believe that more critical thought concerning mechanical properties is =
again called for. With such thought the need exists to dispense with 'blin=
ders' associated with specific (previous) individual failures to achieve pe=
rformance levels that were being sought. For example, most everybody has d=
eparted from the use of viscous damping in favor of powerful rare earth mag=
nets, functioning by means of induced eddy currents in highly conductive (c=
opper) plates. I believe there is merit to the consideration of both uncon=
ventional magnet use coupled with a viscous fluid to provide both damping a=
nd also buoyancy for period lengthening; which I now describe.
In the paper that John Lee and I wrote (mouse sensor for pendu=
lum measurements), arxiv.org/html/0904.3070
one will find measurements in which a magnet was used with a ball-point pen=
to support the pendulum, while providing a low-friction axis of rotation. =
The magnet/pen-point is responsible for a 'destoring' force (away from ver=
tical equilibrium), whereas gravity provides a 'restoring' force (torque to=
ward vertical when disturbed). The torques from the two act in opposition =
to one another, so there is a period lengthening that is beneficial.
A commercial horizontal seismometer that successfully uses the =
same physical principle is the Kinemetrics SH-1. The torque acting in oppo=
sition to gravity is provided by means of a vertical elastic metal strip (s=
pring) that holds the seismic mass in an equilibrium position directly abov=
e it. The instrument employs two magnet/coil subsystems: (i) one for Farad=
ay law detection with which every amateur seismologist is familiar, and the=
(ii) other one for calibration (by introducing a current to the coil) and =
for damping when operational (by placing a resistor across the coil).
What I propose for some of you to experiment with is the follow=
ing. Use my magnet/ball-pen for top support and a viscous fluid (such as m=
ineral oil) for the 'bob' at the bottom to move in. If the bob is suitably=
shaped, and if the amount of its submersion is judiciously chosen, then it=
should be possible to get period lengthening of beneficial type from both =
the fluid and the pen/magnet. The fluid will serve two purposes, to assist=
the magnet in period lengthening and to also provide damping.
As the period gets longer, such systems are increasingly sensit=
ive to pendulum-structure variations. In other words, just as the sensitiv=
ity to an ideal pendulum is proportional to the square of the period, so th=
e adverse sensitivity of a non-ideal pendulum to the influence of its struc=
tural imperfections. Without some means to minimize the influence of imper=
fections, the tendency for the system to 'creep to the mechanical rails' in=
creases dramatically as the period lengthens. The conventional means for o=
vercoming this creep is to use electronic-based force feedback with a magne=
t/coil actuator. I believe, however, that amateur experimentation might di=
scover a bob shape with useful liquid-buoyancy (like that of a properly de=
signed hull of a ship that is stable rather than unstable). If so, it shou=
ld greatly facilitate keeping the pendulum in an operational position.
I have not here mentioned sensor type. For those with interest=
I can offer thoughts based on my long-time experience with both velocity a=
nd position types.
Randall Peters
=
; After
following recent discussions concerned with the tradeoffs that regulate
seismometer performance, I decided to offer the following. Just as Ch=
ris
has years of significant experience concerned with one of the two primarily
critical factors (electronics), so my physics research for more than two
decades has been focused on the other one of these two factors (mechanical
properties, especially 'imperfections'). It is the property of real
(structural) materials (imperfections associated with creep) that motivated=
the
development of the professional standard of force feedback (relatively
complicated electronics).
=
; Why
are the mechanical challenges so important? The answer has to do with
instrument sensitivity. For a 'pure' pendulum, the sensitivity is pro=
portional
to the square of its natural period. If one could readily make an ide=
al
simple pendulum with a period of 10 s (very long indeed), its ability to de=
tect
earthquakes would be a hundred times that of a pendulum having a period of =
1 s.
It must be understood, however, that period lengthening (though necessary) =
is
not a sufficient condition for improved (horizontal) acceleration
sensitivity. It accomplishes nothing (for conventional seismology
purposes) when the means for period lengthening involves moving the center =
of
mass close to the axis of rotation. Such a device can be used for
measuring rotational effects of an earthquake, but that is not yet a part o=
f
routine amateur studies. To appreciate what I've just said, take a
uniform stick (such as a meter sick) and hold it vertical with forefinger a=
nd
thumb just above its center (functioning as a physical pendulum).
Rotating the stick away from vertical by pushing with the other hand at the=
top
(or bottom), you will see that it can oscillate with a longer period than a
string/bob (simple pendulum) of the same length. Now laterally
(horizontally) accelerate your hand holding the stick and you will see that
there is no rotation of the stick whatsoever as your hold-point approaches =
the
stick-center. If creep were not an issue, this system would have zero
(conventional) earthquake sensitivity even as its period approaches
infinity! For those who want a math-treatment of what I've just
described, I can give you a copy of the invited tutorial that I wrote for t=
he
BSSA Special issue concerned with rotation, titled "Tutorial on
gravitational pendulum theory applied to seismic sensing of translation and
rotation". http://www.bssaonline.org/cgi/content/abstract/99/2B/1050
=
; So
what should be the primary concern of amateur seismologists. I believ=
e
that more critical thought concerning mechanical properties is again called
for. With such thought the need exists to dispense with 'blinders'
associated with specific (previous) individual failures to achieve performa=
nce
levels that were being sought. For example, most everybody has depart=
ed
from the use of viscous damping in favor of powerful rare earth magnets,
functioning by means of induced eddy currents in highly conductive (copper)
plates. I believe there is merit to the consideration of both unconve=
ntional
magnet use coupled with a viscous fluid to provide both damping and also
buoyancy for period lengthening; which I now describe.
=
; In
the paper that John Lee and I wrote (mouse sensor for pendulum measurements=
), arxiv.org/html/0904.3070
one will find measurements in which a mag=
net
was used with a ball-point pen to support the pendulum, while providing a
low-friction axis of rotation. The magnet/pen-point is responsible fo=
r a
'destoring' force (away from vertical equilibrium), whereas gravity provide=
s a
'restoring' force (torque toward vertical when disturbed). The torque=
s
from the two act in opposition to one another, so there is a period lengthe=
ning
that is beneficial.
=
; A
commercial horizontal seismometer that successfully uses the same physical
principle is the Kinemetrics SH-1. The torque acting in opposition to
gravity is provided by means of a vertical elastic metal strip (spring) tha=
t
holds the seismic mass in an equilibrium position directly above it. =
The
instrument employs two magnet/coil subsystems: (i) one for Faraday law dete=
ction
with which every amateur seismologist is familiar, and the (ii) other one f=
or
calibration (by introducing a current to the coil) and for damping when
operational (by placing a resistor across the coil).
=
; What
I propose for some of you to experiment with is the following. Use my
magnet/ball-pen for top support and a viscous fluid (such as mineral oil) f=
or
the 'bob' at the bottom to move in. If the bob is suitably shaped, an=
d if
the amount of its submersion is judiciously chosen, then it should be possi=
ble
to get period lengthening of beneficial type from both the fluid and the
pen/magnet. The fluid will serve two purposes, to assist the magnet i=
n
period lengthening and to also provide damping.
&nb=
sp; As
the period gets longer, such systems are increasingly sensitive to
pendulum-structure variations. In other words, just as the sensitivit=
y to
an ideal pendulum is proportional to the square of the period, so the adver=
se
sensitivity of a non-ideal pendulum to the influence of its structural
imperfections. Without some means to minimize the influence of
imperfections, the tendency for the system to 'creep to the mechanical rail=
s'
increases dramatically as the period lengthens. The conventional mean=
s
for overcoming this creep is to use electronic-based force feedback with a
magnet/coil actuator. I believe, however, that amateur experimentatio=
n
might discover a bob shape with useful liquid-buoyancy (like that of =
a
properly designed hull of a ship that is stable rather than unstable). =
; If
so, it should greatly facilitate keeping the pendulum in an operational
position.
&nb=
sp; I
have not here mentioned sensor type. For those with interest I can of=
fer
thoughts based on my long-time experience with both velocity and position
types.
Randall Peters