Brett,
But think about what you've just said; that after setup (for at least =
a while) an instrument like yours with the 'necessary evil spring' will gen=
erate 'pops', as well as forever generate unavoidable long period noise. T=
he broadband conference that I attended would never have happened were it n=
ot for the professionals' desire to improve performance toward frequencies=
below the lowest corner frequency of your instrument. Yes, you have recen=
tly seen eigenmode oscillations with your instrument after one of the big e=
arthquakes, but superconducting gravimeters that operate on the basis of th=
e Meisner effect have seen not just the same, but also earth 'hum' at mHz f=
requencies on a regular basis after 1998. It is also interesting to note =
that the 'brains' behind probably the most significant commercial instrumen=
t of feedback type (Erhard Wielandt, who holds (if I'm remembering correctl=
y) a patent on the spring used in instruments built by Streckeisen), has st=
ated that sometimes 'hastening toward stability against 'pops' can be manag=
ed by beating with a hammer on the pier that holds the instrument! Whether=
that type of impulsive beating (which I have also used in experiments of y=
ears past, to hasten the evolution of secondary creep) is mainly due to def=
ects in the pier or to defects in the instrument (or both) is beside the po=
int. The fact remains that both the pops and the noise of the spring are t=
he result of real materials having defect structures. And what makes their=
study so hideously complex is that they are not static. Incidentally, the=
paper that 'launched' great interest in your type of instrument was the 19=
82 BSSA article by Wielandt and Streckeisen titled "The Leaf spring Seismom=
eter: Design and Performance", the abstract of which can be read at http://=
bssa.geoscienceworld.org/content/72/6A/2349.abstract
Their abstract mentions that this development 'took off' in 1976, yielding =
the instrument's outstanding performance; i.e., 140 dB dynamic range over p=
eriods from 0.3 s to beyond 300 s, for which ground noise was resolvable.
I am not at all contesting the wonderful properties of force feedback o=
f the type you and Dave Nelson have mastered for purpose of observing earth=
quakes in a way that has not been realized with any other type of seismogra=
ph-excellence that has not to my knowledge been exceeded by anybody other t=
han maybe one professional with skills like that of a master watchmaker, na=
med Gunar Streckeisen. I envy all three of you, for reason of my tendency =
toward something akin (in a figure of speech) to two left thumbs. ( I even=
have trouble 'stringing' the popular computerized Cavendish balance, whose=
design I created, and which is sold by Tel-Atomic Inc to measure 'big G'--=
because the 25 micron diameter tungsten torsion wire that supports the sma=
ll spheres on an aluminum plate, is a difficult component with which to wor=
k (for most folks other than some dexterous women). But force feedback ins=
truments as presently configured cannot be the 'end all' of technology when=
it comes to concern with bandwidth extension below a few hundred seconds. =
I believe this to be the case, not only for reason of the numerous experim=
ents that I've conducted for about two decades, but also because of the sta=
tement that I've highlighted below, taken from Wielandt's well known work =
titled "Seismic sensors and their calibration".
The force-balance principle
In a conventional passive seismometer, the inertial force produced by a sei=
smic ground motion deflects the mass from its equilibrium position, and the=
displacement or velocity of the mass is then converted into an electric si=
gnal. This principle of measurement is now used for short-period seismomete=
rs only. Long-period or broadband seismometers are built according to the `=
force-balance' principle. It means that the inertial force is compensated (=
or `balanced') with an electrically generated force so that the seismic mas=
s moves as little as possible; of course some small motion is still require=
d because otherwise the inertial force could not be observed. The feedback =
force is generated with an electromagnetic force transducer or `forcer' (Fi=
g. 14). The electronic circuit is a servo loop (Fig. 16) like=
in an analog chart recorder. A servo loop is most efficient when it contai=
ns an integrator, in which case the offset of the mass is exactly nulled in=
the time average. Due to unavoidable delays in the feedback loop, force-ba=
lance systems have a limited bandwidth; however at frequencies where they a=
re effective, they force the mass to move with the ground by generating a f=
eedback force strictly proportional to ground acceleration. When the force =
is proportional to the current in the transducer, then the current, the vol=
tage across the feedback resistor R, and the output voltage are all proport=
ional to ground acceleration. We have thus converted the acceleration into =
an electric signal without depending on the precision of a mechanical suspe=
nsion.
My career has focused on the complexities that can accompany the 'small mo=
tions' that Wielandt has mentioned, as another 'necessary evil'-- that beco=
me ever more significant as the frequency of external acceleration drive ge=
ts ever lower-toward the range of lower/lowest eigenmodes --that need addit=
ional study. I see two ways that the unavoidable influence of these defect=
complexities can be made less significant: (i) reducing the 'load' on the=
offending part(s), and (ii) lowering to near absolute zero the temperature=
of the instrument. The latter is not feasible for the vast majority of pe=
ople, but I believe that we amateurs can do some things pertaining to the f=
ormer-if you and others are interested in some of my thoughts on the matter=
.
Randall
Brett,
But think about what you=
’ve just said; that after setup (for at least a while) an instrument =
like yours with the ‘necessary evil spring’ will generate ̵=
6;pops’, as well as forever generate unavoidable long period noise.&n=
bsp; The broadband conference that I attended would never have happened wer=
e it not for the professionals’ desire to improve performance toward =
frequencies below the lowest corner frequency of your instrument.&nbs=
p; Yes, you have recently seen eigenmode oscillations with your instrument =
after one of the big earthquakes, but superconducting gravimeters that oper=
ate on the basis of the Meisner effect have seen not just the same, but als=
o earth ‘hum’ at mHz frequencies on a regular basis after 1998.=
It is also interesting to note that the ‘brains’ b=
ehind probably the most significant commercial instrument of feedback type =
(Erhard Wielandt, who holds (if I’m remembering correctly) a patent o=
n the spring used in instruments built by Streckeisen), has stated that som=
etimes ‘hastening toward stability against ‘pops’ can be =
managed by beating with a hammer on the pier that holds the instrument!&nbs=
p; Whether that type of impulsive beating (which I have also used in experi=
ments of years past, to hasten the evolution of secondary creep) is mainly =
due to defects in the pier or to defects in the instrument (or both) is bes=
ide the point. The fact remains that both the pops and the noise of t=
he spring are the result of real materials having defect structures. =
And what makes their study so hideously complex is that they are not static=
.. Incidentally, the paper that ‘launched’ great interest =
in your type of instrument was the 1982 BSSA article by Wielandt and Streck=
eisen titled “The Leaf spring Seismometer: Design and PerformanceR=
21;, the abstract of which can be read at http://bssa.geoscienceworld.org/cont=
ent/72/6A/2349.abstract
Their abstra=
ct mentions that this development ‘took off’ in 1976, yielding =
the instrument’s outstanding performance; i.e., 140 dB dynamic range =
over periods from 0.3 s to beyond 300 s, for which ground noise was resolva=
ble.
I am=
not at all contesting the wonderful properties of force feedback of the ty=
pe you and Dave Nelson have mastered for purpose of observing earthquakes i=
n a way that has not been realized with any other type of seismograph—=
;excellence that has not to my knowledge been exceeded by anybody other tha=
n maybe one professional with skills like that of a master watchmaker, name=
d Gunar Streckeisen. I envy all three of you, for reason of my tenden=
cy toward something akin (in a figure of speech) to two left thumbs. =
( I even have trouble ‘stringing’ the popular computerized Cave=
ndish balance, whose design I created, and which is sold by Tel-Atomic Inc =
to measure ‘big G’-- because the 25 micron diameter tungsten to=
rsion wire that supports the small spheres on an aluminum plate, is a diffi=
cult component with which to work (for most folks other than some dexterous=
women). But force feedback instruments as presently configured canno=
t be the ‘end all’ of technology when it comes to concern with =
bandwidth extension below a few hundred seconds. I believe this to be=
the case, not only for reason of the numerous experiments that I’ve =
conducted for about two decades, but also because of the statement that I&#=
8217;ve highlighted below, taken from Wielandt’s well known wor=
k titled “Seismic sensors and their calibration”.
The force-balance principle
In a conventional passi=
ve seismometer, the inertial force produced by a seismic ground motion defl=
ects the mass from its equilibrium position, and the displacement or veloci=
ty of the mass is then converted into an electric signal. This principle of=
measurement is now used for short-period seismometers only. Long-period or=
broadband seismometers are built according to the `force-balance' principl=
e. It means that the inertial force is compensated (or `balanced') with an =
electrically generated force so that the seismic mass moves as little as po=
ssible; of course some small motion is still requ=
ired because otherwise the inertial force could not be observed. The=
feedback force is generated with an electromagnetic force transducer or `f=
orcer' (Fig. 14). The electronic circuit is a servo =
loop (Fig. 16) like in an analog chart recorder. A ser=
vo loop is most efficient when it contains an integrator, in which case the=
offset of the mass is exactly nulled in the time average. Due to unavoidab=
le delays in the feedback loop, force-balance systems have a limited bandwi=
dth; however at frequencies where they are effective, they force the mass t=
o move with the ground by generating a feedback force strictly proportional=
to ground acceleration. When the force is proportional to the current in t=
he transducer, then the current, the voltage across the feedback resistor <=
i>R, and the output voltage are all proportional to ground acceleration=
.. We have thus converted the acceleration into an electric signal without d=
epending on the precision of a mechanical suspension.
My =
career has focused on the complexities that can accompany the ‘=
small motions’ that Wielandt has mentioned, as another ‘necessa=
ry evil’-- that become ever more significant as the frequency of exte=
rnal acceleration drive gets ever lower—toward the range of lower/low=
est eigenmodes --that need additional study. I see two ways that the =
unavoidable influence of these defect complexities can be made less signifi=
cant: (i) reducing the ‘load’ on the offending part(s), a=
nd (ii) lowering to near absolute zero the temperature of the instrument.&n=
bsp; The latter is not feasible for the vast majority of people, but I beli=
eve that we amateurs can do some things pertaining to the former—if y=
ou and others are interested in some of my thoughts on the matter.
Randall
=