YOl. 3()) ~o, 1
TRA:\~ArTro~s, A:\fERICA~ GEOPlIVSIL\L U~{o:l
Fcbru"ry 1958
A Long-Period Seismograph System *
FRA~K PRESS, ::\L~URICE EWIXG, AXD FRA~CIS LEU::C>ER
Abslracl--c\ matched three-component seismograph system operating with pendulum
period of 30 sec and galvanometer period of 90 sec is described. As a result of increased sensi­
for periods greater than 20 sec, seismograms from these instruments contain information
not previously available.
Iniroduclioll-The advantages of a seismograph
system consisting of three matched components
are well known. Instruments incorporating this
feature were built by Galitzin, Bcnio.U [1955],
and Hiller. In this paper we describe a system of
longer period than any of these, designed to im­
prove the recording of seismic body and surface
waves of longer period.
It is desirable to limit the response of such a
system in the short-period range. This would
minimize the response to microseisms (which
limit the usable sensitivity of any seismograph
system) and bring out long-period
which
are obscured by shorter period \-ibratiolls on
standard seismograms. The major problem was
that of constructing a stable vertical pendulum
with a period of about 30 sec. A solution was
found
using the zero-length spring principle of
La Coste [193'*l and the recently developed isoelastic
spring materials, and by compensating for the
buoyant effect of atmospheric pressure fluctu­
ations.
Since 1951, a buoyancy compensated vertical
seismometer [£-icing and Press, 1953] with periods
of 15 and 90 sec for pendulum and galvanometer,
respectively, has been operated at Palisades. The
expected ad van tages of the instrument were
quickly realized and, in addition, waves of several
types, previously unknown, were recorded. In
1953, the matched horizontal components were
were
added and the expected advantages
soon realized. During the next few years, ad­
ditional instruments of this type were installed
in Bermuda, Pietermaritzberg, Perth (Western
Australia), and Waynesburg (Pennsylvania).
Excellent results having been obtained in all
cases, it was decided to make available a standard
design of this seismograph system. This design
was engineered by Lehner and Griffith, Inc., and
is being manufactured by them as the Press­
* Contribution No. 8-!2, Division of Geological
Sciences of California Institute of Technology; Con­
tribution No. 264 Lamont Geological Observatory of
Columbia University.
E\ving seismograph system. These instruments
have been recording in Pasadena since 1956 with
To = 30 sec, Tg
90 sec, pendulum and
vanometer being critically damped. They will be
operated at a number of stations during the In­
ternational Geophysical Year.
Description-Figures 1 and 2 are photographs
of the vertical and horizontal components, re­
spectinly. The
sphere on the vertical pen·
dulum serves as the buoyancy compensator.
The upright is made of Invar for the vertical and
of brass for the horizontal components. It is
bolted to the pier through clearance holes in the
base plate. The transducer is of the moving coil
type, \vith a coil of 5000 turns of 34-gage magnet
wire with total resistance 500 ohms. The magnet
is made of Alnico V alloy with pole pieces ar·
ranged to produce a radial magnetic field of 2000
gauss.
The brass boom is 1'*1 inches long and the
inertial mass is 15 lb. The pendulum period is
adjustable for stable operation in the range ]0
to 30 sec. Damping of the pendulum is accom­
plished by shunting the coil, approximately 3800
ohms being required for critical damping for
operation at a pendulum period of 30 sec. Sensi­
th·ity is adequate even when operated in a highly
overdamped condition. Space is available on the
base plate for installing an additional transducer,
for example, to drive a second galvanometer or,
after suitable amplification, an ink recorder.
All structural elements of the horizontal pendu­
lum have equal coefficients of thermal expansion,
the upright and boom being brass, and the suspen·
sion wire phosphor bronze. The boom length,
inertial mass, transducer, free period, damping,
and sensitivity ar~ identical with those of the
vertical.
The cover for each instrument is fabricated from
sheet aluminum with a plexiglass window. All
joints are sealed against draft with felt gaskets.
/\. Lehner galvanometer with free period 90
sec, coil resistance 500 ohms, critical damping,
current sensitivity 10-10 amperes per millimeter
106
FIG. 2 - The
deflection at one
ducer through a
provides for ,""r;~t;()I11
tivity.
A triple drum
perfect alignment
to facili tate
ing with pendulum
damping is shown
curves for other
The method of
Murphy and others
The peak magnificaLJ
of 25 sec. The
... periods less than
longer periods.
cation in the period
strument produces
formation previously
Performance-AI
A
Februa ry 1958
LO~G-PERIOD
SEISMOGRAPH SYSTE;\!
1 pendulum reased sensi­
iniormation These instruments
Jcna since 1956 with
, pendulum and gal­
amped. They will be
Hions during the In­
1-'1(;."
The t ripJc component rccordin!(
107
d!'UDl
To,,1
FlG.
1 _. The
vertical component scismometer
TOd
r.
Id 2 are photographs
ntal components, re­
~ on the vertical pen­
oyancy compensator.
ar for the vertical and
al components. It is
clearance holes in the
is of the moving coil
lrns of 34-gage magnet
500 ohms. The magnet
y with pole pieces ar­
magnetic field of 2000
1 inches long and the
he pendulum period is
ration in the range 10
Ie pendulum is accom­
oil, approximately 3800
critical damping for
period of 30 sec. Sen5i­
.en operated in a highly
ace is available on the
1 additional transducer,
!cond galvanometer or,
on, an ink recorder.
,f the horizontal pcndu­
i of thermal expansion,
~ brass, and the suspen­
~e. The boom length,
free period, damping,
cal with those of the
ment is fabricated from
llexiglass window. All
raft with felt gaskets.
with free period 90
lms, critical damping,
mperes per millimeter
Te .. GROUND PERIOD IN SECONDS
1'1(;
4
Calibration curves oi yarious seismograph
systems
research, seismogram,.: from this instrument have
been surpri5ingly useful in routine analysis by
proYiding clear recordings of later hody phases,
G wanes, and the impulsive initiation of surface
FIG. 2 - The horizontal compoll<.'nt seismomcter
deflection at one meter is cOllllected to the tran,·
ducer through a rcsistin: netl\'ork. The network
provides for \·ariation 0\ damping ~l1d of sensi·
tivity.
A triple drum recorder b used (Fig. J), providing
perfect alignment between the three seismogram;;;,
to facilitate comparisons of phases and determina·
tion of orbital motion.
The calibration curve of the int'trumen ts operat·
ing with pendulum and gah-anometcr at critical
damping is shown in Figurc 4. :\150 shown are
curves for other seismographs currently in usc.
The method of calibration was essentially tha t of
Murphy and others [1954J.
The peak magnification of 4200 occurs for periods
of 25 sec. The response diminishes rapidly for
periods less than 25 sec and more gradually for
longer periods. Owing to its superior magnifi­
cation in the perioo range 20-200 sec, this in­
strument produces seismograms containing in­
formation previously unavailable,
Pe~rorma1Zce-Although primarily
useo for
"-aYes.
The recording of mantle Rayleigh "aves from
the great Kamchatka earthquake of Xovember
4, 1952, made possible an extension of the disper­
sion curyes for these ,\·aves as ,,'ell as a determi­
nation of internal friction in the mantle [Ewing
mId Prrss, 1954).
Press and Ewillg \ 1955J found clear indications
of the new ph<1sc~ Pa and Sa on recordings of
these seisllIographs. Oli!'er and Ewing [1957J
reported on obscrv<1 tions of unusually long-period
1111croselS1115,
These example" show how the nc\\' instruments
have contributed to seismological research by
providing increased sensitivity in a period range
!lot co\-ercri hy other seismographs.
j ckrtmc!cdgmml- That part of this research con·
ductcd in Pasadena was supportcd hy a grant from
the Alfred' P. Sloan Foundation.
REl;'ERENCES
11.. Earthquake seismographs and associa1l6od
instruments, Admnces in Geophysics, 2, 219-275,
1955.
E\\,ji,n, ;\f., ,1:\11 F, PRESS, Further study of atmos­
BENIOFF,
1 pendulum
reased sensi­
information
Th ese instrumcnts
lena since 1956 with
pendulum and gal­
iffiped. They will be
.tions during the 1n­
fIr.. 3 - The triple component recording drum
FIG.
1 - The "ertical component seismometer
d 2 are photographs
ltal componen ts, rc­
on the vertical pen­
'yancy compensator.
.r for the vertical and
.I components. It is
:learance holes in the
is of the moving coil
'ns of 34-gage magnet
30 ohms. The magnet
with pole pieces ar­
magnetic field of 2000
inches long and the
e pendulum period is
Ition in the range 10
~ pendulum is accom­
ii, approximately 3800
critical damping for
,eriod of 30 sec. Sensi­
~n opera ted in a highly
.ce is available on the
additional tran sducer,
:ond galvanometer or,
n, an ink recorder.
the horizon tal pend uof thermal expansio n,
brass, and th e suspen­
:. The boom length,
ree period, damping,
II wi th those of the
len t is fabricated fr om
exig lass window. All
1ft with felt gaskets.
wi th free period 90
as, critical damping,
<peres per millimeter
To:2
0.\
F<G
-l ­ Calibration curves of various seismograph
systems
FIG.
2 - The horizontal component seismometer
deflection at one meter is connccted to Lhe tran :'­
ducer through a rcsisti\'e net \york. The network
prov ides ior variation of damping and of scns i­
tivity.
A triple drum reeon]cl i, u,ed (Fig. 3), providing
perfect alignment between the three seismograms,
to facili tate comparisons of rha,cs ,U1U detcrmin:J.­
tion of orbital motion ..
The calibration curve of Lhe: instruments operat­
ing with pendulum a.nd gah'anometer at critical
dampj~ is show n in Figure -le. j\l so shown are
curves for other seismographs curren tly in usc.
The method of calibration \\'as essen tially tha t of
Murphy and others [1954].
The peak magnification oi -le200 occurs ior periods
of 2S sec. The response diminishes rapidly for
periods less than 25 sec and more gradually for
lon~cr periods. Owing to its superior magnifi­
cation in the period range 20-200 sec, this In­
,trument produces .< cismograms containing in­
formation previously unavailable.
Pel:rormance-Although primarily used for
•
research, " ci~mogralll s from this instrument ha\'e
been surpri:.;i,ngly useful in rou tine analysis by
providint: clear recordings of later body phases,
G waves, and the impUlsive initia Lion of surface
waves.
The recording of mantle Rayleigh \I'aves from
the great Kamchatka earthquake of l'iovember
-le, 1952, madc possible an extension of the disper­
sion curves for these \\'av es as well as a determi­
nation of internal iri ction in the mantle [Ewing
a.lid Press, 195-le].
Prcss and Ewing [1955J fou nd clear indications
of the new pha ses Pa. and Sa on recordings o f
these seismographs. Oliver and Ewing [1957]
reporteci on obser\'a tions of unusually long-period
mlcroselsms.
These examples sho\\' bow the new instruments
have contributed to seismological research by
providing increased sens itivity in it period range
not covered bv other se ismographs.
.1 ckn<ho!e.dg lllcnl- That part of this research con­
ducted in Pasarlena was supported hy a grant from
the Alfred P. Sloan Foundation.
R EfEHENCES
H ., Earthquake seismographs and associated
instrum ents, A dvances in Geophysics, 2, 219-275,
1955.
EWING , ;\1. , ,\ ~D f. PRESS, further study of atmos­
BEN"IOH,
~
108
PRESS, EWING, AND LEHNER
pheric pressure fluctuations recorded on seismo­
graphs, Trans. Amer. Geoplzys. Union, 34, 95-100,
1953.
EWING, M., AND F. PRESS, 1\:Iantle Rayleigh waves
from the Kamchatka earthquake of November 4,
1952, Bul. Seis. Soc. Amer., 44, 471-479, 1954.
LA COSTE, L. J. B., JR., A new type long-period vertical
seismograph, Physics, 5, 178-180, 1934.
MURPHY, L. M., R. M. WILSON, L. R. BURGESS, AND
T. H. PEARCE. Response curves of an electromagnetic
seismograph by sine wave simulator method, But.
Seis. Soc. A mer., 44, 7-19, 1954.
OLIVER, J., AND M. EWING. Microseisms in the 11- to
18-second period range, Bztl. Seis. Soc. Amer., 47,
111-128, 1957.
PRESS, F., AND sf. EWING. Waves with Pn and Sn
velocity at great distances, Proc. Nat. A cad. Sci.,
41, 24-27, 1955.
Seismological Laboratory. Division of Geologica/. ScUnCtS,
California Institute of Technology, Pasadena, Cali­
fornia, (F. P. and F. L.) and Lamont Geological
Observatory, Colltlnbia University, Palisades, NC'J!
York (M. E.)
(Communicated manuscript received September 23,
1957; open for formal discussion until July 1, 1958.)
Vo!' 39. Xo. 1
Sho
Abstracl
high-frequ
by tides.!
transducel
mitted to
I ntrodltctioll-TI
paper was designe
whose frequency Ii
waves and those (
waves fall into tWI
noise, which is all
associated with
tsunami waves, wI
are associated wj
floor.
The existence 0
long been known
but these are not
particularly the 1
pier-mounted ga~
filter [Munk, 19u.
stalled at the
Oceanography;