Jim,

I was just cleaning e-house and found your question of 21 Nov regarding
recording teleseismic or long-period (eg. 20 seconds) data from
a short-period seismometer, like at 1 second.

This has been explored somewhat on the PSN as members try to get
some interesting data out of cheap 4.5hz geophones. The answer is
that for a velocity sensing seismometer, it WILL respond to a wide
range of ground motion periods, with the maximum velocity response
(ie voltage output) at the mechanical period (or frequency) of the sensor.

The Riedesel paper gives a nice treatment of this.

"Limits of Sensitivity of Inertial Seismometers with
Velocity Transducers and Electronic Amplifiers";
by Mark A Riedesel, R.D.Moore, and J.A.Orcutt; Bulletin of
the Seismological Society of America, Vol. 80, No. 6,
December 1990.

also: 

"The Sensitivity and Dynamic Range of Inertial Seismographs";
by Ben S. Melton; Reviews of Geophysics and Space Physics, 
Vol 14, No. 1; February 1976; C. American Geophysical Union.

In theory, the ideal seismometer will respond to all periods, so
one only has to shape the passband filter to enhance the data of
interest by attenuating what is not wanted.  The major difficulty
here is that the velocity response decreases by the square of the 
period from either side of the natural period. For a 1-second L4-C,
the output at 1 second (for a special high output coil) might be
about 8000 volts/meter/second, but at 10 hz AND 10 seconds it is
only about 100 V/m/sec., and 10 V/m/sec at 33 seconds. So to get
good sensitivity at 30 seconds for distant earthquakes, the sensitivity
at 30 seconds needs to be increased by about 1000. This is impossible
to do without amplifying all the associated mechanical, thermal, and
amplifier noises, as Riedesel documents.


Unfortunately, a real seismometer or geophone has major limitations 
caused mostly by a small mass and thermo-mechanical and micro-positioning
suspension noise. Even a fedback system can't make these go away.
I've made a 4.5hz geophone operate quite reliably at 20 seconds 
by adding a displacement transducer and fedback electronics. But it
was not very useful for detecting moderate teleseisms since its
self noise level was about 100 times the expected signal. I have
discussed this previously and can repeat it if necessary. It does
make a reliable "strong motion" broadband sensor.

So this is why we have to build instruments that have an inherently
long natural period. Or at least a period long enough that it can
be extended with broadband feedback electronics.

Regards,
Sean-Thomas

_____________________________________________________________________

Public Seismic Network Mailing List (PSN-L)