PSN-L Email List Message

Subject: Pendulum Q
From: Larry Cochrane lcochrane@..............
Date: Thu, 18 Jan 2007 21:48:09 -0800

Hi everyone,

Dr. Randall Peters has ask me to forward the following to the list.


   I have noticed that many of you have great interest in detecting small teleseismic
earthquakes, but probably don't have a great deal of interest in carefully studying
the temporal features of what gets recorded.  Thus I have a recommendation for those 
of you who are using long-period instruments, such as the Lehman or a folded 
pendulum.  If your instrument's natural period is in the vicinity of about 20 s, then 
try the following simple experiment--remove all external damping from your 
instrument. In other words, take off the magnets, or remove the oil pot.   I have 
already learned to wear the label comfortably, but before you view me as a heretic 
(which some will do, since anything other than near-critical damping is viewed as 
ridiculous) consider the following.

   Your undamped pendulum's acceleration responsively at 20 s will be increased by the
value of Q.  My guess is that many of you will thus see at least a 10- to 50-fold
increase in your ability to then detect small, distant earthquakes.  No doubt the
objection to be raised by some to my recommendation will be--BUT the transient 
response of the pendulum will 'mess up' everything!   Yes, you will have immeasurably 
complicated things if your interest is traditional.  This traditional viewpoint is 
one that I believe will become for some observers increasingly antiquated and 
irrelevant.  In the days before computers, such a mode of operation was indeed 
ridiculous for those who wanted to see all the intricacies of arrival time features 
governed by phase, etc.  But now with the computer we have the ability to graph what 
is probably the single-most-important feature (to one with a single instrument as 
opposed to an array of them) of what the seismometer is telling us about the earth's 
motion--the POWER.  By using the computer to compensate for the transfer function of 
the pendulum, one can generate a reasonably good power spectral density no matter 
whether the instrument is damped or undamped.  I will provide details on this 
important calculation for those who are interested.

    What is the biggest problem with the traditional approach which I'm suggesting 
that some of you abandon?.  Consider an analogy from the electromagnetic case.  How 
does everybody detect radio (TV or countless other types of similar) waves?  The 
answer--by 'TUNING' their receiver with a relatively high Q.  Of course, if you want 
to see a whole lot of different regions of the E-M spectrum you must be able to 
adjust the bandpass of your receiver.  In the case of the earth, a whole lot of 
different natural period pendulums would be necessary to see everything of interest. 
   Think about the following:  If communication systems were to operate in the manner 
of seismology, we would just pick up everything in the electromagnetic spectrum with 
a broadband antenna and try to sort things out with sophisticated electronics!

      I have already demonstrated the viability of this method in the case of 
microseisms with periods around 4 s; c.f. the paper "Compound Pendulum to Monitor 
Hurricanes and Tropical Storms"  There's no
reason the same thing can't be done for the longer period region of teleseismic
earthquakes (around 20 s).  I predict you will be AMAZED at what starts to show up in
your records.

   I hope that some of you will try this surprisingly simple experiment and tell me of
your excitement after a few days watching the output from your seismograph.

Randall Peters


Public Seismic Network Mailing List (PSN-L)

[ Top ] [ Back ] [ Home Page ]