Seismometry Adventure

Last Updated 12/28/01

Here is the story of a fun project in which we set out to build an amateur seismometry station and record the signals from distant earthquakes. Much useful related information can be obtained from the website of the Pacific Seismic Network (PSN).

The the complete project will include the following elements:

The Seismometer

The seismometer design is based on one by Pete Rowe, WA6WOA who was kind enough to share many details of his efforts with me. It consists of a pendulum that is substantially horizontal, such that the natural frequency of oscillation is on the order of 15 seconds or so. A coil of many turns of wire is mounted on the bob of the pendulum and is coupled to a strong magnet such that relative motion between the coil and the magnet create a voltage across the coil. The magnet is fixed to the frame of the seismometer and thus shakes as the earth moves. The coil on the other hand is very weakly coupled to the frame (as is evidenced by the long natural period of oscillation). Thus the coil largely stays still (in the axis along which it is free to move) while the coil shakes. It is the relative motion (velocity) between the magnet and coil that leads to a voltage signal at the terminals of the moving coil.


Here is a picture of the horizontal
pendulum in development showing
the supporting hinge in the foreground,
the golden colored boom,
and a lead weight for a bob.

The supporting hinge is special. It is a clever flexure-based hinge which uses some fine music wire to make a substantially frictionless hinge. The photo below shows the arrangement.


Detail of Hinge Flexure

The geometry of the hinge is such
that the two wires are both in tension
the pendulum bob tries to rotate
the hinge under gravity. The wire is
stiff enough, however, to support the load
vertically, but can bend with a predictable
spring constant as the bob swings on the
hinge. The spring constant is not a
negligible factor in determining the
period of oscillation....More about that

The equation for the voltage developed by the moving magnet in the coil is given by:

V = v (dB/dx)*N*A where:

V = voltage across the coil, v is relative velocity of coil and magnet,
dB/dx is the gradient of the magnetic field, N is the number of turns in
the coil, and A is the average area of a turn in the coil

The astute reader will recognize some approximations and simplifications in this expression, but the spirit of the interaction is mostly represented.

Some characteristics that lead to a good signal-to-noise ratio for the seismometer include:
Long period of oscillation, large number of turns of wire in the coil, a strong magnetic field gradient in the direction that the coil is free to move. The area of the coil should be big enough to catch all the available flux from the magnet, but not so big as to become a large antenna for noise pickup.

My magnet

The photo below shows my magnet. It is made out of some steel bar that has been cut and ground into pieces that bolt together to form a C-shaped flux return structure. To this are attached two very strong Neodymium-Iron-Boron permanent magnets that have diameters of about 3/4 inch and length about 1.25 inch.


The steel bar is 3/8 inch thick
and 2 inches wide. The short
ends of the C structure are about
2 3/8 inches long, the back of the
C is 3 1/4 inches long so that the
gap in the magnet is 3/4 inch.

The coil that I am using is shown below. I pressed it out of an old relay that had a 115 volt coil with a nice geometry. I have yet to determine the number of turns, but I plan to do that. I painted the coil with Corona Dope, a high voltage insulating enamel to stabilize the fuzzy insulation on the surface of the coil and to improve the electrical insulation.


Here you can see how the coil
size compares to the magnet
pole size and gap.


Seismometry Amplifier/Filter

The voltage output from the coil is connected to an amplifier/filter circuit. The purpose of this circuit is to provide a voltage gain within the bandwidth of about .05 to 10 Hz and very little gain outside that bandwidth. This improves the signal-to-noise ratio in that the interesting earthquake related signals are expected to lie within this bandwidth, and extraneous noise of higher and lower frequencies will be ignored. The circuit I am using is based on and very similar to that available for sale by Larry Cochrane.

The output of the amplifier/filter circuit is connected to the input of the Analog-to-Digital converter (ADC) card in the data acquisition computer.

Data acquisition computer and ADC sub-system

The computer that we are using is an old cast-off 486 system running an old version of Windows 3.1 or DOS depending on my mood. A Pentium would be nice, but it really doesn't matter for the data logger. The data acquisition card is a PcLabCard PC-711s which happens to be supported by the EMON program which is available for free off the web. I had a couple of these ADC cards in my junk pile from previous projects, so that set the choice of ADC and logging program. If I hadn't had those, I probably would have purchased and ADC card from Larry Cochrane and used his SDR program. My ADC card permits up to 16 inputs which can be digitized to 12 bit resolution (one part in 4096). This is plenty good to see earthquake signals. Lots of guys have done it with only 8 bit resolution.

Seismic data logging computer program

As I mentioned above, I am using EMON because it supports my ADC cards (thus saving a couple hundred dollars on the project). It can be downloaded from the software section of this PSN page.

Current Project Status as of 12/28/01:

The computer runs and I can wave the magnet in the coil and see wiggles going into the data logging program. This is with the coil connected directly to the input of the ADC card. I can log data and recall it from disk files using the various utility programs that I downloaded from the sites indicated above. The magnet is now mounted on the seismometer, and the boom is equipped with a paddle to hold the coil. I have mounted the whole pendulum assembly on a separate plate which has leveling screws for adjustment, and springs to keep the whole thing together.

Next I need to finish the mounting of the coil and its wiring, and I plan to attach some adjustable boom stops to facilitate transporting the seismometer. Then I need to finish mounting the whole thing in the windproof box, and finish the power supplies and amplifier electronics. At that point I will be ready to see how noisy my neighborhood is.

The photo at the left shows the current status of the seismometer as of 12/28/01

This project is a work in progress...please check in from time to time to see the developments.