return to home

Amateur Seismology

for about 5 years, starting around 1970 :-))

by Ed Thelen

Welcome to probably the world's most boring hobby
     * Amateur Seismology *
Well, at least you get to relax and watch nothing happen a lot ;-))

This paper presents ideas, and is not a how-to construction document

Modern (since the 1980s) commercial broad band seismometers are three dimension, force-balanced (use electro-dynamical feedback) and are smaller and much easier to install and operate. Examples are here, a leaflet (local copy).

A Seismology History from USGS, (local copy) including that in the early 1900s, "B.B. Galitzen develops the the first electromagnetic seismograph ..." widely used even today, and discussed in this paper.

Earthquake Information Bulletin, 1972 Jan/Feb

Table of Contents
     What style and direction sensitivity to build?
     Sensor Design Considerations
     Sensor Mounting and Adjustments
     Amplifier and/or Signal Converter
     Time Tagging

     Another type, using LED with LDR Sensor with Data Logging
     An amateur looks at professional needs and equipment
     Toward Sensor Linearization
     Another Amateur Machine, Lyle Denny, Jan 30, 2017

A "seismograph" can be thought of as having three parts:
1) The seismometer, which senses earth motion, usually producing a weak electrical output 2) An amplifier, which makes the weak electrical output much stronger
Or more recently, a signal converter which digitizes the seismometer output

3) A recorder, often this popular helicorder, which graphs the electrical output, which adds timing marks from a clock synchronized the radio time signals ( WWV ) accurate to about 0.1 seconds.
Or more recently, a computer which stores, displays, and analyzes the digitized seismometer output, which incorporates time signals from GPS accurate to 0.000001 second (actually better ;-)) .

With some persistence, skill, luck, and other useful things, you too can make a seismic recording
A one day ( a good day in the 1970s ) recording of several distant earthquakes. The tick marks are one minute apart, each horizontal line is one hour (some overlap in the picture). The recording started 7:16 Greenwich Mean Time, December 27, of some year ??? The paper is three sheets of continuous form sprocket feed printer paper, still available -

Our family moved to California - "Earthquake Country" - in 1970. Having read a number of "do-it-yourself-seismology" articles in "Scientific American" and hearing of a nice integrated circuit amplifier, I started building. Soon I had a seismomter, an amplifier (I was an electronics junkie), and a then a recording unit, all working well.
    OK, I was not too demanding ;-))

I told my Los Angeles cousin Rita that I was interested in seismology - Rita had gone to school with Tom Sanders of Palo Alto. Tom was interested in amateur seismology.

I contacted Tom, and found that he was a member of
     the "World Amateur Seismological Society" ( WASS )
and was the current treasurer and also editor/publisher of its quarterly journal.

Soon I was a member and contributing articles. I seemed to be the only member contributing articles. (There were about 180 "members" paying annual dues (mostly going into postage for the quarterly newsletter) but not saying or volunteering anything.)
Tom tired of publishing the WASS journal, and offered me the editorship - I seemed the only "victim".

After about two years of rather solitary work, as treasurer, author, editor, and publisher, I too tired of publishing the WASS journal- but could not find a victim - and closed WASS.

After a few years I "down sized" and gave the WASS notebooks of publications to Larry Cochrane of the Redwood City Public Seismic Network.

Preparing to write this article, I asked Larry for the WASS notebooks, but he too had "down sized", and they are no more :-((

I am a techie - encouraged by a gift of the monthly "Scientific American" magazine while in high school in 1946 - and a faithful reader of its column "Amateur Scientist" until about 1991 when the liberal arts types ( who didn't seem to know high school science ) took over the magazine, and the "Amateur Scientist" column was no longer supported.

One of many fascinating articles in "Amateur Scientist" was how to build a seismograph sensitive in a horizontal direction to very low frequency waves, like a wave that takes 10 seconds. The output signals needed amplification to be seen an/or recorded. At the time, vacuum tube amplifiers were in, transistors flaky and expensive. The project seemed too much bother.

We moved to earthquake country, lots of earthquake activity, in 1972, just about the same year a lovely little integrated transistor amplifier, the 741, became available, for about $ 2 instead of say $300 for the previous hot, power hungry, troublesome, vacuum tube direct current amplifiers.

All of the above, plus meeting Tom Sanders ( above ) got me started building ;-))

What style and direction sensitivity to build?
The drawing to the left, from the British Geological Survey. I had heard that making a long period vertical seismometer was is a major engineering challenge :-( , so lets see if we can build a longer period horizontal instrument sensitive down to 1/20 Hz, (wave per second).

I had studied the following "Amateur Scientist" articles in "Scientific American"

  1. April, 1952. Larkin's Sprengnether Seismometer
    describes a Wood-Anderson torsion seismometer - owned and operated by Harry H. Larkin, Jr.
  2. June, 1953. About an Ingenious Electronic Seismograph and ... see picture
  3. July, 1957. Concerning Simple and Ingenious Devices to Record the Waves Made By Earthquakes
    see picture 1, picture 2
  4. August, 1970. Devices for Listening to Sounds Both in Water and in the Solid Earth
    - w transistorized DC amplifier - 10 transistors -
    Note that C.L. Stong calls this type of seismometer "Galitzin" after its inventor.

The "garden gate" type horizontal seismometer ( the sensor of a seismograph system ) seemed promising, reasonably simple, and the "Scientific American" articles had shown examples.

For your background, you can view
and for that lovely little DC amplifier, the 741 ( not cutting edge any more ;-)

There is also the "style" of the sensor. One of the early seismometers used a mirror on a wire. The mirror would be an inertial mass twisted by the wire shaking with the earth. A beam of light bouncing from the mirror would darken moving photographic film to record the earthquake. This system was used by Richter to define the "Richter Scale".

Using a coil interacting with a magnetic field is widely used now, in part because of availability of convenient and stable electronic amplifiers. It is easy to get a signal from a coil near a moving magnet, * but * to get a reasonably linear calibrated output over a reasonable range of movement requires some engineering and care - well beyond the scope of this little paper.

Positioning one edge of your round coil of wire in the center of your magnetic field will hopefully at least get your polarity consistent with the motion, even if not very linear :-))

Sensor Design Considerations
OK, now we are into it - the goal - a home built "garden gate" type horizontal seismometer with as long a period as practical.

Basic Idea
Here is a nice, informative picture with text from (Local copy of "Building Your Own Simple Seismometer") which states that the effective top pivot will be about 2 mm from vertical over the effective bottom pivot for a 20 second period. As the restoring forces are very slight, maybe 0.5 percent of the effective arm mass, the pivots must work almost flawlessly ...

Seismometer Overview
The long dimension of the base is usually about 18 inches for convenience. Adjusting the Period Screw adjusts the period of the swing. The Leveling Screws help position the magnet in the coil assembly. At 15 second period, 1/32 of a turn makes a significant difference. The pivot details are not shown here, but are critical to successful performance. Another requirement is that all structures other than the pendulum have mechanical resonances damped and at a frequency well above the highest of interest.
    I made my frame from 1.5 inch copper tubing using a 4 way junction - easy, nice - the pendulum L shaped of 1 inch copper tubing (no diagonal bar).

This is not a construction article, but you will likely spend a lot of time thinking about and experimenting with pivots - knife edges sound nice for bottom pivots but the must be carefully aligned vertically. Points flatten, wires harden, and on and on ;-))

Sensing Coil - outputs the voltage to the amplifier -
Get a half pound of the finest enamel wire you can wind with your electric drill ( maybe # 38 ? ) onto a non-magnetic non-conducting bobbin. When winding feel the wire carefully as there may be a non-conductive knot connecting two wires in that pound spool of wire :-|

The magnets and magnet housing from a defunct microwave oven magnetron work well ;-))

Leveling and Period screws and nuts
You will spend significant time
     a) Leveling the seismometer to get the magnet and/or mass aligned
     b) Adjusting the period of the pendulum.

These adjustments interact somewhat, as the period gets longer the leveling gets touchier.

It is very convenient that:
     a) The adjustment knobs you glue onto the screw heads are large
     b) Use fine thread screws for easier fine adjustments
         - some professional machines used differential threads
     c) Use flat smooth material (window glass?) under the screw points.

Choice of moving magnet or moving coil ??
  1. Many people put the magnet on the moving boom as part of the mass, and place the coil on the fixed frame. I worried that moving iron near by - such as an automobile if the seismometer is in a garage that is used.
  2. I put the coil on the moving boom, and the magnet on the frame, hoping that a moving car near by would not cause as much false signal. But you are faced with passing a pair of very fine wires near the upper pivot.
  3. If you are not worried about moving magnetic material near your seismometer, choose option "A" above ;-))

Damping the Pendulum
A mass, suspended from a spring or pendulum, will tend to swing for a while after getting pushed. This is great for clocks but extremely undesirable for seismometers. All seismometers have some means of damping out an "impulse" quickly, but not too quickly -

There is a concept of "Critical Damping" that is interesting ( external ) study. To achieve approximately correct damping various methods are used.

  1. Some older seismometers had a little pan of (motor) oil on the frame, with a little vane positioned parallel to motion, in the oil. This would absorb the energy of the swinging pendulum.
  2. Another possibility is a vane of conductive material ( aluminum or copper ) in a magnetic field, maybe a separate magnet to prevent interaction with the sensing coil. Adjust the position of the conducting vane to achieve correct damping.
  3. A currently widely used damping system is to use the power generated by the sensing coil to be dissipated in resistance in the coil and a resistor in parallel with the sensing coil. Since different resistance is required for different conditions, and a the usual variable resistor can be noisy and unstable, I use a selection of fixed resistors and an aligator clip lead. The resistor assembly can be fastened to the seismometer frame for convenience and easy adjustment. 10 % Resistors in the series from 10 K to 200 K should be just fine :-)) standard values, comment, Or you can start with a 200 K potentiometer, "audio taper", for a start. See further discussion under adjustments.

Sensor Mounting and Adjustments
Seismometer Location
Most of us live in urban settings, some in apartments not even directly fixed to "Mother Earth", susceptible to all kinds of urban seismic noise - trucks hitting bumps, trees swaying, ...

As part of the fun of this hobby is to record events thousands of miles away, whose signals coming to you are very slight, local noise is a big problem. As an example of a serious seismometer setting, the University of California has one of its seismometers in an abandoned gold mine far from any mining activity. Should be low seismic noise there :-))

Also related to location is the seismometer should ideally placed on or in an enormous rock. Soil expands and contracts and tilts with changes in humidity, gets into your equipment, and is your enemy.

Most of the rest of us must compromise - I placed mine on the garage concrete slab in a corner. Oddly, the car movement did not disturb it noticeably, but when the weather changed from dry to rainy or rainy to dry ( California climate cycles ) the concrete corner tilted and the seismometer needed frequent ( weekly ) re-leveling. Also, there was a train that ran twice a day about 5 blocks away which I claimed I could "see" seismicly.

Another annoyance is the wave action on an ocean coast. This is reputed to generate the "microseisms" of about 1/7 Hz. I was about 10 miles from the coast, but during winter storms the "microseism" noise ruined most of the fun. I tried to make a 1/7 Hz "notch" filter, but that is another tale.

Seismometer Protection
Obviously your sensitive seismometer must be physically protect from children, curious adults, cows, hunters, ...

And weather, water, air movements, spiders, ants, mould, heat sources like amplifiers, ...

A tale - I had my seismometer mounted on concrete, protected by insecticide, under a foam box 1.5 inches thick. It was working well - *except* - that about 3 AM some mornings there would be LARGE slow swings of the pendulum recorded. First I figured some spider had odd hours, but no spider. Then I wondered if the air cooled outside the box there must be a temperature inversion in the box which moved to non-inversion - causing the large slow movement. I installed a 2 watt lamp inside the top of the box, and that problem went away. :-))

Seismometer Adjustment
This is the basic scheme, does not describe all possibilities, "an exercize for the student".
  1. Disconnect the seismometer from the amplifier
  2. Disconnect the seismometer's damping system
  3. Adjust a Leveling Screw to position the magnet where you want it in the coil area
  4. Adjust the Period Screw to make the period more desirable.
    Tap the pendulum with a little stick or straw to get it swinging
  5. Adjust a Leveling Screw to position the magnet where you want it in the coil area
  6. If the period is not as desired, go back to 4 above
  7. Connect the damping system
  8. Tap the pendulum with a little stick or straw to get it swinging
  9. If the seismometer is under-damped, reduce the resistance, go to 8 above
    Correctly damped is just a little overshoot -
  10. If the seismometer is over-damped, increase the resistance, go to 8 above
  11. Reconnect the seismometer to the amplifier, you are done :-))

Calibration of this device is highly problematical -
     a) If you are careful, and lucky, your sensor linearity might approach 50%
     b) This sensor is a velocity, not a position sensor -
          Trying to get a known velocity, say 1 cm/sec, by hand, is highly approximate ;-))
     c) Trying to read your digital or analog voltmeter with precision while you shakily ...
However, it is comforting to get even a wild guess of say cm/sec to volts.

     - with your handy voltmeter set to say 2 volts DC,
     - the probes connected across the coil and damping circuit,
try to move the magnet at 1 cm/sec and see what you get. In my 1970s pre-metric dayze, I usually got about 0.5 volts when moving the pendulum at 1/2 inch per second.

Enjoy ;-))

Amplifier and/or Signal Converter
This is where we moderns have it all over the ancients ( we had vacuum tubes from the 1920s through the 1960s , current semi-conductors are even better ;-))

The June 1953 Amateur Scientist a schematic of a vacuum tube amplifier with 30 uf coupling capacitors and 5 meg grid resistors - but you don't want to deal with that !!

I used the 741 op amp in the 1970s, great in its day, but by current standards had large offset (to be manually adjusted out), high temperature drift, and high input current.

I used a non-inverting logarithmic amplifier circuit similar to this. Values are not shown as this is not a construction article.

A Logarithmic Amplifier

Input vs. Output curve
This non-linearity permits recording regional as well as distant earthquakes without going off the recording surface or throwing ink. The disadvantage is difficult calibration. Oddly, the interpretation does not suffer much. The peaks are somewhat flattened, but the general form is shown.

OH - How could I forget the battle I lost !! -- I live about 20 miles from the Pacific coast, where the (sometimes large) waves break upon the shore. This results in noise in the 1/7 Hz region. When there was a big storm in the Pacific within 1000 miles, the 1/7 Hz noise - called "microseisms" - was very noticeable. :-((

I tried to make a notch filter that would reduce the amplitude of signals in the 1/5 to 1/10 Hz region. In the analog world this requires high ohmage resistors and "large" non-polarized capacitors - along with amplifiers of much higher input impedance than the 741 op amps. At the time, Radio Shack was selling counterfit FET input op amps (relabeled 741s) which gave me fits.
When I went digital I didn't even try to make a notch filter program.

Currently there are two good ways to get useful outputs from your seismometer:
     - for analog output, very nice chopper stabilized amplifiers with negligable offset and drift for $1
     - for digital output, very high resolution A to D converters, don't need amplifiers for $10
In either case, pay attention to the application notes, and also use twisted pair wires from the seismometer to the electronics to reduce noise and hum pickup. Shielded twisted pair is even better ;-)) Using a connector at each end eases handling and transportation problems.

I am assuming that you don't want to become an electronics bug complete with soldering iron, so lets just buy an already made board with components already installed and tested, with output to your PC for easy recording and possible post-processing the seismic activity you collect.

Larry Cochrane of the Redwood City Public Seismic Network has designed and sold ready made electronics for years. He also links to free PC software that operates this equipment and records and presents the seismic information.

A note on these wonderful high resolution Analog to Digital Converters - say 24 bit ADCs - A 24 bit ADC can (should ;-)) resolve its legal input values into over 16 million different digital values. It can produce useful digital output from a 5 microvolt wave - plenty good enough for almost anyone ;-))
And it outputs to USB to interface to any computer made in the past 10 years ;-))

I'm stressing digital output because the continual changing paper and other servicing of analog outputs is really a drag !!

Time Tagging
For anything more serious than amusing your neighbors with your squiggles, you should include time information into your data. In the 1970s I used a little synchronous timer motor that rotated once a minute, operating a switch. This I adjusted by listening to WWV on a cheap short wave radio. I also had an electric clock set to "Greenwich Mean Time" for convenience :-))

Now fancy GPS equipment, accurate to nanoseconds, is available for about $80 -

And speaking of time - since seismology is world wide folks like to use what used to be called Greenwich Mean Time ( GMT ). Time turns out to be rather complicated - in the 1970s, a friend gave me a 40 page document document describing time used in earth satellites :-|

There are at least UT1, UT2, and GPS times :-| And leap seconds ... Enjoy ;-))


Recording onto Paper
Not trying to tell you how to lead your life ;-))
     But given modern alternatives, you don't want to do this
for the following reasons:
- expensive - either to purchase or build ( I actually did - once - )
- must change paper after every recording day
- dealing with ink can be interesting :-((
- - - - and on a close earthquake, the pen can throw ink about :-((((
- driving a galvanometer movement is interesting
- - - - a combination of inductance and back EMF due to movement -
- - - - think of it as a current driven, not voltage driven, device -

Recording into a Computer
You have several choices:
1) put an application into your regular computer to read from the USB, record, ...
2) buy an old separate computer system, and put in the above "app"
3) buy an old computer, and share monitor, keyboard and mouse (saves space)
     using this works for me for another purpose, I don't seismology any more.
Larry Cochrane of the Redwood City Public Seismic Network links to free PC software that operates his equipment and records and presents the seismic information.

Another type, using LED with LDR Sensor with Data Logging - Jan 12, 2017
Lyle Denny of Michigan constructed a seismic system based on this web site.
Also discussed are:
a) Data Acquisition, input analog voltage value into your computer
b) Data Logging, showing and saving seismometer output on your computer.

An amateur looks at professional needs and equipment - Jan 13, 2017
I am thinking of adding a section
     "an amateur looks at professional needs and equipment"
discussing that an amateur is happy to see that his/her equipment has "captured" an "event" from far away ::

But the professional is interested in more:
     - the amplitude and frequency of the signals with some precision
     - the timing to seconds and milliseconds
Science is all about measurement.
     - how big, how fast, what is the energy and where did it come from, ...
     - this wave must have reflected off of something to arrive at this time, ...
Professional papers need facts to help support theories.

Missing from most (?all?) amateur reports is calibration data

  1. velocity per inch or volt or ..., from a magnetic detector or displacement from a photo detector
  2. calibration of the instrument, such as
    - natural period - of the swinging arm
    - damping of the swinging arm
  3. calibration of the electronics - frequency limits and quirks
    RC coupled circuits are especially suspect.
For the above reason, most long distance seismic instrumentation is effectively DC coupled - not through a capacitor --
(If you are looking for oil, with dynamite, your frequencies are higher and life is likely easier in this respect :-)) Of course you have other troubles --

Local earthquakes can easily sent these horizontal instruments we have been talking about off their physical scale.
Lets say that the linear range of our seismometer is one inch, and the local earth jumps 10 feet -
The boom of the instrument has been pushed into its stops, the instrument has probably slid and been knocked about a bit.
For this event, you need a "strong motion" machine -

Toward Sensor Linearization
In the early 1970s I was allowed to closely examine a Sprengnether long period horizontal seismograph. Other than solid beautiful constuction, I noticed the magnet and coil sensor. This roughly what I remember from 45 years ago. (I'm guessing at the polarity of the radial magnet structure.)

Guessing at the dimensions, I estimate 3 inches diameter, 5 inches long. I forget if the magnetic structure or the the wire wound solenoid was attached to the boom. I do remember the wire was a larger diameter than I expected - maybe #32 enameled wire.

Note that for most of the travel of the solenoid in and out of the magnetic structure, you can expect a rather linear change of magnetic flux per millimeter of movement. I imagine that the pole pieces could be tapered to improve linearity.

Most amateurs (certainly myself) did not try for this level of sensor linearity. If using magnets, we tend to use circular magnets and circular coils and use an alignment similar to the figure on the left. (click on the figures to enlarge)

The equation (not including fringing effects) for the magnetic flux change through the coils per millimeter of offset from this position is unpleasant, but definitely non-linear. If we calibrate carefully with the components in say this position, and the components shift due to changes in leveling, then the calibration is no longer accurate. This limits the usefulness of our data for serious seismic work, such as velocity or displacement estimates.

However, lets make the coil narrow, with straight sides. A person would guess that the elongated coil would provide quite good linearity (and relatively constant resistive damping) over quite a range of positions. :-))
(Next time ;-)

Another Amateur Machine from Lyle Denny, Jan 30, 2017
Mr. Thelen,
I thought you might like to see my home brew seismometer. So here is a series of photos. You can see the LED and photo cell arrangement. The graphic is of a 4.5 mag earthquake that occurred about 60 mile south(near Galesburg Mi) of my home on May 2, 2015. It was a classic wave form, you can see all the elements of the wave. The boom is suspended with a guitar string. I let the five pound mass hang on it for two weeks to get the stretch out of it. On the end of the boom I have a pan of mineral oil and a paddle for a damper. I have almost zero maintenance. Twice a year I take my laser level and check the boom for level and the offset of the pivot points to be sure I have 1mm offset. It is hard to get the offset as I use a ball bearing on a flat surface for the top pivot, but it works. I added 50 pounds of lead to the base to increase it's sensitivity and it did. The seismometer is in my old basement shower room away from drafts and other interferences, foot traffic. You can believe this or not but when someone walks around upstairs the house and foundation tip and this thing will detect it and when strong winds come along same thing will happen. I have been very satisfied with it's performance or the years.
Thank you for putting up with me as there isn't many people that have this boring hobby.


Lyle Denny
Saranac Michigan

Web page started Dec 10, 2012
updated Jan 30, 2017