Lehman Seismometer

From the July, 1979 Scientific American Magazine
"The Amateur Scientist"

How to build a simple seismograph to record earthquake waves at home

by Jearl Walker

Figure 1 (16k GIF)

This month I shall describe a seismograph built by James D. Lehman of James Madision University in Harrisonburg,VA. He has sent me his design for making the apparatus. The machine can be built quite easily and is sufficiently sensitive to record North American earthquakes of magnitude 4.8 or more on the Richter scale and earthquakes elsewhere of magnitude 6 or more. Lehman has obtained records of an underground test of a nuclear device in Nevada and a severe earthquake in Turkey.

The main objective in the design of any seismograph is to translate a slight movement of the earth into a larger motion that can be put on some kind of permanent record, which can then be analyzed. The waves from a distant earthquake do not all arrive at the same time, some take longer paths through the earth and some travel more slowly. Hence the record should be plotted as a funtion of time so that the arrival of the various components of the waves can be noted and timed. The difference in arrival times of the types of waves can be utilized to determine the distance of the earthquake from the instrument.

From a variety of possible seismograph designs Lehman has chosen a cantilever type coupled to a strip-chart recorder. Basically his seismograph consists of a weighted boom, a horseshoe magnet mounted on the boom and a coil of wire positioned between the poles of the magnet. When seismic waves pass through the room, the magnet and the coil move with respect to each other, thereby changing the magnetic flux through the coil. The voltage generated in the coil by the variation in flux is then amplified and recorded on the strip-chart recorder. A wiggle of the floor therefore ends up as a larger wiggle on the paper. Since the paper is moving through the recorder at a specified rate, the clock time and the duration of the wiggle can be determined from it.

The arm of the boom rests on a knife edge that is held against a frame. The other end of the boom is supported by a wire attached to the top of the frame. The two support points for the boom (the knife-edge and the point where the wire is attached to the frame) are not on a vertical line. Therefore when the boom is set in oscillation by a seismic wave, it will swing horizontally in a pendulumlike motion around its rest position. If such a motion is allowed, about all that the seismographic record will show is the natural swinging of the boom excited by passing seismic waves. The periods and arrival times of the individual waves will be lost.

To avoid such loss, Lehman has designed the seismograph to damp the boom's swing quickly. This damping is accomplished by causing a piece of aluminum mounted on the far end of the boom to swing through a magnetic field. The motion generates currents in the aluminum that in turn generate magnetic fields of their own. The magnetic fields induced by the motion oppose further motion of the boom, which soon stops swinging.

If the natural swing is stopped, what the seismograph records is the relative velocity between the boom and the floor as the boom and the pickup coil move with respect to each other because of their different inertial masses. The current created in the pickup coil is proportional to the velocity of this motion.

If there is no damping, the boom will be most sensitive to seismic waves that have the same period as the natural pendulum period of the boom. With damping it will be most sensitive to waves with periods about half the length of the boom's natural pendulum period. An instrument built to respond to a seismic wave period of about a second is called a short-period seismograph, and one that responds to periods of from 10 to 20 seconds is a long-period seismograph. Lehman adjusts his boom for a natural period of from 12 to 18 seconds, and so he has a long-period instrument.

The main frame of the seismograph consists of a base and two upright sections of pipe. The base is best made of metal for rigid support, but seasoned wood or thick pressed board could also serve, provided you guard against warping from the humidity in the room. Lehman has a metal base of about 40 by 25 centimeters, supported by three 3/8-inch machine nuts he has glued to it, two at the rear and one in the middle of the front. The three-point support of those nuts should provide a level base free of any rocking. Shims of thin metal can be put under any of the legs to tilt the base as desired when the seismograph is prepared for operation.

The boom consists of a steel rod 5/16-inch in diameter and 75 centimeters long. One end is ground to a knife-edge. The other end is threaded for about eight centimeters. At the threaded end a nut is screwed on as far as the threads go and then a piece of plumber's lead weighing five to six pounds is slipped over the end, followed by another nut. The two nuts are screwed against the lead, holding it in place on the rod.

The uprights of the seismograph are attached to the base by means of two 1/2-inch pipe flanges, which are bolted to the base so that their edges are about half and inch from its edge. Large wood screws or 1/4-inch machine bolts will serve to fasten the flanges to the base, but make certain that the fasteners do not interfere with the legs supporting the base. Counter- sinking the heads of the fasteners may be necessary.

Two lengths of 1/2-inch pipe, each 18 inches long and threaded at both ends, are screwed into the flanges. Crosspieces will run horizontally between these two uprights at two places: where the knife-edge of the boom rocks against the frame and where the wire supporting the other end of the boom is attached. The lower crosspiece is a metal bar bolted onto the upright pipes. The upper crosspiece is a 1/2-inch pipe that is 3 1/2 inches long and threaded at both ends. It is attached to the upright pipes by two 1/2-inch pipe elbows.

Midway across the higher crosspiece the wire supporting the boom must be attached. Lehman uses a nozzle from an oil burner as a pivot point for the wire. The standard type of nozzle has a diameter of 9/16-inch. Take out the inside filter because all that is needed is the tiny hole through which the wire is to run. Assemble the base, the flanges, the uprights and the upper crosspiece and determine where the wire from the boom should be attached. You will want it midway across the crosspiece and on the pipe so that when the rig is assembled, the wire will form an angle of from 30 to 40 degrees with the horizontal. Mark the proper location on the cross pipe and then drill a 9/16-inch hole in it. Glue the nozzle into the hole, run the wire through the nozzle and through the pipe and then anchor the wire on the other side of the pipe.

Lehman suggests that you use either No. 26 Nichrome wire or piano wire. One length of the wire will come down from the nozzle and another piece will come up from the far end of the boom,where it is attached through a hole drilled in the boom. The two ends are attached and tightened through a turnbuckle about halfway up.

The lower crosspiece is a flat length of metal bolted to the upright pipes about two inches above the base. To it is attached a piece of hard metal such as the head of a bolt ground flat. The knife-edge of the boom must rest on the hard metal plate with the knife-edge vertical. When the seismograph is finally assembled, the nozzle on the upper cross- piece should be about a centimeter forward of the vertical from the knife- edge plate. An adjustable metal plate would be helpful here.

Attach to the boom a horseshoe alnico magnet. One can be obtained from the Edmund Scientific Company (6982 Edscorp Building, Barrington, N.J. 08007). Lehman has a magnet with a 25-pound pull. The magnet should have a gap of about 7/8-inch. In the gap will rest the pickup coil, from which signals about the relative motion of the magnet and the coil are fed to the electronic equipment. Lehman uses a rod clamp and pieces of wood to attach the magnet to the boom near the piece of lead.

The pickup coil is made by winding 10,000 or more turns of scramble- wound No. 34 magnet wire on a coil form. If you have an Edmund magnet (No. 41949), the coil form should be 3/4-inch wide and 3-inches in diameter, so that the pickup coil can fit from a 1/4-inch to a 1/2- inch into the gap of the magnet with suffcient clearance on the sides. Lehman made his coil form by gluing two circular pieces of 1/16-inch plastic to a wood cylinder 1-inch in diameter. The form was mounted on a drill press and then the wire was carefully wound onto the form will the press was run a low speed. (Be careful about running the press: it is not difficult to get your fingers caught in the wire.) Lehman brought the ends of the coil out to the sides of the form and then ran lamp cord from the ends to the amplifier. The cord should not be much longer than 10 feet: otherwise it will have to much electrical resistance.

The coil is mounted on a stable base that has the same type of three- leg support as the main base. Be sure to avoid installing any material that can become magnetized. For the base you could use a piece of aluminum with brass hardware to fasten the coil in place.

At the far end of the boom you will need another magnetic field in order to damp the swing of the boom. Lehman mounts two of the Edmund magnets on a wood base, orienting them so that they attract each other. He screws a wood dowel on the threaded end of the boom. On the outer end of the dowel he hangs a flat piece of aluminum (3-inches on a side) that is either glued to the wood or held in place by a brass wood screw. The idea is to have the aluminum hang down between the two facing magnets so that when the boom swings, the aluminum swings through their magnetic field. The eddy currents created in the aluminum set up magnetic fields of their own that oppose the motion and thereby damp out the swing of the boom.

The entire seismograph is covered to protect it from air currents that arise either from convection in the room or from the air-conditioning system. The top of the covering should be glass or clear plastic so that you can inspect the apparatus.

The voltage on the pickup coil is read through an amplifier. You could buy a commercial direct-current amplifier or build the one shown in the illustration. Lehman says that his circuit is normally operated to multiply the signal by a factor of 10. The circuit is mounted on fiber- board and placed inside a small box. The 60 Hz noise from the wiring of the room is filtered out in the input stage of the amplifier, as is any mechanical noise in the frequency range of 10 Hz or more. The seismic signals fall in the range from 1 to 1/15 Hz and therefore pass through the filter. The first 741 gate acts as a voltage follower for the voltage on the pickup coil. The second 741 gate acts as a hundredfold amplifier. Lehman says that two six-volt batteries of the lantern type can power the amplifier for several months.

The output of the amplifier has been designed to match the 10-millivolt level of Lehman's strip-chart recorder (a Heath IR 18M model). He says that the time base of the paper flow through that model is so good there was no need for an external clock system to put time marks on the paper. Felt-tip pens were found satisfactory for the inking. The ink cartridges could be reused by employing a medicine dropper to refill them with ink. One day's run typically consumed about two-thirds of a cartridge.

If you want to buy another type of graphic recorder. Lehman emphasizes the importance of finding one with independent controls for the paper flow and the coordination of the pen and the paper. The penholder should be universally jointed, and the chart should be easily removable for rewinding and other adjustments.

Lehman obtained standard 10-inch chart paper for his machine, running it through at the rate of a foot (30 centimeters) per hour so that the time spacings on the paper were a convenient one inch per five minutes (.5 centimeter per minute). By running the paper through on both sides, Lehman cuts the cost of paper in half. He estimtes that a full year of daily recording needs only $10 worth of paper.

To set up the seismograph, detach the boom and its wire support and then level the base by placing metal washers or shims as needed under the legs. A bubble level is helpful in this operation. The nozzle opening on the upper crosspiece should be about a centimeter off the verticl from the place where the knife-edge will eventually rest on the lower crosspiece. To achieve this arrangement you may have to change the lower crosspiece or put shims under the two rear legs on the base.

Add the boom and attach its wire support to the turnbuckle so that the boom is horizontal and the knife-edge is vertical on its plate. Give the boom a slight push to check its period of swing, which should be in the range from 4 to 8 seconds. Also check its rest position: it should be horizontal and perpendicular to the metal plate on which the knife-edge rests. At the other end of the boom you may want to add stops about a 1/4-inch from the rest position of the pendulum on each side to prevent the pendulum from swinging widly as you make adjustments to the seismo- graph. With the stops in place the rest position should, of course, be midway between them.

Leave the seismograph idle for several days so that the framework can adjust to the stress and the wire can stretch. Then adjust the boom so that its natural swing period is from 12 to 18 seconds. To increase the swing period to this value put a bushing either under the metal plate on which the knife-edge rests or under the front leg of the base. With such a step you reduce the distance by which the nozzle is off the vertical from the knife-edge point. If you were to tilt the assembly back far enough for the nozzle to be exactly on the vertical from the knife-edge, the boom would no longer swing like a normal pendulum because there would be no restoring force tending to return it to its rest position. With the assembly in proper alignment the nozzle point is off the vertical from the knife-edge, and the boom constantly tends to return to its rest position because of its own weight. The more the nozzle point is off the vertical from the knife-edge, the greater the restoring force is, and the smaller the natural swing period of the boom will be. You should adjust the alignment of the nozzle and the knife-edge so that the natural period is from 12 to 18 seconds. The boom will thus be sensitive to seismic waves of slightly shorter period.

The next task is to adjust the damping. First check it by placing the damping magnets in their proper position around the vertical aluminum plate and then swinging the end of the boom over a distance of about a centimeter. When you release the end, it will swing past the rest position. With proper damping the overshoot should be about two milli- meters. You can adjust the amount of damping by adjusting the gap be- tween the magnets. Closing the gap increases the magnetic field and the damping. You could also adjust the damping by varying the natural swing period of the boom. For example, a longer period would mean less restoring force, so that (with the same magnetic field) the damping would increase.

These steps constitute a rough tuning of the damping. Now position the pickup coil with respect to the magnet attached to the boom so that they are both at the proper height from the floor and so that the coil is about a 1/4 inch inside the gap of the magnet when the boom is in its rest position. Then cover the entire assembly and adjust the gain on the amplifier until you can see background noise recorded on the strip-chart recorder. Lehman says that if you then approach the seismograph, the tilting of the floor by your weight should give you a readout on the strip-chart recorder with an amplitude of about three inches.

To check the damping, walk up to the seismograph, wait a few seconds and then walk away. The paper on the recorder should show a pair of peaks for your advance and a pair for your retreat. The first peak in each pair is due to the tilting of the floor as you advance. The second peak, which is smaller, represents the overshoot of the boom past its rest position. If the boom is properly damped, the ratio of the two peaks should be between 6:1 and 10:1. If it is not, you will have to adjust the damping. Lehman points out that this may take some patience until you finally get the damping just right, but once all the adjustments are made the instrument will not need further attention for weeks or even months.

The best place for your seismograph would be in a room undisturbed by either people or thermal changes. Do not put the instrument where sunlight will fall on it or the floor, because the resulting variations in the stress of the floor will show up in your readings. Even some unexposed floors may show thermal variations because of sunlight striking the building. Lehman has had good success with one of his instruments on a gravel floor in an unfinished part of his campus. Since the seismograph is most sensitive to waves passing perpendicularly to the boom and since most earthquakes occur on an east-west azimuth, you should set up your seismograph with the boom running north and south.

When you first begin to record, your will notice a constant background noise in the readings. Much of it is from microseisms, which are small oscillations of the earth that are still largely unexplained. Invest- igators have long tried to link microseisms to natural phenomena such as the beating of surf on a cliff-lined shore or the presence of a cyclone over the ocean. Although at times the connection between microseisms and such a cause has seemed clear, at other times the microseisms have seemed totally unrelated to the proposed cause.

In some of the data Lehman has sent me, one can identify several other kinds of noise in the records. One day a brief set of spikes revealed a blast at a local quarry or construction site. A winter cold front appeared on another record as a set of continuous but random spikes. On at least two other occasions Lehman has been able to display the effects of hurricanes in the Atlantic.

You will probably be less interested in these kinds of waves in your recordings than in seismic waves. Such waves are of four main types (and several other less important types and subclassifications I shall not mention). Two types propagate through the body of the earth but in different ways and at different speeds. One of them is the transverse wave designated S. The material part- icipating in the wave motion oscillates perpendicularly to the direction of travel of the wave. The term transverse is meant to convey that sense of oscillation. Most of the waves with which you are familiar are transverse waves, for example waves on a guitar string or on the surface of a body of water.

The other type of wave tht travels through the earth is the longitudinal wave designated P. Here the material participating in the wave motion oscillates back and forth in the direction in which the wave is traveling. The term longitudinal is meant to convey the fact that the oscillations are parallel to the wave direction.

The remaining two general types of seismic waves propagate along the surface of the earth. With one type, the Love wave (L), the material participating in the wave motion oscillates transversely to the direction of travel of the wave and horizontally with respect to the earths's surface. With the other type, called Rayleigh waves (R), the material moves in elliptical paths perpendicular to the surface and aligned along the direction of travel of the wave.

If all these waves arrived at your seismograph at the same time, you would not be able to distinquish them, but they arrive at different times because they either take different routes to your instrument or travel at different speeds. The first to arrive are the P waves because they both take a direct route through the earth and travel faster than the S waves. Next comes the S waves, followed by the two types of surface waves, which are delayed by traveling over a longer path along the surface of the earth.

The amount of time between the arrival of the P and the S waves can be used to determine the distance between you and the center of the earthquakes. Tables relating the difference in arrival times and the corresponding dis- tance are available in seismograph references such as the book by Markus Bath cited in the bibliography for this issue. The distance is often quoted in terms of the angular distance between the focus of the earthquake and the seismograph as measured from the center of the earth. For example, if the difference in arrival times is about 7 minutes 11 seconds, the focus is about 50 degrees away from you.

With a single instrument you can only guess at the direction. Rough determina- tion of the direction can be made with two instruments oriented perpendicularly to each other (one seismograph on a north-south axis, the other east-west). Precise determinations of the distance, direction and depth of the focus are obtained only when several widely separated seismograph stations record the event. The difference in arrival times of the waves at the stations provides the clue to where the focus is.

Several periodicals can help you to determine the locations of major earth- quakes. 'Preliminary Determination of Epicenters, and Earthquake Information Bulletin' are both available from the Superintendent of Documents, Government Printing Office, Washington, D.C. 20402. 'Earth Science Event Reports' can be obtained from the Centr for Shortlived Phenomena, 195 Alewife Brook Parkway, Cambridge, Mass. 02138. Another valuable reference is 'Principles Underlying the Interpretation of Seismographs. Special Publication No. 254 (revised in 1966), which can be obtained from the U.S. Coast and Geodetic Survey, Rockville Md. 20852. Other references are in the bibliography for this issue. You should also have a globe in order to measure the distance to earthquake epicenters. Coordinated Universal Time (UTC) can be got by radio on stations WWV & WWVH.

Background noise and earthquakes are not the only things you may record. Underground nuclear explosions also send seismic waves through the earth and along its surface. Lehman has recorded a number of such explosions that were set off in Nevada, seven in the first three months of 1976 alone.

Lehman sent me several examples of earthquake data, two of which I find parti- cularly striking. One is a record of the earthquake that occured in the Ga- lapagos on March 29, 1976, at 05:39:33 UTC. The quake had a shallow focus, was of magnitude 6.7 and apparently resulted from slippage along the boundary between the Cocos and the Nazca tectonic plates. Six minutes 57 seconds after the quake began the P waves reached Lehman's recorder. They had periods of from one second to three seconds and were recorded for more than five minutes. It took the S waves 12 minutes 28 seconds to arrive, since they had longer periods and greater amplitudes than the P waves. The surface waves look about 15 minutes to arrive and had periods of from 12 to 15 seconds. Oscillations continued for about an hour. The difference in the arrival times of the P and S waves (5 minutes 31 seconds) indicated that the epicenter was about 2,440 miles (3.930 kilometers) from Lehman's apparatus.

In another example Lehman recorded the waves from a devastating earthquake in Turkey on November 24, 1976 at 12:30 UTC. It was the worst quake in the Anatolian fault zone since 1939. Lehman recorded the first P wave at 12:35:10 UTC and the first S wave at 12:45:30. Activity was still visible an hour and a half after the first P wave arrived.

Lehman is prepared to answer questions on the construction or operation of the seismograph. He is particularly interested in exchanging seismic records. Lehmans address is Physics Department, James Madison University, Harrisonburg, VA. 22807.


[ Back ] [ Home ]

Larry Cochrane - www.seismicnet.com/contact.html