November 1995

by Allan Coleman

Download this article and drawings (75k ZIP file).

The following article describes how it is possible to lengthen the natural period of a velocity type seismometer using electronic feedback. This article assumes that the reader has made, or has the knowledge to make, a horizontal long period (5-15 seconds) velocity type seismometer. A good article which describes to how make such an instrument for amateur use was published in the Scientific American, July 1979 edition, in the column titled "The Amateur Scientist". It detailed the fabrication of an instrument as well as how to record the ground motion it detects. The device can be adapted to that seismometer design.


The design of the prototype horizontal seismometer is similar to the one written about in the Scientific American magazine's article referenced above. The major differences are as follows:



The feedback coil is made on a PVC plastic former as shown in Figure 1. The coil was made from 38 AWG enameled magnet wire, weighing approximately 2 ozs, with a resistance of about 1500 ohms. All materials should be non-magnetic that hold the coil in place, otherwise the magnet will suck up to it. The feedback coil is connected to the pre-amp with the same wiring scheme that's used for the seismic coil, with some flexible fine wire adjacent to the pendulum pivot point. For whichever item is attached to the pendulum, either feedback magnet or coil, the cantilevered mass MUST be counter-balanced to prevent a rotational torque on the boom. See the set-up illustrated in Figure 1. It has a brass counter weight on the opposite side of the boom. Threaded rod allows for convenient adjustment.

The magnet (made by General Hardware Mfg. Co. Inc.), whose proportions can be taken from Figure 1, is rated with a holding force of 22 lb. and made of Alnico. To the magnet was bonded a mild steel pole piece to help shape the magnetic field within the coil. An air gap exists between the sides of the pole piece and tube bore. A 3/8" space between the end of the pole piece and end plug allows for ample movement of the pendulum.


The circuit may be broken down into 2 sections, the seismometer pre-amp (Figure 2) located adjacent to the seismometer, and the filter circuitry (Figure 3) located adjacent to the recorder. The seismic signal is fed into a difference amplifier (IC1) and then amplified 100x (IC2) before going to another signal amp (IC4). Depending on the type of recorder used it may be necessary to eliminate IC4 from the circuit and alter the resistor values as connected to pins 2 and 6 on IC2. To eliminate problems with high frequency noise the low pass filter (IC5) is used and the high pass filter (IC6) keeps signal drift in check.

Electronic feedback is achieved by monitoring the voltage generated by the seismic coil. The signal is cleaned up by a low pass filter and sent to a differentiator (IC3A). The differentiator is used to produce a signal that is proportional to the rate of change of the input signal. So with a constant DC level the output of the differentiator is zero. As the rate of change increases, the output voltage also increases. The input signal is differentiated for up to 100 seconds, the RC time constant. The differentiator's output is converted to a current (IC3B) sufficiently large enough to power the feedback coil and thereby slowing down the pendulum motion.


Figure 4 shows different pendulum responses as they were made on a strip chart recorder. NOTE: the signal amplitude is not supposed to be the same in the 3 examples, it's the period duration that we need to review. In Figure 4A the pendulum was free to swing with no damping at all. The natural period for the seismometer was originally adjusted to 10 seconds but was shortened to 6.5 seconds to show the possible effects of period lengthening on a smaller, simpler, designed instrument.

A shunt resistor (R1) of 1 K ohm was put across the seismic coil. The pendulum was tapped with a finger to observe how long it would take for it to come to a rest. Figure 4B shows that it took 18 seconds to return to its rest position with minimal overshoot. This test was repeated 3 times, as shown.

The feedback circuit was hooked up to the feedback coil and the results are presented in Figure 4C. The pendulum was clamped to a travel limit stop, about 1/2" off center, and released. The output signal seen at pin 6 on IC6 shows that the combination of damping resistor R1 and period adjustment resistor R2 presently selected gives the required damping with a 35 second period.


First build the circuits shown in Figure 2 and Figure 3. Lay out the boards where it is easy to swap out resistors R1 & R2. To achieve the desired results with the instrument that you apply this device to, some experimentation should be expected. For resistor R2, the value chosen for the prototype was 20K ohm. When this value increases, the period is lengthened because the feedback coil receives more current to hold back the pendulum. This is the equivalent of adding more mass to the pendulum. As the period lengthens (by increasing the resistance of R2) the shunt resistor R1 value needs to be decreased to compensate for the added inertia, thereby providing more damping. When R1 resistance is decreased, it robs the signal going to IC1 and IC3A.

When the electronics are powered up while connected to the coils and the pendulum starts to oscillate on its own, it indicates that the feedback coil is wired backwards. Simply reverse the polarity of the wires connected to the feedback coil.


If you want to try for extra long periods of 60 seconds or more it may be necessary to add a non-inverting signal amplifier into the circuit between the pick-off point at R1 (point A) and the 100 K resistor of the low pass filter. This amp will help boost the voltage signal going to the differentiator because R1 will become a very low value resistor, robbing signal strength from the seismic coil. The 10 Meg ohm resistor on IC3A should be increased to 15 Meg ohm to lengthen the differentiator time constant to 150 seconds. Either horizontal or verticaI velocity seismometers can have the period lengthened with electronic feedback. A long period vertical seismometer can be a real nightmare to operate due to the fact that they are very sensitive to barometric pressure changes. I have been informed that the effect is proportional to the square of the period, that is to say a 30 second vertical is 900 times more sensitive than a 1 second vertical. The pendulum "floats" in the air. Even building such an instrument out of extremely dense material will not help. The seismometer must be put in an air-tight, rigid wall enclosure. The long period horizontal seismometer is mostly effected by tilting. An enclosure for it may be best left vented to the atmosphere, otherwise the warping of the enclosure will cause the base to warp, tilting the instrument.

I hope that you achieve the same results as seen with the prototype seismometer. Please let me know of any problems that you find with this design so the original document can be corrected before any other copies are handed out. Thank you.

[ Back ] [ Home ]

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