If some routine check (such as an impulse or period test) has indicated that the instrumental response has drifted outside acceptable limits, it will be necessary to recalibrate, and we must consider the possibility of restoring the equipment to its original condition. The most likely situation is that the natural period of either the seismometer or the galvanometer has altered, or that a bad connection has changed the resistance of part of the circuit. We therefore proceed as follows:

Check as in Section 4.3.2 or 4.4. If too short, look for dirt on the coil, or spider's web between coil and boom. If these cannot be found, the cause probably resides in a change of level or in spring adjustment. In horizontal instruments the period is increased by tilting the instrument backwards so as to raise the centre of gravity of the boom. In vertical instruments, period is lengthened by adjusting the spring system.

Excessively short values are probably due to the coil touching the magnet. Long or short values may be caused by dirt on the coil or in the gap, or by excessive dryness causing the accumulation of static electricity. If the desired value cannot be restored by attention to these matters the only possibilities which remain are to replace the galvanometer, or to correct for the effects of moderate drift by adjusting the seismometer period and the galvanometer and seismometer damping. If the effect of the galvanometer drift has been to make the band-width narrower, the resistances in the coupling network should be reduced, so as to increase the damping. In instruments for which the recommended periods of the seismometer and galvanometer are close together, the peak of the sensitivity curve can be maintained at the proper period by altering the seismometer period in the reverse direction to the drift of the galvanometer.

If either of the main elements seems heavily underdamped, there is probably a faulty contact in the wiring, or a break in the coil itself. If over-damped, there is probably a short-circuit in the wiring or in one of the coils.

The parameters which are readily adjustable in an electromagnetic seismograph are the period of the seismometer, the resistors in the coupling network, which jointly determine the overall magnification, and the damping constants of the seismometer and galvanometer. The galvanometer period and the dynamic constants of the galvanometer and seismometer are not usually at the disposal of the operator. Consistency of performance of manufactured systems therefore tends to depend on the fit of the galvanometer period to the specified value, and on the ability of the system as a whole to deliver sufficient overall magnification to enable the coupling network to be adjusted for the damping constants of the seismometer and the galvanometer, and for the necessary overall magnification. If adjustable resistors are used the procedure of trial and error can be adopted whereby the resistors `R` and `r` are reduced if it is desired to increase the damping constant of the seismometer and galvanometer respectively. The shunt resistance `S` is increased to increase the overall gain, but will interact on the damping constants when it is changed.

If the galvanometer period does not coincide with the specified value, the specified degree of galvanometer damping will give an inaccurate bandpass characteristic. A compromise solution for moderate degrees of mismatch can be achieved if somewhat excessive values of galvanometer damping are used when the galvanometer period is too close to the designed period of the seismometer, or if reduced galvanometer damping is used when the period ratio is excessive.

The complete theory of the electromagnetic seismograph system (see Grenet and Coulomb, 1935, or Willmore, 1961) enables considerable variations in response to be coped with in this way, especially for large values of the coupling constant a and in circumstances where the galvanometer period is very close to that of the seismometer in the sample system.

It is possible (Coulomb and Grenet 1935) to set up a 'false Galitzin' combination having exactly the theoretical characteristics of a Galitzin with a period of `T` seconds, by using a pendulum of period `T`/`a` and a galvanometer of period `aT`, where `a` must satisfy the relation

Thus, it is possible to use a pendulum of 6 seconds and a galvanometer of 24 seconds period to obtain the theoretical characteristics of a Galitzin apparatus of 12 seconds period, but the coupling coefficient must be adjusted to a particular value and it is not possible to adjust the magnification by modifying the coupling. To obtain a different magnification, it is necessary to change the optical level or the mass of the pendulum.

The conditions which have to be satisfied become much less rigorous if one is content with an approximate match in place of a mathematically exact one, and if the value of `a` is not very different from unity. Thus, if one was aiming to simulate a standard 15-second Galitzin and had a 17-second galvanometer, one could obtain quite a close match by setting the seismometer period to 13.25 s and adjusting the shunt to give optimum damping.

On the basis of the analysis of seismic noise spectra, four standard amplitude characteristics (Fig. 5.3) of short-period seismographs were proposed (Aranovich and others, 1968a) for Soviet and associated networks. They are defined for negligible coupling coefficient (`sigma`^{2} = 0) by the constants of seismometer `T _{s}`,

The procedure is in two stages, as follows:

The general formula for the magnification `V` (displacement sensitivity) can be written in the form

where the normal magnification `V-overbar` and the amplitude characteristic `U-overbar` are

The level of magnification is regulated by setting up the coupling coefficient in the range 0 < `sigma`^{2} < 1. To preserve the seismograph response for non-negligible magnitude of coupling coefficient, it is necessary to change the period and damping constant of the seismometer and galvanometer, respectively (Fig. 5.3.1a-d). The constants `C _{i}` can be calculated by polynomials of the form

The coefficients `C _{i}` in Table 5.3.1a meet the following criteria:

Deviations of the approximate value of constant from the exact value must not exceed 1%.

The amplitude response for the approximate constants must not differ by more than 1% over the whole interval of periods 0 <

`T`< infinity from the standard amplitude response [3] defined by the constants in Table 5.3 (Tobyas, 1974).

For the determination of the coupling coefficient for the required maximum magnification `V _{m}` and a particular seismometer and galvanometer, a parameter

The values of maximum standard amplitude response `U-overbar _{m}` with corresponding period

The relative error of this approximation using coefficients given in Table 5.3.1b is smaller than 1%. The required magnification can be achieved only if `w` is smaller than `w _{m}` given in Table 5.3, which corresponds to theoretical maximum coupling

To adjust necessary damping constants of the seismometer and galvanometer and the coupling between these systems, the resistances `R` and `S` (see INST 3.2_generalized coupling network) of the signal circuit will be (Aranovich and others 1968b)

and the total resistance `R _{d}` of the second coil used for additional electromagnetic damping of seismometer is

`a`_{s} is the critical resistance of the signal coil (i.e. the total circuit resistance for which the electromagnetic damping constant is equal to 1), `a _{d}` is the critical resistance of the damping coil and

The characteristic A IV with magnification 50 000 can be achieved because both periods `T _{s}` and

The SKM-3 seismometers with GK-VIIM galvanometers enable one to use the above-mentioned standard characteristics up to maximum magnification 100 000. Other types of seismographs are applicable for this purpose if the following conditions are fulfilled:

`w`<`w`_{m}`T`and_{s}`T`are adjustable to required values_{g}`S`> 0`R`,`R`and_{d}`r`are not smaller than the internal resistance of signal, damping and galvanometer coil, respectively.

After adjustment of constants it is advisable to check the function of the seismograph. The calibration constants for the determination of absolute magnification are given in Tobyas (1975).

Date created: 1/7/97 Last modified: 9/9/97 Copyright © 1997, Global Seismological Services Maintained by: Eric Bergman bergman@seismo.com