Many of the seismometers in our network are of the magnet-coil-spring type. This type of instrument consists of a permanent magnet and a coil of wire. The coil, which is wound around a rather massive core, is suspended by a spring. When the ground moves, the coil tends to remain in place due to its mass, while the magnet, which is rigidly attached to the seismometer housing, moves relative to the coil. The relative motion produces a current in the coil, and it is this electrical signal that ends up being recorded as a seismogram.
The mechanical response of the magnet, springs and coil, as well as the electronics that amplifies the current, all affect the final signal. If the springs weaken or the electronics drifts, for example, the seismogram will not be accurate. Since these seismometers are located all over northern California, it is not practical to visit each one to check its performance.
Instead, the seismometer is programmed to check itself. Once a day, the electronics in the seismometer sends a controlled current through the coil. The response of the magnet-spring-coil system to this test signal is sent back as a calibration pulse. These pulses can be measured at the central recording site in Menlo Park, California, to assure that each seismometer is functioning properly.