Bob Hutt provided me with FORTRAN subroutines for computing the power spectral density (PSD) of a seismogram. I've created a program that will compute the PSD of a seismogram generated on the AS-1 seismic system. The program and files related to the following graph are included in this zip file: as1psdnew.zip.
In order to determine the electronic noise of the AS-1 system a sample of 22 minutes of data was processed during which time the magnet and boom were removed from the AS-1 sensor. The PSD of that sample is indicated above as "No Magnet."
The magnet was replaced and another sample was taken during a time interval with no earthquake signals. That is the "Noise" PSD above.
For comparison I've also plotted the standard low high noise model (NLNM) developed by the Albuquerque Seismological Laboratory (ASL).
Clearly the AS-1 system has a high level of system electronic noise outside the range of about 0.1 to 1 Hz. The increased noise between .01 and .1 Hz is clearly one factor that makes it difficult to see surface waves from all but the largest teleseisms.