Jim,

Regarding instability at long natural periods:
  
For your information and encouragement, I would like to point out
that when the WWNSS (Worldwide network of standardized seismograph
stations) was first deployed, the horizontal long period (LP) seismometers
(Sprengnether 5100s) were set at 30 seconds period. But after about a 
year, it became evident that this resulted in many against the stops
for much of the time, in spite of being an elegant instrument installed
on massive piers in carefully designed vaults (with notable exceptions).

So the standard period for the long-period WWNSS was set to 15 seconds.
Recording was done with a 90-second galvanometer on photographic paper.
The seis, galvo, and coupling/attenuation network impedances were 500 ohms.

So I would also suggest that strong over-damping coupled with a good
amplifier can achieve a similar result as a much longer period, since 
over-damping makes a somewhat broad response, but at reduced output, hence 
the need for a proper amplifier, especially one with very low input
offset current, since such currents WILL flow through the signal coil
of the seis and push the mass around. These alone may cause substantial
apparent drift of the mass position.

In two stations where I am currently operating LP seismometers (with
natural periods set to 15 seconds), I am damping them with 500 ohms 
(the same as the coil resistance), which is a damping coefficient of 2 times
critical. I am using the amplifier posted on the web site. With no
galvanometer, the response is fairly broadband from 10 hz to 25 seconds.

Regards,
Sean-Thomas

PS: You have not mentioned the thermal protection that you are providing
for your seismometer. Even the large 5100 LP seismometers have 4" to 8"
of insulation. I use an indoor/outdoor thermometer to monitor the instrument 
temperature, which changes no more than 1 degree C / week, and about 1/10th 
that per day.

PS2: I notice that you mention moving your boom 1" off center to measure
the period. This may result in major non-linearities in the restoring
forces that control the period. The period is rarely measured visually; 
when it is, displacements of less than a mm are used, and a magnifying
reticule is used to time the zero crossings going in the same direction
(a full cycle of motion represents the period). Preferably an oscilloscope
or the seismic digitizer is used to measure the period from the output
of the moving coil, using very small displacements. It may be that your
period is actually much longer than what you get by moving it 1".


_____________________________________________________________________

Public Seismic Network Mailing List (PSN-L)