I read Bob Barns' 1996 article about seismograph calibration
and I wonder if there is not a much simpler method that does
not require any special equipment.  

Suppose you just take the damping off a Lehman pendulum and 
give it a little shove.  It will begin to oscillate and if you
know the period and the amplitude then you can calculate
the peak velocity in nm/s.  Then you read the peak value
from your ADC and that allows you to compute the
velocity sensitivity of your seismograph in nm/s per bit.
By observing your backgraound noise level in bits during normal
damped operation you can convert it to nm/s.   

Or am I making a stupid mistake?

And what is a good value for sensitivity and background noise level?
When I do the tests above, it looks like I have a sensitivity of
about 120nm/s/bit and my noise level is about +/- 3.  I have
a 16 bit ADC so should I crank up the analog gain a tad?

The reason I am trying to calibrate my Lehman is that I just
got it working and I can not figure why it detects some quakes
well and others poorly.  I am comparing my results with those
posted on seismicnet.  Especially the results from Herndon, VA
which is just a few miles away from me at Bailey's Xroads, VA.
I am trying to figure out whether I need more gain or is my location
on a office building basement slab next to a busy street just too noisy? 

TIA for any help,

Dave Saum

__________________________________________________________

Public Seismic Network Mailing List (PSN-L)