Hi,
  I would reiterate the other replies in stating the importance of a
well insulated box.  All long period sensors are extremely sensitive to
air currents and tilting.  As a final touch, add a heater, say 10 watts,
to the inside of the top of the box.  This will stagnate the air in the
box and eliminate convection currents.  This can be as simple as a
resistor powered by a wall-wart.  A small lamp running at half its rated
voltage (for nearly infinite life) will also work.
  My rig works exactly like Conklin's--my basement floor bends when
there is foot traffic in the bedroom above.  This tilts the sensor. 
Since these signals don't look anything like a 'quake, just ignore them.
  I think that a good way to set the gain is to adjust it until you get
+/- 10-20 counts on the quietest days.  
  Larry's amp has a low-pass filter which has a cutoff frequency which
is much too high for good signal/noise performance with a Lehman.  I
suggest adding a 6 or 8 pole filter with a cutoff of about 0.1Hz.  This
will reduce the noise due to microseisms which peak at about 0.16 Hz. 
Filter design is well described in Lancaster's book, see  www.tinaja.com
Bob Barns
> Bryan & Regina Goss wrote:
> 
> Ok I got my seismograph built and I used Larry's A/D card and
> filter/amp, My question is how do I know if I have to much gain?
> It looked as if I did so I put 15k res at r29 this reduced noise but I
> don't know if it was to much. I can walk across my shop floor
> 30x 40x concrete slab and going to seismograph I see the trace go up
> about an inch then slowly go back to center when I walk away just the
> opposite, The trace seems to drift up and down a little guess this is
> normal??????
> I really need help
> 
> this is before the damper and cover
> http://www.tsixroads.com/~spardue/seismograph/seismograph.htm
>
__________________________________________________________

Public Seismic Network Mailing List (PSN-L)