Following Jim H.s comments about using a speaker as a seismometer,
I have also experimented with the idea by laying a car-stereo
speaker on its back (magnet side) and putting half a dozen 1/4"
lead shot balls (Sports Authority) in the cone so that it drops
about half its range of movement. It does make a nice VERY short
period sensor, since the rigidity of the cone support is pretty high.
The mechanical resonance of most speakers is in the range of 30
to 100 hz, depending on the size. Adding mass lowers the frequency,
but this is limited by the cone's range of motion or the strength
of the support in the case of a horizontal configuration where the
added mass has to be glued to the cone and makes it sag. I did not
estimate the actual output, but the 60db (x1000) seismic amplifier
produced a reasonable signal.

I also experimented with a 6" speaker by adding a small leaf spring
to raise the cone so that more mass could be added. The spring was
a tapered inverted V, with a fine wire from the point to a bracket
epoxied to the center of the cone, and pulled tight so the tapered
spring looks like a fishing pole with a giant fish on the line.
This looked promising as  a way to increase the added mass to lengthen
the period, but it never achieved anything near 1 second. Maybe
a 15" speaker might get there.

I considered adding a VRDT displacement transducer to a weighted
speaker to use feedback to extend the period, but never did. I
suspect that the compliance of the speaker suspension has a large
thermal coefficient. 

Another experiment I did to make a strong-motion seis or sorts was
to take a taught-band micro-ammeter and turn it on edge with "O" up
and add a small weight (fine solder) to the pointer to bring it down 
to mid range. With a 2000 ohm coil it has considerable output. The
mechanical period was about 0.5 seconds (2hz). I tried to fix the
core of a miniature LVDT (Schaevitz 005MHR) to the pointer, but gave
up on the alignment problem. A VRDT, with its open construction, 
would be better if a lightweight sensing vane could be used.
A horizontal configuration requires that the mechanical centering
be over-adjusted to move the pointer to center scale.
Unfortunately, a quality 4" taught-band meter costs over $100.
I expect that a jeweled movement would have too much stiction.

Regards,
Sean-Thomas

_____________________________________________________________________

Public Seismic Network Mailing List (PSN-L)