At 09:28 PM 10/17/99 -0500, Jim Hannon wrote:

>Ok Larry now I see where you are coming from.
>Only covering the strong motion range say +/- 2 g it is not hard to come
up with a
>lot of different ideas for sensor. The real challenge as I see it is in the
>details of the design. That is picking a design that is easy to
manufacture the
>parts or using a design that can be build from common parts such as the
STM-8.
>

Exactly! 

>The very first sensor I built consisted of a small weight glued to the
cones of
>two speakers. It easily detected footsteps across the basement on a concrete
>floor.
>

I like that. Maybe you are on to something here. If one could add some mass
and a position loop could a speaker be user as s SM detector? Was two
speakers used for X and Y motion or were they used together somehow?

>We should try and nail down a few specifications so that the discussion
don't get
>too far out in left field. How about +/- 2 g max acceleration with 16 bits of
>dynamic range. (How does this match the low end of a geophone
sensitivity?) What
>about frequency response? ( 0.01 to 10.0 Hz).

The SM sensor should have +- 2 g max acceleration, this should produce a +-
5 or +-10 volt output. It should be accurate to within +- 5% or 10%?? over
the operating temperature range of the device. I don't think the operating
temp  should be too extreme. If needed power can be used to either cool or
heat the device. 

I don't think super low power should be a big design criteria. Since the
data will most likely be collected using a PC, 110 v (or 220 v) should be
handy. A UPS should be added by the user so that the sensor and computer
will keep recording if the power goes down. 

The dynamic range should be 16 bits or better. If a sensor could be built
with very good dynamic range then the unit could have two outputs. One at
+- 2g and another more sensitive channel with maybe .2 or .02 g max output.
This maybe asking too much for one sensor? Maybe the same mechanical and
electronics parts, with a little tweaking, could be made to look like a
short-period, higher gain device that could be used to record the smaller
events?

I think the frequency response should be higher, in the 25 hz range. The
low end of .01 hz sounds good. The FBA-23 I have has a DC output. If I
rotate one of the horizontal channels vertically I get a constant 1 g
output, or about 5 volts DC. The sensor I have has a +- 2 g = +- 10 v
output. I high and low pass the signal out of the FBA-23 at .07 hz (~14
sec) and 20 hz.

Regards,
Larry Cochrane
Redwood City, PSN


_____________________________________________________________________

Public Seismic Network Mailing List (PSN-L)