In a message dated 13/09/02, meredithlamb@............. writes:

> I would suggest "S-G's" (Shackelford-Gunderson) type seismo's as being 
> perhaps the easiest to build.  They simply are pendulums that hang 
> downward, and have used a variety of sensors that I've noted. These are 
> horizontal sensitive instruments.  About the most important part 
> mechanically seems to have been just the 0.001" thick X ~1.5" length brass 
> hinge/s I used; it seemed to need the flexibility there for longer period 
> (electronic capacitive) response to distant quake phases with periods 
> ranging up to > ~30 seconds.
> 
> I've little electronics experience per'se, but the Hall sensor type 
> designed by my (deceased) brother Robert back in ~1997, seems to be another 
> design that any interested partys might wish to review.... if for nothing 
> more than the basic simpicity of the mechanical design....
> Being as the Hall S-G is a stable gravity related instrument, I've seen 
> that after construction and a reasonable period of mechanical settling in 
> time, that its been much more dependable than any of my Sprengnethers and 
> their "hanging gate" pendulums for temperature and even moreso for ground 
> water related tilt effects tolerance.
> 
> http://www.geocities.com/meredithlamb/index.html

       The resolution and the linearity that you can get from a $1.83 
A3515LUA sensor from http://www.newark.com/ and a $2 type #2 NdBFe four pole 
magnet from www.wondermagnet.com are really quite good.

     The minimum movement which can be measured depends on the sensitivity 
and on the noise level present on the output signal. With a three pole 10 Hz 
Butterworth filter and the NdBFe magnets set for a sensitivity of 5.8 mV / 
micron, an A3515 sensor gave a bit under +/-20 nano metres resolution and the 
linear range of 0.8 mm. If I increased the sensitivity to 10 mV / micron by 
adjusting the magnetic field, I got down to about +/-10 nano metres 
resolution, but I had only ~0.5 mm total linear range. This is still OK for a 
pendulum sensor.

     See http://mariottim.interfree.it/doc11_e.htm for a typical calibration 
graph and some further notes on these sensors.

     Regards,

     Chris Chapman
In a message dated 13/09/02, meredithlamb@............. writes:



I would suggest "S-G's" (Shackelford-Gunderson) type seismo's as being perhaps the easiest to build.  They simply are pendulums that hang downward, and have used a variety of sensors that I've noted. These are horizontal sensitive instruments.  About the most important part mechanically seems to have been just the 0.001" thick X ~1.5" length brass hinge/s I used; it seemed to need the flexibility there for longer period (electronic capacitive) response to distant quake phases with periods ranging up to > ~30 seconds.



I've little electronics experience per'se, but the Hall sensor type designed by my (deceased) brother Robert back in ~1997, seems to be another design that any interested partys might wish to review.... if for nothing more than the basic simpicity of the mechanical design....

Being as the Hall S-G is a stable gravity related instrument, I've seen that after construction and a reasonable period of mechanical settling in time, that its been much more dependable than any of my Sprengnethers and their "hanging gate" pendulums for temperature and even moreso for ground water related tilt effects tolerance.



http://www.geocities.com/meredithlamb/index.html



       The resolution and the linearity that you can get from a $1.83 A3515LUA sensor from http://www.newark.com/ and a $2 type #2 NdBFe four pole magnet from www.wondermagnet.com are really quite good.



     The minimum movement which can be measured depends on the sensitivity and on the noise level present on the output signal. With a three pole 10 Hz Butterworth filter and the NdBFe magnets set for a sensitivity of 5.8 mV / micron, an A3515 sensor gave a bit under +/-20 nano metres resolution and the linear range of 0.8 mm. If I increased the sensitivity to 10 mV / micron by adjusting the magnetic field, I got down to about +/-10 nano metres resolution, but I had only ~0.5 mm total linear range. This is still OK for a pendulum sensor.



     See http://mariottim.interfree.it/doc11_e.htm for a typical calibration graph and some further notes on these sensors.



     Regards,



     Chris Chapman