In a message dated 2008/02/17, barry_lotz@............. writes:

> Something has confused me about the volksmeter or the SG sensor. For a 
> passive sensor I always thought the response of the sensor dropped by (I forget 
> the db slope) for frequencies below the natural frequency of the sensor. 
> Therefore I would think the output would have to amplified accordingly for low 
> frequencies. Is the noise amplified also? Is the drop-off mathematically known 
> so the amplification can be correctly compensated? I didn't think a simple 
> intergator in the output of the SG provided a simple flat response of output vs 
> frequency.

Hi Barry,

       If you use a velocity sensor, the output falls off below resonance at 
x100 per decade and stays level above resonance. However, both the SG and the 
Volksmeter use position sensors which fall off at x10 per decade below 
resonance. Hence you can extend the period by about 1000 and still get an OK noise 
level, using a really low noise system. Your 0.9 Hz Volksmeter really will work 
out to 1000 second periods. They also fall off x10 per decade above 
resonance..... but you can also compensate OK for that.
       See Brett's recent posting on feedback systems?
       http://bnordgren.org/seismo/feedback_in_seismic_sensors3.pdf

       Regards,

       Chris Chapman   
In a me=
ssage dated 2008/02/17, barry_lotz@............. writes:



Something has confused me about=
 the volksmeter or the SG sensor. For a passive sensor I always thought the=20=
response of the sensor dropped by (I forget the db slope) for frequencies be=
low the natural frequency of the sensor. Therefore I would think the output=20=
would have to amplified accordingly for low frequencies. Is the noise amplif=
ied also? Is the drop-off mathematically known so the amplification can be c=
orrectly compensated? I didn't think a simple intergator in the output of th=
e SG provided a simple flat response of output vs frequency.



Hi Barry,



       If you use a velocity sensor, the outpu=
t falls off below resonance at x100 per decade and stays level above resonan=
ce. However, both the SG and the Volksmeter use position sensors which fall=20=
off at x10 per decade below resonance. Hence you can extend the period by ab=
out 1000 and still get an OK noise level, using a really low noise system. Y=
our 0.9 Hz Volksmeter really will work out to 1000 second periods. They also=
 fall off x10 per decade above resonance..... but you can also compensate OK=
 for that.

       See Brett's recent posting on feedback=20=
systems?

       http://bnordgren.org/seismo/feedback_in=
_seismic_sensors3.pdf



       Regards,



       Chris Chapman