In a message dated 2006/12/08, dickthomas01@............. writes:

> Have you followed this? Am I missing something? 

       I have followed it. And I suspect that you are missing 'quite a lot' !

This instrument is broadband -- that is great, but this device is horizontal. 
I thought 
> we were interested in P wave transmission ....first. Even if this is truly 
> broadbanded, it isn't going to pick up P wave propagation as well as vertical 
> orientation will?
> 
    If you just want to pick up vertical P waves, a quite narrow band 
response of 0.5 to about 3 Hz, correctly damped, will give optimum results. You don't 
need a broad band vertical seismometer to do it. But you may well need the 
associated surface waves, either L or R, to 'mark' the event clearly. P waves on 
their own can be quite difficult to pick out of the background and you can 
only receive them directly up to ~103 degrees.

The Volksmeter.

       The Volksmeter is a clever new development which can allow a 
''simple'' pendulum sensor to register seismic movements out to 3000 sec period. There 
is growing interest in the Earth's natural vibrations (earth hum) from 0.3 to 
several tens of milli Hertz. If it does prove possible to predict at least 
some earthquakes, this is a likely sensor to be able to do it. Transient 
vibrations of just a few cycles have be observed as the Earth's crust flexes during 
Earth Tide oscillations.   
       Commercial production seismometers are limited to about 360 secs max, 
more typically 120 sec. For Amateur Lehman designs, it is relatively difficult 
to get stable operation much above 30 secs and the instrument is highly tilt 
sensitive. Amateur instruments also use coil + magnet velocity coupling and 
the output voltage falls as the signal frequency decreases. In commercial terms, 
the Volksmeter is inexpensive and it offers very long period, very high 
sensitivity responses.
       This new design ALSO allows the tilt to be measured directly. This 
could be very useful in predicting / monitoring which quakes are likely to 
produce tsunamis, particularly for 'near shore' quakes from the subduction zones off 
the west coast of north and south America. 
     
History and Background.

       For the last 30 years seismometer development has mostly been 
concentrated on a few designs using sprung elements. A lot of effort has gone into the 
design and manufacture of 'better' springs, reducing / compensating the 
effects of temperature, providing prestressed shapes, reducing natural resonance 
effects and reducing the 'internal noise' by alloying and heat treatment. 
However, metals are not 'perfectly elastic', but show very tiny 'steps' in their 
extension / relaxation behaviour which ultimately limits their performance - it 
is not completely smooth (Portevin Le Chatelier effect) This is all reflected 
in the high cost of production of the very long period instruments.

       A precursor to the Volksmeter design is the Shackleford-Gundersen 
Seismometer  which also used a simple pendulum. I suspect that the full potential 
of the S-G design was not realised at the time due to drift and noise in it's 
resonant circuit variable capacitance distance transducer and in the 
difficulty at that time of making a long period electronic integrator. This 1970s 
design was moderately temperature sensitive. When you are looking for ppm 
stability, it is not too helpful if your capacitors and inductors are all temperature 
sensitive and the rectified output voltage depends both on the temperatures of 
two diodes and on their power dissipation. The original integrator and other 
circuits were later updated with FET input and low noise opamps. However, the 
capacitance variation was still provided by varying the separation of ~parallel 
capacitor plates. This gives a non linear position output. As the central 
plate moves, air is pumped out of one side and into the other. This air pumping 
effect is highly non linear with varying gap width and it can seriously effect 
the damping. Negative feedback was provided, but this also effects the signal 
to noise ratio. 
       
       The new capacitative sensor uses a central moving 'electrostatic 
shadow mask' plate sliding in between an AC driver plate with two parallel strips 
and a sensor plate with four cross linked elements. The shadow mask plate does 
not need to be earthed, so it is completely free to move and it experiences no 
air damping. The output is highly linear with displacement over several mm. 
Electromagnetic damping is provided.

       A specialised capacitative bridge measurement chip, the AD7745, has 
been used. This and the 24 bit ADC are controlled by a dedicated microprocessor. 
The extremely high resolution and the use of position sensing, rather than a 
velocity output, allows digital processing of the signals to hugely extend the 
period. For surface waves with periods above about 300 sec, the tilt 
component of the signal becomes very significant and this can be measured independent 
of the period. 

       This new sensor development should be a very significant addition to 
the science of Seismology. See 
http://www.seismicnet.com/volksmeter/State-of-the-art_Digital_Seismograph.pdf and http://physics.mercer.edu/hpage/peters.html

       Regards,

       Chris Chapman MA
In a me=
ssage dated 2006/12/08, dickthomas01@............. writes:



Have you followed this? Am I mi=
ssing something? 




       I have followed it. And I suspect that=20=
you are missing 'quite a lot' !



This instrument is broadband -- that is great, but this device is horizontal=
.. I thought 

we were interested in P wave tr=
ansmission ....first. Even if this is truly broadbanded, it isn't going to p=
ick up P wave propagation as well as vertical orientation will?



    If you just want to pick up vertical P waves, a quite na=
rrow band response of 0.5 to about 3 Hz, correctly damped, will give optimum=
 results. You don't need a broad band vertical seismometer to do it. But you=
 may well need the associated surface waves, either L or R, to 'mark' the ev=
ent clearly. P waves on their own can be quite difficult to pick out of the=20=
background and you can only receive them directly up to ~103 degrees.



The Volksmeter.



       The Volksmeter is a clever new develop=
ment which can allow a ''simple'' pendulum sensor to register seismic moveme=
nts out to 3000 sec period. There is growing interest in the Earth's natural=
 vibrations (earth hum) from 0.3 to several tens of milli Hertz. If it does=20=
prove possible to predict at least some earthquakes, this is a likely sensor=
 to be able to do it. Transient vibrations of just a few cycles have be obse=
rved as the Earth's crust flexes during Earth Tide oscillations.  =
 

       Commercial production seismometers are=20=
limited to about 360 secs max, more typically 120 sec. For Amateur Lehman de=
signs, it is relatively difficult to get stable operation much above 30 secs=
 and the instrument is highly tilt sensitive. Amateur instruments also use c=
oil + magnet velocity coupling and the output voltage falls as the signal fr=
equency decreases. In commercial terms, the Volksmeter is inexpensive and it=
 offers very long period, very high sensitivity responses.

       This new design ALSO allows the tilt to=
 be measured directly. This could be very useful in predicting / monitoring=20=
which quakes are likely to produce tsunamis, particularly for 'near shore' q=
uakes from the subduction zones off the west coast of north and south Americ=
a. 

     

History and Background.



       For the last 30 years seismometer devel=
opment has mostly been concentrated on a few designs using sprung elements.=20=
A lot of effort has gone into the design and manufacture of 'better' springs=
, reducing / compensating the effects of temperature, providing prestressed=20=
shapes, reducing natural resonance effects and reducing the 'internal noise'=
 by alloying and heat treatment. However, metals are not 'perfectly elastic'=
, but show very tiny 'steps' in their extension / relaxation behaviour which=
 ultimately limits their performance - it is not completely smooth (Portevin=
 Le Chatelier effect) This is all reflected in the high cost of production o=
f the very long period instruments.



       A precursor to the Volksmeter design is=
 the Shackleford-Gundersen Sei=
smometer  which also used a simple pendulum. I suspect that the ful=
l potential of the S-G design was not realised at the time due to drift and=20=
noise in it's resonant circuit variable capacitance distance transducer and=20=
in the difficulty at that time of making a long period electronic integrator=
.. This 1970s design was moderately temperature sensitive. When you are looki=
ng for ppm stability, it is not too helpful if your capacitors and inductors=
 are all temperature sensitive and the rectified output voltage depends both=
 on the temperatures of two diodes and on their power dissipation. The origi=
nal integrator and other circuits were later updated with FET input and low=20=
noise opamps. However, the capacitance variation was still provided by varyi=
ng the separation of ~parallel capacitor plates. This gives a non linear pos=
ition output. As the central plate moves, air is pumped out of one side and=20=
into the other. This air pumping effect is highly non linear with varying ga=
p width and it can seriously effect the damping. Negative feedback was provi=
ded, but this also effects the signal to noise ratio. 

       

       The new capacitative sensor uses a cent=
ral moving 'electrostatic shadow mask' plate sliding in between an AC driver=
 plate with two parallel strips and a sensor plate with four cross linked el=
ements. The shadow mask plate does not need to be earthed, so it is complete=
ly free to move and it experiences no air damping. The output is highly line=
ar with displacement over several mm. Electromagnetic damping is provided.


       A specialised capacitative bridge measu=
rement chip, the AD7745, has been used. This and the 24 bit ADC are controll=
ed by a dedicated microprocessor. The extremely high resolution and the use=20=
of position sensing, rather than a velocity output, allows digital processin=
g of the signals to hugely extend the period. For surface waves with periods=
 above about 300 sec, the tilt component of the signal becomes very signific=
ant and this can be measured independent of the period. 



       This new sensor development should be a=
 very significant addition to the science of Seismology. See http://www.seis=
micnet.com/volksmeter/State-of-the-art_Digital_Seismograph.pdf and http://ph=
ysics.mercer.edu/hpage/peters.html



       Regards,



       Chris Chapman MA