In a message dated 20/09/02, charles.r.patton@........ writes:

Hi Charles,

       Great effort in assembling the very informative EMail! Good info and=20
thought provoking!=20
=20
> Chris wrote, "There is NO PROBLEM in this case. Didn't you check the
> data sheet?" =20
> I have to admit I was trying to take the lazy way out and just ask the=20
> question and so I didn't check the spec sheet.  I went back and checked ou=
t=20
> the data sheet for the 3515 and 3516, (27501.10B) and did not see any=20
> statements about magnetic or non-magnetic leads - so where does that=20
>=20

       There is no reference to a flux concentrator since none is used. If=20
one is used, details are given on the data sheet and on the drawings.

       I have checked the on-line A3515 data sheet and you are perfectly=20
correct that the lead material is not mentioned. My apologies! It was=20
specified in application information relating the A3505. I confirm that the=20
A3515 leads are non magnetic.

>  "What other companies are producing devices similar to the MS3110, please=
?"=20
>  =20
> Analog Microelectronics GmbH in Mainz, Germany, has developed a family
> of analog ASICs, the CAV404, CAV 414 and CAV424.  Downloadable spec
> sheets at: http://www.servoflo.com/analog/data_ho.htm

       Thanks. Google came up with XEMICS, fairly similar to the MS3110, but=
=20
they do quote resolutions, etc. and seem to be clued up on getting high=20
accuracy; see http://www.xemics.com/internet/index2.jsp

> So on to the theory.  The "New Manual of Seismological Observatory
> Practice" which covers many topics is at:
> http://www.seismo.com/msop/nmsop/nmsop.html and discusses many areas of=20
> seismometer construction, but in particular the noise floor discussion at:
>=20

       See also http://www.ifjf.uib.no/seismo/SOFTWARE/OTHER/instrument.pdf
'Instrumentation in Earth Seismology' is a very good account and it is=20
complete, but it is a large file. The Stuttgart files seem to be all divided=
=20
into short sections.=20

> And in a discussion on electronic displacement sensors the statement,
> "Their sensitivity is limited by the ratio between the electronic noise of=
=20
> the demodulator and the electrical field strength; it can be a hundred=20
>=20

       However, both capacitative and inductive types can give resolutions o=
f=20
less than 1 nano metre, so while I have no doubt that this is correct, it=20
need not effect us. You need a very quiet site indeed to be able to make use=
=20
of this sort of resolution.=20

> Part of the reason a capacitive, or even an inductive sensor, can surpass=20
> the Hall effect or LED shutter types is that at the conversion point from=20=
a=20
> mechanical quantity to an electrical one, there is no noise in the=20
> capacitive sensor because it is reactive.  A bit of explanation.=20
>     There can be several noise sources in a sensor application such as a=20
> seismometer.  External, such as the one discussed by Sean-Thomas (in the=20
> email mentioned below) set one limit.  If your electronics were perfect,=20
> you can't do any better than your external noise floor.  But many times=20
> there are additional local sources, which can amount to traffic in the=20
> street, trees waving in the breeze, etc.  Again these can become the noise=
=20
> floor.
>     But next comes the internally generated noise sources. The first one i=
s=20
> thermal.  The arm mass of your seismometer has a temperature and is=20
> "vibrating" with Brownian motion. The air around that mass also has=20
> Brownian motion which can push on the mass generating noise. This source i=
s=20
>=20

       The Brownian motion of the sensor may require you to use a seismic=20
mass of more than 1 oz using feedback type damping, with typical masses=20
several times this. Systems with passive damping may need to be heavier.

       The air may also develop motion due to thermal gradients, whichcan be=
=20
severe and may be more of a problem with long period / large instruments. Th=
e=20
movements may be continuous, only appear every few minutes, intermittently,=20
or when the outside air cools down.=20
=20
>    Barometric pressure changes "float" the mass which can translate into=20
> movement, especially on vertical seismometers. =20

       This is a major source of noise on verticals which are not in a seale=
d=20
container and may include pressure fluctuations due to wind.

Next comes the=20
> electronic noise -- the heart of the discussion that prompts this
> reply.  A non-reactive (resistive) sensor such as a Hall effect sensor or=20
> photosensor vs a reactive sensor such as capacitive or inductive sensor=20
> suffer from at least two major noise sources, thermal and current noise. =20
> These are physical, inherent properties.  The thermal noise is proportiona=
l=20
> to the square root of the product of the resistance, temperature and BW. =20
> Amounting to 1.29 nV for 100 ohms, 1 Hz BW, and 295 degrees K. =20
> Current noise is proportional to current and inversely proportional to the=
=20
> frequency of measurement.  I.e., it goes up as the frequency goes down and=
=20
>=20


> So glancing at the A3515LUA, we find that this floor is 400 uV for a 10 to=
=20
> 10KHz BW in a +/- 2V range.  This limits your total range to 10,000 to 1,=20
> or about 80 dB.                      Capacitive and inductive displacement=
=20
> sensors do not have to have either of these noise sources since they are=20
>=20

       Perhaps I should have made it clearer that the resolution that I=20
quoted for the A3515 based linear magnetic detector was actually measured an=
d=20
was determined for 30 sec periods, peak to peak. The above characteristic=20
agrees with my readings fairly well. The 400 micro V RMS =3D 566 micro V pk=20
(presumably over 1/10 sec), but the bandwidth is only 10 Hz so the rough=20
noise estimate is ~ 18 micro V and this is a 'typical' single deviation valu=
e=20
only. You may get about 2 standard deviations over longer periods. After=20
taking into account some increase in low frequency noise levels and adding=20
power supplies, an amplifier and filter, my <100 micro V resolution is not=20
too bad. It can be further reduced if you opt for 3 Hz or less bandwidth. Th=
e=20
overall range is ~40,000, which is better than the 14 true bits (16,384 -=20
84dB) you get out of most nominally 16 bit A/D converters.

>     A more subtle problem is another noise called the 1/f or flicker noise=
=20
> problem in semiconductors.  This noise also rises as the frequency goes=20
> down.  So the numbers for the A3515LUA will almost surely get worse as we=20
> go down in frequency for a seismometer.  In particular, 10 Hz is high, whe=
n=20
> we want to look at teleseismic events, which is exactly where we need good=
=20
>=20

       This was why I was so pleased when I learnt about the A3515s! I had=20
been working with A3505s previously and they were significantly noisier at=20
the low frequencies. However, Allegro put a chopper amplifier in the A3515=20
which rejects most 1/f noise and this is exactly what we need for the longer=
=20
period seismic observations! They are uniquely well suited to this=20
application.

       One note: the output impedance of an A3515 is very low and there is=20
some very wide band noise on it. This can be rejected by connecting a 4.7 mH=
=20
choke and a 10 nF ceramic capacitor to the output pin. A 100 nF ceramic=20
capacitor should be connected across the supply lines close to the device.=20
=20
>     One way around 1/f noise is the carrier amplifier, i.e., if you put th=
e=20
> information on a carrier frequency, amplify this carrier frequency, then=20
> demodulate, you step around the 1/f problem.  Coincidentally, this is=20
> exactly what happens in capacitative and inductive displacement sensors.=20
> A high frequency is impressed across the sensor capacitor (or inductor). =20
> This capacitor changes value according to the displacement, yielding=20
> (modulating) that high frequency (carrier) which is amplified by a carrier=
=20
> amplifier. The output of that carrier amplifier is then detected and=20
> filtered at a high level where the 1/f, thermal and current noises are wel=
l=20
> below the output.  So, you'll note that Sean-Thomas used an inductive=20
> sensor on his vertical seismometer.  That is one route.
> Another is that you could easily build the design idea "Circuit resolves
> 0.1-fF change from 100 pF by Derek Redmayne, Linear Technology Corp,
> Milpitas, CA  (from EDN Access Design Ideas1/6/2000)" at:
> www.e-insite.net/ednmag/index.asp?layout=3Darticle&articleId=3DCA46462&
> pubdate=3D01/06/2000
> The major cost of the semis in quantities of one would be about $32, but=20
> this includes a 24 bit digitizer sampling at 7.5 samples/second.=20
> Resolving a 0.1 fF change in 100 pF is a 1,000,000 to 1, or 120 dB
> range. =20

       This looks like the most promising candidate for the differential=20
capacity measurements that we need. The gain is only two, but this could=20
easily be increased and two off two pole filters, a high pass filter ~40 sec=
=20
and an output amplifier added to give a complete system. If you used a +/- 1=
0=20
mm range for a Lehman, the step size for a 14 bit true A/D would be about 1.=
2=20
microns. To allow for drift you add a high pass filter and then amplify the=20
signal x100, or as convenient. The gives high resolution and is not sensitiv=
e=20
to slow drifts. You could add a 'cheap' tuning 10 V movng coil meter to the=20
direct output to read off the position of the arm, to know when it needs=20
re-centering.

One measure to decide if this is sufficient. Sean-Thomas=20
> mentioned in an email post of Nov 15, 2000 to the PSN list that, "The norm=
al=20
> background noise for a seismometer like a Lehman are the 6-second=20
> microseisms, usually caused by storms off the east coast.  Away from the=20
> immediate shore (100km) these run 2 to 4 microns peak-to-peak=E2=80=A6"  S=
o for=20
> ease of computation, let's set a goal of 1 micron.  Therefore the capacito=
r=20
> plates could be set apart by 1e6 * 1micron, or 1 meter and we could still=20
> resolve the mechanical background.  Practical construction might be to hav=
e=20
> parallel plates about 10 mm away on each side (~1/2 inch) from a moving=20
> plate on a SG pendulum.  Then for 100 pf between the plates we need: C=20=
=3D=20
> 2.249E-13 * Er * sq.in * (N-1) / (in of separation) so rearranging,=20
> substituting and solving we get that we need 222 sq.in
> of area.  Not good.  but lets reduce the spacing to about 0.05" and we get=
=20
> 22 sq.in or a plate a bit over 4.5" square. =20

       You need to check on air damping for sizeable capacitor plates which=20
move toward or away from each other. This can be a serious limitation and=20
there are 'guideline' estimates.=20

 A variation would be to salvage the parallel plates of a variable tuning=20
capacitor such=20
> as in a broadcast receiver.  Fully meshed, the large gang was 365 pF.  So,=
=20
> mount so the pendulum meshes the group.  Now you have perhaps a 1" range=20
>=20

       Maintaining the alignment could be just a bit difficult with an ex=20
radio variable capacitor and they are of a rather odd shape, designed to giv=
e=20
a ~linear wavelength scale. The high voltage transmitter types have larger=20
spacing. It would probably be both easier and more satisfactory to make a=20
variable area plate capacitor out of sheet brass and spacers.=20

Another solution would be to implement a slightly different circuit such=20
> as the one in the SETRA patents: 4,054,833: 5,194,819: and 6,316,948. Now=20
> the amplifier can provide gain, and the voltage tracks the ratio of bridge=
=20
> caps, so small capacitance changes generate larger voltages and therefore=20
> you could use smaller single plate sensors.
> An example of this was the design idea "Bridge Measures Small Capacitance"=
=20
> by Jeff Witt, Linear Technology Corp, in Electronic Design, Nov. 4, 1996,=20
>=20

       See http://www.linear.com/pdf/an87.pdf page 87 for this article.

       This circuit is designed to measure a variable capacity with referenc=
e=20
to a fixed capacity using an integrating feedback path. One problem with sei=
s=20
sensors is that the variable capacity also varies with temperature, at about=
=20
20 ppm / C Deg for Brass. This does not sound a lot, but if the full sensor=20
range was 20 mm, this would represent 0.4 micron / C deg, when we may be=20
interested in quake signals down to 1/40 of this or less. You could, however=
,=20
make a fixed reference capacitor out of the same materials. A differential=20
capacity arrangement only experiences the 20 ppm variation on the difference=
,=20
not on the overall capacity - the zero should be designed to be ~constant=20
with temperature. They are also less sensitive to eccentric movement /=20
wobbles at right angles to the normal movement. The 'professional' type=20
sensor that I looked at was made from silver plated invar.

 This circuit seemed to be similar to the SETRA circuits and allow gain. =20
> This final circuit is probably the best compromise of the lot.  Take it an=
d=20
> add the 24 bit A/D, LTC2400, and you have a powerful combination for low=20
> cost.  Since the initial output is analog, you have a node that can be fed=
=20
> back to the force coil on the SG and at the same time you get a simple=20
>=20

       The signals which you wish to resolve and record will be of different=
=20
frequency and may be far lower in amplitude than the ocean microseisms -=20
maybe 1% the size. Having to record large 'noise' signals at the same time a=
s=20
much smaller quake signals puts definite limits on the signal processing.=20
 =20
       It is a pity that the LTC2400 is only available in a SOIC miniature=20
package.

> Oh, yes, an excellent reference for capacitance sensor design: "Capacitive=
=20
>=20

       I agree! It is very good and describes a lot of very diverse practica=
l=20
applications in detail. ISBN =3D 078035351X. Unfortunately, the paperback=20
edition is $99-95 new, but you might try your library.

       Regards,

       Chris Chapman
In a message dated 20/09/=
02, charles.r.patton@........ writes:



Hi Charles,



       Great effort in assembling the very=
 informative EMail! Good info and thought provoking!=20

=20

Chris wrote, "There is NO P=
ROBLEM in this case. Didn't you check the

data sheet?"  

I have to admit I was trying to take the lazy way out and just ask the q=
uestion and so I didn't check the spec sheet.  I went back and checked=20=
out the data sheet for the 3515 and 3516, (27501.10B) and did not see any st=
atements about magnetic or non-magnetic leads - so where does that informati=
on appear?



       There is no reference to a flux con=
centrator since none is used. If one is used, details are given on the data=20=
sheet and on the drawings.



       I have checked the on-line A3515 da=
ta sheet and you are perfectly correct that the lead material is not mention=
ed. My apologies! It was specified in application information relating the A=
3505. I confirm that the A3515 leads are non magnetic.



 "What other companies are=20=
producing devices similar to the MS3110, please?"   

Analog Microelectronics GmbH in Mainz, Germany, has developed a family

of analog ASICs, the CAV404, CAV 414 and CAV424.  Downloadable spec

sheets at: http://www.servoflo.com/analog/data_ho.htm



       Thanks. Google came up with XEMICS,=
 fairly similar to the MS3110, but they do quote resolutions, etc. and seem=20=
to be clued up on getting high accuracy; see http://www.xemics.com/internet/=
index2.jsp



So on to the theory.  =
The "New Manual of Seismological Observatory

Practice" which covers many topics is at:

http://www.seismo.com/msop/nmsop/nmsop.html and discusses many areas of=20=
seismometer construction, but in particular the noise floor discussion at:

http://www.geophys.uni-stuttgart.de/seismometry/man_html/node28.html



       See also http://www.ifjf.uib.no/sei=
smo/SOFTWARE/OTHER/instrument.pdf

'Instrumentation in Earth Seismology' is a very good accou=
nt and it is complete, but it is a large file. The Stuttgart files seem to b=
e all divided into short sections.=20



And in a discussion on elec=
tronic displacement sensors the statement,

"Their sensitivity is limited by the ratio between the electronic noise=20=
of the demodulator and the electrical field strength; it can be a hundred ti=
mes better than that of the inductive type. 



       However, both capacitative and indu=
ctive types can give resolutions of less than 1 nano metre, so while I have=20=
no doubt that this is correct, it need not effect us. You need a very=20=
quiet site indeed to be able to make use of this sort of resolution.=
=20



Part of the reason a capaci=
tive, or even an inductive sensor, can surpass the Hall effect or LED shutte=
r types is that at the conversion point from a mechanical quantity to an ele=
ctrical one, there is no noise in the capacitive sensor because it is reacti=
ve.  A bit of explanation.=20

    There can be several noise sources in a sensor applic=
ation such as a seismometer.  External, such as the one discussed by Se=
an-Thomas (in the email mentioned below) set one limit.  If your electr=
onics were perfect, you can't do any better than your external noise floor.=20=
 But many times there are additional local sources, which can amount to=
 traffic in the street, trees waving in the breeze, etc.  Again these c=
an become the noise floor.

    But next comes the internally generated noise sources=
.. The first one is thermal.  The arm mass of your seismometer has a tem=
perature and is "vibrating" with Brownian motion. The air around that mass a=
lso has Brownian motion which can push on the mass generating noise. This so=
urce is problem for miniature sensors such as the Analog Devices MEMs types.=
 



       The Brownian motion of the sensor m=
ay require you to use a seismic mass of more than 1 oz using feedback type d=
amping, with typical masses several times this. Systems with passive damping=
 may need to be heavier.



       The air may also develop motion due=
 to thermal gradients, whichcan be severe and may be more of a problem with=20=
long period / large instruments. The movements may be continuous, only appea=
r every few minutes, intermittently, or when the outside air cools down.=20

=20

   Barometric pre=
ssure changes "float" the mass which can translate into movement, especially=
 on vertical seismometers.  



       This is a major source of noise on=20=
verticals which are not in a sealed container and may include pressure fluct=
uations due to wind.



Next comes the=20

electronic noise -- the hea=
rt of the discussion that prompts this

reply.  A non-reactive (resistive) sensor such as a Hall effect sen=
sor or photosensor vs a reactive sensor such as capacitive or inductive sens=
or suffer from at least two major noise sources, thermal and current noise.=20=
 These are physical, inherent properties.  The thermal noise is pr=
oportional to the square root of the product of the resistance, temperature=20=
and BW.  Amounting to 1.29 nV for 100 ohms, 1 Hz BW, and 295 degrees K.=
  

Current noise is proportional to current and inversely proportional to t=
he frequency of measurement.  I.e., it goes up as the frequency goes do=
wn and is somewhat dependent on the construction/material of the resistance.=
 





So glancing at the A3515LUA=
, we find that this floor is 400 uV for a 10 to 10KHz BW in a +/- 2V range.=20=
 This limits your total range to 10,000 to 1, or about 80 dB.  &nb=
sp;            &=
nbsp;      Capacitive and inductive displaceme=
nt sensors do not have to have either of these noise sources since they are=20=
essentially reactive, and the A3515LUA noise sources are due to resistance.<=
/BLOCKQUOTE>



       Perhaps I should have made it clear=
er that the resolution that I quoted for the A3515 based linear magnetic det=
ector was actually measured and was determined for 30 sec periods, peak to p=
eak. The above characteristic agrees with my readings fairly well. The 400 m=
icro V RMS =3D 566 micro V pk (presumably over 1/10 sec), but the bandwidth=20=
is only 10 Hz so the rough noise estimate is ~ 18 micro V and this is a 'typ=
ical' single deviation value only. You may get about 2 standard deviations o=
ver longer periods. After taking into account some increase in low frequency=
 noise levels and adding power supplies, an amplifier and filter, my <100=
 micro V resolution is not too bad. It can be further reduced if you opt for=
 3 Hz or less bandwidth. The overall range is ~40,000, which is better than=20=
the 14 true bits (16,384 - 84dB) you get out of most nominally 16 bit A/D co=
nverters.



    A more s=
ubtle problem is another noise called the 1/f or flicker noise problem in se=
miconductors.  This noise also rises as the frequency goes down.  =
So the numbers for the A3515LUA will almost surely get worse as we go down i=
n frequency for a seismometer.  In particular, 10 Hz is high, when we w=
ant to look at teleseismic events, which is exactly where we need good low n=
oise performance. 



       This was why I was so pleased when=20=
I learnt about the A3515s! I had been working with A3505s previously and the=
y were significantly noisier at the low frequencies. However, Allegro put a=20=
chopper amplifier in the A3515 which rejects most 1/f no=
ise and this is exactly what we need for the longer pe=
riod seismic observations! They are uniquely well suited to th=
is application.



       One note: the output impedance of a=
n A3515 is very low and there is some very wide band noise on it. This can b=
e rejected by connecting a 4.7 mH choke and a 10 nF ceramic capacitor to the=
 output pin. A 100 nF ceramic capacitor should be connected across the suppl=
y lines close to the device.=20

=20

    One way=20=
around 1/f noise is the carrier amplifier, i.e., if you put the information=20=
on a carrier frequency, amplify this carrier frequency, then demodulate, you=
 step around the 1/f problem.  Coincidentally, this is exactly what hap=
pens in capacitative and inductive displacement sensors.=20

A high frequency is impressed across the sensor capacitor (or inductor).=
  This capacitor changes value according to the displacement, yielding=20=
(modulating) that high frequency (carrier) which is amplified by a carrier a=
mplifier. The output of that carrier amplifier is then detected and filtered=
 at a high level where the 1/f, thermal and current noises are well below th=
e output.  So, you'll note that Sean-Thomas used an inductive sensor on=
 his vertical seismometer.  That is one route.

Another is that you could easily build the design idea "Circuit resolves

0.1-fF change from 100 pF by Derek Redmayne, Linear Technology Corp,

Milpitas, CA  (from EDN Access Design Ideas1/6/2000)" at:

www.e-insite.net/ednmag/index.asp?layout=3Darticle&articleId=3DCA464=
62&pubdate=3D01/06/2000

The major cost of the semis in quantities of one would be about $32, but=
 this includes a 24 bit digitizer sampling at 7.5 samples/second.=20

Resolving a 0.1 fF change in 100 pF is a 1,000,000 to 1, or 120 dB

range.  



       This looks like the most promising=20=
candidate for the differential capacity measurements that we need. The gain=20=
is only two, but this could easily be increased and two off two pole filters=
, a high pass filter ~40 sec and an output amplifier added to give a complet=
e system. If you used a +/- 10 mm range for a Lehman, the step size for a 14=
 bit true A/D would be about 1.2 microns. To allow for drift you add a high=20=
pass filter and then amplify the signal x100, or as convenient. The gives hi=
gh resolution and is not sensitive to slow drifts. You could add a 'cheap' t=
uning 10 V movng coil meter to the direct output to read off the position of=
 the arm, to know when it needs re-centering.



One measure to decide if this is sufficient. Sean-Thomas=20

mentioned in an email post=20=
of Nov 15, 2000 to the PSN list that, "The normal background noise for a sei=
smometer like a Lehman are the 6-second microseisms, usually caused by storm=
s off the east coast.  Away from the immediate shore (100km) these run=20=
2 to 4 microns peak-to-peak=E2=80=A6"  So for ease of computation, let'=
s set a goal of 1 micron.  Therefore the capacitor plates could be set=20=
apart by 1e6 * 1micron, or 1 meter and we could still resolve the mechanical=
 background.  Practical construction might be to have parallel plates a=
bout 10 mm away on each side (~1/2 inch) from a moving plate on a SG pendulu=
m.  Then for 100 pf between the plates we need: C =3D 2.249E-13 * Er *=20=
sq.in * (N-1) / (in of separation) so rearranging, substituting and solving=20=
we get that we need 222 sq.in

of area.  Not good.  but lets reduce the spacing to about 0.05=
" and we get 22 sq.in or a plate a bit over 4.5" square.  This is doabl=
e. 



       You need to check on air damping fo=
r sizeable capacitor plates which move toward or away from each other. This=20=
can be a serious limitation and there are 'guideline' estimates.=20



 A variation would be to salvage the parallel plates of a variable tunin=
g capacitor such=20

as in a broadcast receiver.=
  Fully meshed, the large gang was 365 pF.  So, mount so the pendu=
lum meshes the group.  Now you have perhaps a 1" range with about 300 p=
F change.



       Maintaining the alignment could be=20=
just a bit difficult with an ex radio variable capacitor and they are of a r=
ather odd shape, designed to give a ~linear wavelength scale. The high volta=
ge transmitter types have larger spacing. It would probably be both easier a=
nd more satisfactory to make a variable area plate capacitor out of sheet br=
ass and spacers.=20



Another solution would be to implement a slightly different circuit such=
=20

as the one in the SETRA pat=
ents: 4,054,833: 5,194,819: and 6,316,948. Now the amplifier can provide gai=
n, and the voltage tracks the ratio of bridge caps, so small capacitance cha=
nges generate larger voltages and therefore you could use smaller single pla=
te sensors.

An example of this was the design idea "Bridge Measures Small Capacitanc=
e" by Jeff Witt, Linear Technology Corp, in Electronic Design, Nov. 4, 1996,=
 pg. 110. 



       See http://www.linear.com/pdf/an87.=
pdf page 87 for this article.



       This circuit is designed to measure=
 a variable capacity with reference to a fixed capacity using an integrating=
 feedback path. One problem with seis sensors is that the variable capacity=20=
also varies with temperature, at about 20 ppm / C Deg for Brass. This does n=
ot sound a lot, but if the full sensor range was 20 mm, this would represent=
 0.4 micron / C deg, when we may be interested in quake signals down to 1/40=
 of this or less. You could, however, make a fixed reference capacitor out o=
f the same materials. A differential capacity arrangement only experiences t=
he 20 ppm variation on the difference, not on the overall capacity - the zer=
o should be designed to be ~constant with temperature. They are also less se=
nsitive to eccentric movement / wobbles at right angles to the normal moveme=
nt. The 'professional' type sensor that I looked at was made from silver pla=
ted invar.



 This circuit seemed to be similar to the SETRA circuits and allow gain.=
  

This final circuit is proba=
bly the best compromise of the lot.  Take it and add the 24 bit A/D, LT=
C2400, and you have a powerful combination for low cost.  Since the ini=
tial output is analog, you have a node that can be fed back to the force coi=
l on the SG and at the same time you get a simple digital stream to feed you=
r computer. 



       The signals which you wish to resol=
ve and record will be of different frequency and may be far lower in amplitu=
de than the ocean microseisms - maybe 1% the size. Having to record large 'n=
oise' signals at the same time as much smaller quake signals puts definite l=
imits on the signal processing.=20

  

       It is a pity that the LTC2400 is on=
ly available in a SOIC miniature package.



Oh, yes, an excellent refer=
ence for capacitance sensor design: "Capacitive Sensors, Design and Applicat=
ions," by Larry K. Baxter, IEEE Press.



       I agree! It is very good and descri=
bes a lot of very diverse practical applications in detail. ISBN =3D 0780353=
51X. Unfortunately, the paperback edition is $99-95 new, but you might try y=
our library.



       Regards,



       Chris Chapman