In a message dated 2006/11/22, gmvoeth@........... writes:

> I have been playing around for the past 10 years (maybe several months of 
> real time) and have learned that the only decent preamps for a geophone seem 
> to be ones with PHANTOM grounding in the input coil. This means no matter what 
> the coil resistance is there exists a mid point inside the coil you might 
> call a ground. You just need to use a totally shielded line to the geophone 
> tied to circuit ground
> to keep out unwanted electrical stuff.

Hi Geoff,

       You get the lowest amplifier noise using a single +ve opamp input. 
Whether you need to use a differential input depends on other factors. 
       There are usually two output pins on a geophone. I provide a two or 
more core cable secreened with woven copper braid. The screen braid is connected 
to the geophone case and connecting this to earth screens the windings. 

> The gain will drift a bit with temp but not the baseline. You seem to need 
> not more than X10 in the preamp to keep the gain changes small.
> Not less than X10 because of effects on CMMR ?

       Low gains should not effect the CMMR. With high gains you may run out 
of feedback voltage if the output goes to the rails.

> So long as you tie the geophone directly across the +/- inputs in a 
> differential fashion you seem to get a response similar to a full blown 
> instrumentation amplifier.

       Assuming that you have a full differential circuit, you need to 
include the damping resistors in series with the opamp inputs.  

> It is the only way to avoid the resistance vs temp changes from affecting 
> the class "A" operational baseline.

       You have 'lost me there'. What is a class A operational baseline? With 
the input to the +ve opamp input and a damping resistor across the geophone, 
you get very little drift with temperature - the value of the damping resistor 
may be 10x the resistence of the geophone winding.

> The circuit you see is just one of many I play around with through 
> curiosity.
> 
> The original ideas most all come from the PDF files that you can download 
> relating to whatever part the company has sold you. They are full of various 
> circuits you can
> experiment with.

       Where are all these circuits?

> The differential amp always shows two devices with a ground in the middle 
> and I have found this is not necessary. Just hook a coil between + and - of an 
> op amp and the gain is (Rf/(0.5 Rgeo)) Very similar to an instrumentation 
> arrangement
> of three opamps.
> 
> The internal resistance between + and - is 40 megs or so.
> Depends upon the op amp in question.
> I use OP177G from the Philippines.
> They can not be loaded greater than about 2K on the output.
> ( greater load means smaller R)
> 
> What I am not certain of is there a natural DC current flow between + and - 
> or do they always float at the same potential ??

       Always at the same potential - unless the output is on the rails. The 
gain is over 1 million.

> The circuit I showed you seemed to work well but because it puts a common 
> mode signal on the input I am reluctant to use it.

       Looking at the circuit values, the +ve input to ground is FAR too 
high. Put in a 1 K ohm potentiometer, set it to zero and then increase it slowly. 
       I suggest that you use the standard circuit given by Ulmann. The 
circuit you gave will not compensate for the resistance of the geophone. You must 
return the geophone to earth NOT to the +ve input !! 
       It only puts a short circuit across the geophone. 
       You will need to use a 1/f amplifier and also a lot more gain. 

> Instead I am simply now using a gain of X10 on the preamp and tying the + 
> side to the ground using a resistor matched with Rf (R feedback) within the 
> limits of my
> test equipment.
> 
> I have found you need to match the + and - sides of these within maybe 0.1% 
> or your common mode rejection ratio is not so good.
> 
> I could not make the negative resistance circuit work for me that is in the 
> paper:
> 'Overdamping geophones using negative impedances' Bernd Ulmann@......... 
> (may no longer be valid)

       Your reference copied incorrectly.

> It really seems to work opposite of what it says by increasing not 
> decreasing the eigen frequency. and at the same time lowering the sensitivity of the 
> sensor device. 

       You have got the circuit wrong. This is why it doesn't work.

> instead of locking the mass I think you need to somehow artificially 
> decrease the
> spring constant or simulate electronically a greater mass to get the desired 
> results.
> 
> by whatever means I need to create a vertical sensor with a free period of
> three to four seconds.

       Why on Earth do you want that?? This puts you into the sensitivity 
range of the microseisms, particularly on the Pacific coast. Do you really want 
to monitor them? 2 seconds should be adequate and you get this on your existing 
1 Hz geophone, although at a factor of x4 down.

    When I extend the geophone response, I tailor it to roll off at about 
0.35 to 0.4 Hz. This way you pick up the teleseismic S waves quite OK, but the 
microseisms are considerably reduced.

> I have been considering a push-pull arrangement using unity gain amps to 
> force / drive the opamp to behave differently than normally possibly to lower 
> the eigen freq abnormally to 4 seconds.
> 
> It is just a thought and have not seriously looked into this yet.
> 
> Can you please tell me how the given circuit affects the geophone
> because I never got above college algebra and that was 20 years ago.

       It gives a short circuit across the geophone, but it doesn't 
compensate for the resistance. I don't know where this circuit came from, but it won't 
do the job that you want.

> I am going to scan all the drawings of my current setup into a zip file and 
> let anyone interested have a copy. I just need help from those who know how 
> to
> properly (electronically) lower the free period of a mass / spring system to 
> about 4 seconds so I can use a straight forward amplification and have it be 
> linear in gain between 3 seconds and 3 hertz.

       OK, what website?

> I think this might really be done somehow with a narrow band high Q 
> feedback circuit into a push pull arrangement of X1 opamps. To artificially lower 
> the spring constant. Never tried this and am afraid because I do not want to 
> accidentally damage my geophone magnet or coil.

       There are two OK ways of extending the geophone response to lower 
frequencies. The sensitivity is normally flat above the resonant frequency and 
falls as f^2 below this. Resonant or high Q circuits are definitely to be 
avoided. Your geophone is not that stable in characteristiocs. The resistance and the 
spring constants both change with temperature.
       You can provide a two stage amplifier to boost the output below the 
resonant frequency. This is good to 1/10 the resonant frequency, but you may run 
into noise problems below this.
       You can run the geophone into a negative impedance circuit to give an 
overall zero impedance input. This flattens out the charactersitic and gives a 
response proportional to f, so you need to amplify the signal with a single 
stage top cut amplifier.

       You can also digitally filter the output to give up to about x10 
increase in period. This would be the cheapest option! But you need a lot more 
gain.
       
    Regards,

    Chris Chapman
In a me=
ssage dated 2006/11/22, gmvoeth@........... writes:



I have been playing around for=20=
the past 10 years (maybe several months of real time) and have learned that=20=
the only decent preamps for a geophone seem to be ones with PHANTOM groundin=
g in the input coil. This means no matter what the coil resistance is there=20=
exists a mid point inside the coil you might call a ground. You just need to=
 use a totally shielded line to the geophone tied to circuit ground

to keep out unwanted electrical stuff.




Hi Geoff,



       You get the lowest amplifier noise usin=
g a single +ve opamp input. Whether you need to use a differential input dep=
ends on other factors. 

       There are usually two output pins on a=20=
geophone. I provide a two or more core cable secreened with woven copper bra=
id. The screen braid is connected to the geophone case and connecting this t=
o earth screens the windings. 



The gain will drift a bit with=20=
temp but not the baseline. You seem to need not more than X10 in the preamp=20=
to keep the gain changes small.

Not less than X10 because of effects on CMMR ?




       Low gains should not effect the CMMR.=20=
With high gains you may run out of feedback voltage if the output goes to th=
e rails.



So long as you tie the geophone=
 directly across the +/- inputs in a differential fashion you seem to get a=20=
response similar to a full blown instrumentation amplifier.



       Assuming that you have a full differen=
tial circuit, you need to include the damping resistors in series with the o=
pamp inputs.  



It is the only way to avoid the=
 resistance vs temp changes from affecting the class "A" operational baselin=
e.




       You have 'lost me there'. What is a cl=
ass A operational baseline? With the input to the +ve opamp input and a damp=
ing resistor across the geophone, you get very little drift with temperature=
 - the value of the damping resistor may be 10x the resistence of the geopho=
ne winding.



The circuit you see is just one=
 of many I play around with through curiosity.



The original ideas most all come from the PDF files that you can download re=
lating to whatever part the company has sold you. They are full of various c=
ircuits you can

experiment with.




       Where are all these circuits?



The differential amp always sho=
ws two devices with a ground in the middle and I have found this is not nece=
ssary. Just hook a coil between + and - of an op amp and the gain is (Rf/(0.=
5 Rgeo)) Very similar to an instrumentation arrangement

of three opamps.



The internal resistance between + and - is 40 megs or so.

Depends upon the op amp in question.

I use OP177G from the Philippines.

They can not be loaded greater than about 2K on the output.

( greater load means smaller R)



What I am not certain of is there a natural DC current flow between + and -=20=
or do they always float at the same potential ??




       Always at the same potential - unless=20=
the output is on the rails. The gain is over 1 million.



The circuit I showed you seeme=
d to work well but because it puts a common mode signal on the input I am re=
luctant to use it.




       Looking at the circuit values, the +ve=
 input to ground is FAR too high. Put in a 1 K ohm potentiometer, set it to=20=
zero and then increase it slowly. 

       I suggest that you use the standard cir=
cuit given by Ulmann. The circuit you gave will not compensate for the resis=
tance of the geophone. You must return the geophone to earth NOT to the +ve=20=
input !! 

       It only puts a short circuit across the=
 geophone. 

       You will need to use a 1/f amplifier an=
d also a lot more gain. 



Instead I am simply now using a=
 gain of X10 on the preamp and tying the + side to the ground using a resist=
or matched with Rf (R feedback) within the limits of my

test equipment.



I have found you need to match the + and - sides of these within maybe 0.1%=20=
or your common mode rejection ratio is not so good.



I could not make the negative resistance circuit work for me that is in the=20=
paper:

'Overdamping geophones using negative impedances' Bernd Ulmann@......... (ma=
y no longer be valid)




       Your reference copied incorrectly.


It really seems to work opposit=
e of what it says by increasing not decreasing the eigen frequency. and at t=
he same time lowering the sensitivity of the sensor device. 



       You have got the circuit wrong. This is=
 why it doesn't work.



instead of locking the mass I t=
hink you need to somehow artificially decrease the

spring constant or simulate electronically a greater mass to get the desired=
 results.



by whatever means I need to create a vertical sensor with a free period of
three to four seconds.




       Why on Earth do you want that?? This p=
uts you into the sensitivity range of the microseisms, particularly on the P=
acific coast. Do you really want to monitor them? 2 seconds should be adequa=
te and you get this on your existing 1 Hz geophone, although at a factor of=20=
x4 down.



    When I extend the geophone response, I tailor it to roll=20=
off at about 0.35 to 0.4 Hz. This way you pick up the teleseismic S waves qu=
ite OK, but the microseisms are considerably reduced.



I have been considering a push-=
pull arrangement using unity gain amps to force / drive the opamp to behave=20=
differently than normally possibly to lower the eigen freq abnormally to 4 s=
econds.



It is just a thought and have not seriously looked into this yet.



Can you please tell me how the given circuit affects the geophone

because I never got above college algebra and that was 20 years ago.<=
FONT  COLOR=3D"#000000" BACK=3D"#ffffff" style=3D"BACKGROUND-COLOR: #ffffff"=
 SIZE=3D2 PTSIZE=3D10 FAMILY=3D"SANSSERIF" FACE=3D"Arial" LANG=3D"0">



       It gives a short circuit across the ge=
ophone, but it doesn't compensate for the resistance. I don't know where thi=
s circuit came from, but it won't do the job that you want.



I am going to scan all the draw=
ings of my current setup into a zip file and let anyone interested have a co=
py. I just need help from those who know how to

properly (electronically) lower the free period of a mass / spring system to=
 about 4 seconds so I can use a straight forward amplification and have it b=
e linear in gain between 3 seconds and 3 hertz.




       OK, what website?



I think this might really be do=
ne somehow with a narrow band high Q feedback circuit into a push pull arran=
gement of X1 opamps. To artificially lower the spring constant. Never tried=20=
this and am afraid because I do not want to accidentally damage my geophone=20=
magnet or coil.




       There are two OK ways of extending the=
 geophone response to lower frequencies. The sensitivity is normally flat ab=
ove the resonant frequency and falls as f^2 below this. Resonant or high Q c=
ircuits are definitely to be avoided. Your geophone is not that stable in ch=
aracteristiocs. The resistance and the spring constants both change with tem=
perature.

       You can provide a two stage amplifier t=
o boost the output below the resonant frequency. This is good to 1/10 the re=
sonant frequency, but you may run into noise problems below this.

       You can run the geophone into a negativ=
e impedance circuit to give an overall zero impedance input. This flattens o=
ut the charactersitic and gives a response proportional to f, so you need to=
 amplify the signal with a single stage top cut amplifier.



       You can also digitally filter the outpu=
t to give up to about x10 increase in period. This would be the cheapest opt=
ion! But you need a lot more gain.

       

    Regards,



    Chris Chapman