In a message dated 06/12/2007, gmvoeth@........... writes:

Has anyone ever used a  spring to take up most the mass weight then maybe 
apply an ounce or less  levitation magnetically to set with electronics the free 
period within a  geophone.

I'd think you would pump a current limited signal into the  geophone and use 
some kind of
feedback circuit that might set the free  period. I am looking only for a 
free period in the 3 to 4 second  range.



Hi All,
 
    Trying to provide magnetic 'springs' for  seismometers tends to quickly 
run into the problem of temperature coefficients  and drifts. Alnico has the 
lowest Tc, being about -1 or -2 x10^-4 / C Deg  depending on the grade. Sm-Co is 
about -3x10^-4 / C Deg and NdFeB is about  -11x10^-4 / C Deg. Ferrite magnets 
are a lot worse.
    Sandwiches of thick Sm-Co and a thin opposing sheet  of NdFeB have been 
used to provide ~high stability quadrupole magnets, but  stabilising the 
temperature of the magnetic material is probably an easier  option.
    The torsion coefficient of Steel Springs is about  -2.6x10^-4 / C Deg. 
Only Elinvar and NiSpanC springs have very low coefficients  in the ppm range.
    If you consider electromagnets, you have the  problem of voltage 
reference drifts and amplifier drifts, as well as expansion  coefficients and the 
thermal convection from heating. The common voltage  reference diodes still have 
drifts of 20 to 150 ppm / C Deg. The few types with  drifts in the low ppm 
range like the LM399 and LT1027 are quite expensive. Type  'prec ref' into 
_www.digikey.com_ (http://www.digikey.com)  
    Weak springs were used to extend the period of the  1 to 3 second 
Willmore seismometers to about 20 seconds. These were produced in  both vertical and 
horizontal versions. 'Trimming' of the period by magnetic  repulsion has also 
been used on horizontal Lehman seismometers, but it can  render them sensitive 
to stray magnetic field changes. A similar repulsion  method has been used on 
clock pendulums to correct for period / swing angle  timing errors. 
    Trying to use the magnetic repulsion of pyrolitic  graphite also tends to 
run into Tc problems as well as giving large stray  magnetic fields.
    Geophones can have their natural periods very  considerably extended (to 
<~1/20 natural frequency), if they are fitted with  an amplifier with a 
negative input impedance roughly equal to their electrical  resistance. The current 
generated by any motion is fed into a current to voltage  converter and this 
output is proportional to frequency. The  armature is effectively held 
stationary with respect to the case. A second  amplifier stage with the output 
proportional to 1/f gives the normal velocity  output. The sensitivity is quite low. 
These are described at 
_http://www.lennartz-electronic.de/PDF_documents/Seismometers.pdf_ (http://www.lennartz-electronic.de/PDF_documents/Seismometers.pdf) 
 
    An alternative method also used by Lennartz is  described by Roberts, 
P.M. 'A Versatile Equilization Circuit for  Increasing Seismometer Velocity 
Response Below the Natural  Frequency'  BSSA Vol 79, no 4, pp 1607-1617 Oct. 1989 
See also _http://jclahr.com/science/psn/roberts/index.html_ 
(http://jclahr.com/science/psn/roberts/index.html)  This  circuit selectively amplifies the f^2 
'tail' of the normal geophone  response below resonance and can provide a 
period extension of ~x10. Both types  require the use of very low noise amplifiers, 
preferably with discreet matched  pair low noise input transistors. 
   Aaron Barzilai modified geophones by adding a  capacitative position 
detector. See 
_http://micromachine.stanford.edu/smssl/projects/Geophones/ASMEWin'98FinalDraftSlides.pdf_ 
(http://micromachine.stanford.edu/smssl/projects/Geophones/ASMEWin'98FinalDraftSlides.pdf)  and  other references cited at 
_http://micromachine.stanford.edu/smssl/projects/Geophones/_ 
(http://micromachine.stanford.edu/smssl/projects/Geophones/)  Unfortunately  he seems to have used a 
difficult mechanical modification and his  digital electronics, also described, was 
more noisy than it needed to be. 
 
    Regards,
 
    Chris Chapman



   






In a message dated 06/12/2007, gmvoeth@........... writes:
<=
FONT=20
  style=3D"BACKGROUND-COLOR: transparent" color=3D#000000>Has anyone ever us=
ed a=20
  spring to take up most the mass weight then maybe apply an ounce or less=20
  levitation magnetically to set with electronics the free period within a=20
  geophone.

I'd think you would pump a current limited signal into th=
e=20
  geophone and use some kind of
feedback circuit that might set the free=20
  period. I am looking only for a free period in the 3 to 4 second=20
  range.


Hi All,
 
    Trying to provide magnetic 'springs' for=20
seismometers tends to quickly run into the problem of temperature coefficien=
ts=20
and drifts. Alnico has the lowest Tc, being about -1 or -2 x10^-4 / C D=
eg=20
depending on the grade. Sm-Co is about -3x10^-4 / C Deg and NdFeB is ab=
out=20
-11x10^-4 / C Deg. Ferrite magnets are a lot worse.
    Sandwiches of thick Sm-Co and a thin opposing s=
heet=20
of NdFeB have been used to provide ~high stability quadrupole magnets,=20=
but=20
stabilising the temperature of the magnetic material is probably an easier=20
option.
    The torsion coefficient of Steel Springs is abo=
ut=20
-2.6x10^-4 / C Deg. Only Elinvar and NiSpanC springs have very low coefficie=
nts=20
in the ppm range.
    If you consider electromagnets, you have the=20
problem of voltage reference drifts and amplifier drifts, as well as expansi=
on=20
coefficients and the thermal convection from heating. The common voltag=
e=20
reference diodes still have drifts of 20 to 150 ppm / C Deg. The few types w=
ith=20
drifts in the low ppm range like the LM399 and LT1027 are quite expensive. T=
ype=20
'prec ref' into www.digikey.com 
    Weak springs were used to extend the period of=20=
the=20
1 to 3 second Willmore seismometers to about 20 seconds. These were produced=
 in=20
both vertical and horizontal versions. 'Trimming' of the period by magnetic=20
repulsion has also been used on horizontal Lehman seismometers, but it can=20
render them sensitive to stray magnetic field changes. A similar repuls=
ion=20
method has been used on clock pendulums to correct for period / swing angle=20
timing errors. 
    Trying to use the magnetic repulsion of pyrolit=
ic=20
graphite also tends to run into Tc problems as well as giving large stray=20
magnetic fields.
    Geophones can have their natural periods very=20
considerably extended (to <~1/20 natural frequency), if they are fitted w=
ith=20
an amplifier with a negative input impedance roughly equal to their electric=
al=20
resistance. The current generated by any motion is fed into a current to vol=
tage=20
converter and this output is proportional to frequency. The=20
armature is effectively held stationary with respect to the case. A second=20
amplifier stage with the output proportional to 1/f gives the normal velocit=
y=20
output. The sensitivity is quite low. These are described at ht=
tp://www.lennartz-electronic.de/PDF_documents/Seismometers.pdf 
    An alternative method also used by Lennartz is=20
described by Roberts, P.M. 'A Versatile Equilization Circuit for=20
Increasing Seismometer Velocity Response Below the Natural=20
Frequency'  BSSA Vol 79, no 4, pp 1607-1617 Oct. 1989 See als=
o http://jclahr.com/=
science/psn/roberts/index.html This=20
circuit selectively amplifies the f^2 'tail' of the normal geophone=20
response below resonance and can provide a period extension of ~x10. Both ty=
pes=20
require the use of very low noise amplifiers, preferably with discreet match=
ed=20
pair low noise input transistors. 
   Aaron Barzilai modified geophones by adding a=20
capacitative position detector. See http://micromachine.stanford.edu/smssl/projects/Geopho=
nes/ASMEWin'98FinalDraftSlides.pdf and=20
other references cited at http://m=
icromachine.stanford.edu/smssl/projects/Geophones/ Unfortunately=20
he seems to have used a difficult mechanical modification and his=20
digital electronics, also described, was more noisy than it needed to b=
e.=20

 
    Regards,
 
    Chris Chapman