In a message dated 10/03/2005, jpopelish@........ writes:

I am  getting pretty deeply into the design process for a velocity coil 
seismometer  amplifier and filter and have decided to try to quantify the relative 
noise  performance of various candidate opamps versus coil resistance.  Here 
is  the list of low noise and general purpose types (which I have on hand) that 
I  am comparing:

LT1007
OP27 
OP07
`    Good luck!
 
    OPA604?
    LT1028
    LT1677
    Linear have noise tables for opamps.
 
    These are the commonly used ones for coil / magnet  systems. The choice 
very much depends on the source  resistance that you need to match.
    Coil systems do not normally sense down to  DC, but to some maximum 
period. You put in a high pass filter to remove the  1/f noise and any drift. The 
1/f noise can be very significant.
    The MAX432 has about the best specifications  of the CAZ group. These are 
useful if you want very long periods, since they do  not have any 1/f noise, 
but they are essentially best for high impedance  inputs. They work quite well 
for DC systems, since they have very low  thermal drift.
    You can get very significantly improved noise  performance using discrete 
transistor inputs, either fet, npn or  pnp and also by using a true chopper 
amplifier.
    
   My biggest difficulty in completing this task has been  resolving the wide 
variation in noise specifications for the various types to  produce a valid 
comparison in the DC to 10 Hz frequency range, especially  comparing the 
chopper types to the others.


The chopper types quoted seem to be all fet  input, which have high impedance 
inputs. The current noise is very low, but  the voltage noise is significant.
 
    There is a paper from the 90's? comparing  noise in seismometer amplifier 
systems, but it is now a bit out of  date. I can't remember the reference off 
hand - I think it was bull seis  soc am. I am away from home at the moment, 
so I can't look it up. I am sure that  this sort of choice has been described 
in psn.quake.net Archives - try  putting in a query? Sean Morrissey?



Regards,
 
    Chris Chapman





In a message dated 10/03/2005, jpopelish@........ writes:
<=
FONT=20
  style=3D"BACKGROUND-COLOR: transparent" face=3DArial color=3D#000000 size=
=3D2>I am=20
  getting pretty deeply into the design process for a velocity coil seismome=
ter=20
  amplifier and filter and have decided to try to quantify the relative nois=
e=20
  performance of various candidate opamps versus coil resistance.  Here=
 is=20
  the list of low noise and general purpose types (which I have on hand) tha=
t I=20
  am comparing:

LT1007
OP27 
OP07
`    Good luck!
 
    OPA604?
    LT1028
    LT1677
    Linear have noise tables for opamps.
 
    These are the commonly used ones for coil / mag=
net=20
systems. The choice very much depends on the source=20
resistance that you need to match.
    Coil systems do not normally sense down to=20
DC, but to some maximum period. You put in a high pass filter to remove=20=
the=20
1/f noise and any drift. The 1/f noise can be very significant.
    The MAX432 has about the best specificat=
ions=20
of the CAZ group. These are useful if you want very long periods, since they=
 do=20
not have any 1/f noise, but they are essentially best for high impedance=20
inputs. They work quite well for DC systems, since they have very low=20
thermal drift.
    You can get very significantly improved noise=20
performance using discrete transistor inputs, either fet, npn or=
=20
pnp and also by using a true chopper amplifier.
    
   My biggest difficulty in completing this task has b=
een=20
resolving the wide variation in noise specifications for the various types t=
o=20
produce a valid comparison in the DC to 10 Hz frequency range, especially=20
comparing the chopper types to the others.

 
    The chopper types quoted seem to be all fet=
=20
input, which have high impedance inputs. The current noise is very low,=
 but=20
the voltage noise is significant.
 

    There is a paper from the 90's? comparing=20
noise in seismometer amplifier systems, but it is now a bit out of=
=20
date. I can't remember the reference off hand - I think it was bull sei=
s=20
soc am. I am away from home at the moment, so I can't look it up. I am sure=20=
that=20
this sort of choice has been described in psn.quake.net Archives - try=20
putting in a query? Sean Morrissey?


 
    Regards,
 
    Chris Chapman