Chris,

Thanks for the reply.  The phrase that caught my eye was "very long 
periods".  I have an SG sensor that is based on Larry's electronics 
board, which doesn't use a coil/magnet sensor.  But I have always had a 
substantial amount of "wandering" with a period of 60 seconds or more. 
I always have attributed it to incompletely controlled air currents 
inside the enclosure.  But after several different enclosures, internal 
baffles, warming the box from the top with a couple of power resistors, 
etc, I have never made much of an impact on the noise.  I wound up by 
usin a 30 second hi-pass filter to knock down my "very long period" noise.

Larry

ChrisAtUpw@....... wrote:
> In a message dated 18/02/2009, lconklin@............ writes:
> 
>     In a recent post, Chris Chapman commented "Coil + magnet systems detect
>     velocity and are likely to suffer noise problems at very long periods".
>     Could you elaborate a little on this?  What is the source of the noise,
>     and what sort of periods are we talking about?
>     Larry Conklin
>     lconklin@............
> 
> Hi Larry,
>  
>     Apart from the background seismic noise, you have to consider the 
> input voltage and current noise of the amplifier, the intrinsic noise of 
> the input resistances on both inputs and the 1/f (flicker) noise of the 
> system. Note that a coil has noise determined by it's resistance. The 
> 1/f noise is likely to be a limiting factor below about a few Hz. It can 
> be avoided by using a chopper or a CAZ amplifier. Note that CAZ 
> amplifiers like the MAX420, MAX430, LTC1150 etc., still have much more 
> noise than true chopper amplifiers, although their drift and VLF noise 
> is reduced when compared to ordinary amplifiers. See AN-45 from 
> www.linear.com  
>  
>     The RMS voltage noise of a resistor = SQRT(4.k.T.R.B) where k = 
> 1.38x10^-23 in J/K, T is the absolute temperature in Deg Kelvin (Deg C + 
> 273.14), R is the resistance on Ohms and B is the bandwidth in Hz.
>  
>     Professional seismometers usually use variable capacitor sensors and 
> chopper amplifiers measuring position. See AN-87, p87 at www.linear.com 
>  There is usually no significant intrinsic noise 
> associated with a capacitor, as there is with a resistor or an inductor. 
> These give the same output per mm of movement whether this occurs over 1 
> or 1000 seconds.  However, a coil + magnet velocity detection system 
> would only give 1/1000 the voltage output for the 1000 second signal as 
> compared to a 1 second signal, so you quickly reach the amplifier noise 
> limits as the period increases.
>  
>     The seismic background noise is likely to be greater than the 
> 'ordinary' amplifier noise for periods up to ~ 30 seconds, maybe quite a 
> bit more. The complicating factor is the variation in the minimum 
> background seismic noise seen at different periods.
>     See 
> http://www.geophys.uni-stuttgart.de/oldwww/seismometry/man_html/node28.html
>  
>     I hope that this helps.
>  
>     Regards,
>  
>     Chris Chapman
__________________________________________________________

Public Seismic Network Mailing List (PSN-L)