In a message dated 2006/09/17, gpayton880@....... writes:

>     Is there a way or ways to improve the SNR (signal to noise ratio) of an 
> incoming event?   
>     I realize that if the incoming rays are not strong enough, they will not 
> overcome the ambient noise of the system.
>      I would further assume that lowering the ambient noise is the best 
> solution, but I understand that can be a real challenge. Personally, I live in a 
> wooded area, remote from vehicle travel and most man-made noises. But, 
> weather and wind will affect my noise significantly......something I have to live 
> with.
>     My vertical and horizontal sensors are simply laying on a concrete floor 
> in an unattached garage with proper covers over each. Any gain adjustments 
> raise or lower the noise factor directly. I guess that finding a "happy level" 
> and living with it is my only solution. But, if there are any suggestions or 
> remedies out there, I'd appreciate knowing about it..

Hi Jerry,

       The usual ways of dealing with excessive ambient noise are to damp / 
remove the noise source if practicable / desirable, or to move the equipment to 
somewhere which is quieter and / or to also consider lowering the roll off of 
the low pass filter. This can be reduced to 2 to 3 Hz OK for practical 
purposes. Below this, the filter transmission delays may start to effect your 
accuracy and the capacitors get physically bulky.

       The problem with wind noise is that it can cover a very wide frequency 
/ period range and the amplitude depends strongly on the wind velocity. Check 
the ambient noise in quiet periods versus windy periods and take FFTs of the 
recordings to see if there are any particular noise peaks involved? Check the 
amplifier gain to ensure that you have a background of 100 to 300 counts with 
the Lehman. You should be picking up microseisms of maybe 4 to 7 sec period 
quite clearly. What sort of vertical sensor / response are you using?
       
       The professionals avoid the surface noise problem by placing their 
seismometers in deep boreholes; but in some cases they use large natural caves, 
or old mine workings.

       You may also want to check for noise on your ADC converter and for any 
amplifier noise. Remove the connecting cable from one ADC channel, short the 
input, take a recording and see how many bits of noise you have. A good 16 bit 
ADC may have less than 1 bit of noise. Poor ones may have +/- 3 bits noise, 
or more. Reconnect the amplifier, but put a metal film resistor across the 
amplifier input with roughly the same value as the resistance of your coil and 
check for amplifier noise. Reconnect your coil, but wedge the arm of the Lehman 
stationary and check for EM noise pickup. If you get any, maybe pulses, try 
putting an earthed steel cake tin or similar over the pickup coil, or connect a 
0.01 to 0.1 mu F ceramic capacitor across the amplifier input terminals. This 
is effective for RF pickup.
       Am I correct in assuming that your sensor magnet is on the ground and 
that the coil is mounted on the seismometer arm? Systems which use a 
stationary coil and an unshielded moving magnet on the arm are sensitive to a range of 
external noise problems, from pulses on the house power wiring, moving ve
hicles and changes in the Earth's magnetic field, amongst many.

       What is the low pass frequency on your amplifier? If you select a 
section of the trace and press the FFT button, you should get a ~spectrum and be 
able to read off the filter cutoff - or look up the specification?

       What quake ranges are you trying ot detect? Far regional and 
teleseismic quakes mostly have P waves below 2 to 3 Hz, often about 1 Hz. Local quakes, 
local quarrying and local volcanic activity may have components over 20 Hz!  

       Regards,

       Chris Chapman
In a me=
ssage dated 2006/09/17, gpayton880@....... writes:



    Is there a w=
ay or ways to improve the SNR (signal to noise ratio) of an incoming event?&=
nbsp;  

    I realize that if the incoming rays are not strong enough=
, they will not overcome the ambient noise of the system.

     I would further assume that lowering the ambient no=
ise is the best solution, but I understand that can be a real challenge. Per=
sonally, I live in a wooded area, remote from vehicle travel and most man-ma=
de noises. But, weather and wind will affect my noise significantly......som=
ething I have to live with.

    My vertical and horizontal sensors are simply laying on a=
 concrete floor in an unattached garage with proper covers over each. Any ga=
in adjustments raise or lower the noise factor directly. I guess that findin=
g a "happy level" and living with it is my only solution. But, if there are=20=
any suggestions or remedies out there, I'd appreciate knowing about it..



Hi Jerry,



       The usual ways of dealing with excessiv=
e ambient noise are to damp / remove the noise source if practicable / desir=
able, or to move the equipment to somewhere which is quieter and / or to als=
o consider lowering the roll off of the low pass filter. This can be reduced=
 to 2 to 3 Hz OK for practical purposes. Below this, the filter transmission=
 delays may start to effect your accuracy and the capacitors get physically=20=
bulky.



       The problem with wind noise is that it=20=
can cover a very wide frequency / period range and the amplitude depends str=
ongly on the wind velocity. Check the ambient noise in quiet periods versus=20=
windy periods and take FFTs of the recordings to see if there are any partic=
ular noise peaks involved? Check the amplifier gain to ensure that you have=20=
a background of 100 to 300 counts with the Lehman. You should be picking up=20=
microseisms of maybe 4 to 7 sec period quite clearly. What sort of vertical=20=
sensor / response are you using?

       

       The professionals avoid the surface noi=
se problem by placing their seismometers in deep boreholes; but in some case=
s they use large natural caves, or old mine workings.



       You may also want to check for noise on=
 your ADC converter and for any amplifier noise. Remove the connecting cable=
 from one ADC channel, short the input, take a recording and see how many bi=
ts of noise you have. A good 16 bit ADC may have less than 1 bit of noise. P=
oor ones may have +/- 3 bits noise, or more. Reconnect the amplifier, but pu=
t a metal film resistor across the amplifier input with roughly the same val=
ue as the resistance of your coil and check for amplifier noise. Reconnect y=
our coil, but wedge the arm of the Lehman stationary and check for EM noise=20=
pickup. If you get any, maybe pulses, try putting an earthed steel cake tin=20=
or similar over the pickup coil, or connect a 0.01 to 0.1 mu F ceramic capac=
itor across the amplifier input terminals. This is effective for RF pickup.<=
BR>
       Am I correct in assuming that your sens=
or magnet is on the ground and that the coil is mounted on the seismometer a=
rm? Systems which use a stationary coil and an unshielded moving magnet on t=
he arm are sensitive to a range of external noise problems, from pulses on t=
he house power wiring, moving vehicles and changes in the Earth's magnetic f=
ield, amongst many.



       What is the low pass frequency on your=20=
amplifier? If you select a section of the trace and press the FFT button, yo=
u should get a ~spectrum and be able to read off the filter cutoff - or look=
 up the specification?



       What quake ranges are you trying ot det=
ect? Far regional and teleseismic quakes mostly have P waves below 2 to 3 Hz=
, often about 1 Hz. Local quakes, local quarrying and local volcanic activit=
y may have components over 20 Hz!  



       Regards,



       Chris Chapman