PSN-L Email List Message

Subject: Re: Seismic sensors
From: ChrisAtUpw@.......
Date: Fri, 29 Jun 2007 09:57:37 EDT


In a message dated 2007/06/29, Tangazazen@....... writes:

> I note the recent interest in sensors for seismic use. The quadrant photo 
> cell is a better ready made precision device. A spot of light from an LED 
> modulated at 10KHz is projected onto the cell and diametrically opposite 
> quadrants are phase inverted to produce an X-Y signal that change phase and falls to 
> zero as they cross the null point. The output signals are processed by phase 
> sensitive detectors. Since these sorts of sensors are high impedance, the 
> first stage needs a device with a high noise equivalent resistance such as an 
> FET. Unfortunate the flicker noise appears in the noise voltage generator which 
> is equal to the shot noise at 1 or 2 KHz. By operating at 10KHz and a narrow 
> band filter (phase lock loop time constant) the flicker noise becomes 
> insignificant. The output in volt / meter depends a lot on the support electronics 
> but in general the quad photo cell approach seems to offer a high output. The 
> LED is somewhat temperature sensitive (approx. -0.7% C ) but by summing the 
> four quadrants and applying this as feedback this temperature dependence can 
> be made negligible.  For a single axis, pairs of quadrants could be used.     
>   
>   This technique is more of a development that a new idea but might be of 
> interest. 

Hi Martin,

       At what sort of current are you running your photo detectors? I used 
about 100 micro A and got a noise level of about 14 nano metres. The noise that 
you see is inversely proportional to the SQRT of the current. You do need to 
use large area photocells to get the noise down. CAZ opamps can be useful 
here, especially MAX430 or LTC1150. You can use a genuine chopper amplifier to 
reduce this further, maybe by x10.

       How 'quiet' or otherwise is your LED? The metal cased IR types are 
quite good, a lot of the others, much less so. I would call a drift of 0.7% / C 
Deg quite sensitive. Is this a measured figure? Most LEDs show maybe a factor 
of ~ x5 over 0 to 100 C?  (You may need  to read the small print in the spec 
sheet - LED manufacturers seem to be quite shy about the drift in their 
products) 
       This is which is why I have used a filament lamp with a stable voltage 
regulator and a bypass transistor to handle the power. It is also relatively 
easy to stabilise the output signal level further using another Si photocell 
and a reference voltage. 

       Regards,

       Chris Chapman   
In a me=
ssage dated 2007/06/29, Tangazazen@....... writes:

I note the recent interest in s= ensors for seismic use. The quadrant photo cell is a better ready made preci= sion device. A spot of light from an LED modulated at 10KHz is projected ont= o the cell and diametrically opposite quadrants are phase inverted to produc= e an X-Y signal that change phase and falls to zero as they cross the null p= oint. The output signals are processed by phase sensitive detectors. Since t= hese sorts of sensors are high impedance, the first stage needs a device wit= h a high noise equivalent resistance such as an FET. Unfortunate the flicker= noise appears in the noise voltage generator which is equal to the shot noi= se at 1 or 2 KHz. By operating at 10KHz and a narrow band filter (phase lock= loop time constant) the flicker noise becomes insignificant. The output in=20= volt / meter depends a lot on the support electronics but in general the qua= d photo cell approach seems to offer a high output. The LED is somewhat temp= erature sensitive (approx. -0.7% C ) but by summing the four quadrants and a= pplying this as feedback this temperature dependence can be made negligible.=   For a single axis, pairs of quadrants could be used.   = ;   
  This technique is more of a development that a new idea but might be=20= of interest.


Hi Martin,

       At what sort of current are you running= your photo detectors? I used about 100 micro A and got a noise level of abo= ut 14 nano metres. The noise that you see is inversely proportional to the S= QRT of the current. You do need to use large area photocells to get the nois= e down. CAZ opamps can be useful here, especially MAX430 or LTC1150. You can= use a genuine chopper amplifier to reduce this further, maybe by x10.

       How 'quiet' or otherwise is your LED? T= he metal cased IR types are quite good, a lot of the others, much less so. I= would call a drift of 0.7% / C Deg quite sensitive. Is this a measured figu= re? Most LEDs show maybe a factor of ~ x5 over 0 to 100 C?  (You may ne= ed  to read the small print in the spec sheet - LED manufacturers seem=20= to be quite shy about the drift in their products)
       This is which is why I have used a fila= ment lamp with a stable voltage regulator and a bypass transistor to handle=20= the power. It is also relatively easy to stabilise the output signal level f= urther using another Si photocell and a reference voltage.

       Regards,

       Chris Chapman

[ Top ] [ Back ] [ Home Page ]