From: antillies79@gmail.com   
      
   I have a fiber optic sensor system with output power fluctuations limited by   
   source RIN.   
      
   I want to subtract out that noise component from the output by measuring the   
   source noise with a reference detector that picks off a portion of the source   
   light and measures its fluctuations.   
      
   The twist is the sensor light is delayed by propagation through L = 100-1000 m   
   of optical fiber while the reference light is not, and the sensor signal is   
   concentrated not at DC but rather around frequency f = 1/(2*dT) where dT is   
   the propagation delay    
   time through the fiber. Thus noise at all-important frequency f experiences a   
   pi phase shift on propagation through the sensor and 0 phase shift in the   
   reference arm, and subtracting the signal and reference actually increases the   
   noise (bad). One way to    
   fix this would be to add L meters of fiber in the reference arm (bad). Another   
   idea is to ADD the sensor and reference signals instead of subtracting and use   
   the pi phase shift to cancel out the noise. Is this latter option a feasible   
   approach?   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)