[evlatests] Estimate of continuum coherence in several backend filters

[Mark Claussen]

One simple estimate of the continuum coherence due only to the 
phase change across the baseband is to assume vectors all with 
the same length and rotated by the phase in each channel.  The 
cosine of the phase multiplied by the length of the vector is 
then the projection of that vector onto the x-axis.  Summing all 
the projected vectors associated with each channel, and normalizing 
this by the sum of all the lengths (number of channels, assuming 
unity for the length) gives a measure of the coherence.  For random 
phases distributed over 2*pi radians, this sum should give something 
like zero.

Anyway I made these calculations; the table below is the result.


              Coherence using above estimate

                 50 MHz   25 MHz  12.5 MHz  6.25 MHz  3.125 MHz

Old EVLA         0.854     0.753    -----    -----     0.540
Corrected EVLA   0.995     0.970    0.893    0.802     0.733
VLA              0.998     0.998    0.999    0.998     0.998



Mark