[evlatests] More on Aug 25 2005 tests

[Barry Clark]

Ken and Mark have compared relative sensitivities of EVLA and VLA antennas
using noise measurements on a point source.  A slightly more direct measurement,
and one that should be undertaken, is to simply observe with Tsys correction
turned off.  Fringe amplitude is then simply proportional to correlation
coefficient, with the same constant of proportionality for all antennas.  
We should also do a quick run with this method, just in case some of our
problem is due to second-timescal phase instability.