[evlatests] First Sept 08 3-bit test

[Rick Perley]

    Ken set up another 3-bit test, this time at 6928 MHz, where there is 
more power available to the sampler. 

    Only 3 antennas were available, as the hardware from antenna 27 is 
being moved to antenna 12. 

    Ken observed on and off 3C286, one minute duration on each, for 
about 30 minutes. 

    The results are interesting -- the RR to LL asymmetry is back ...

    Below is a matrix showing the rms noise, in Jy, for each correlator, 
with RR in the upper right, and LL in the lower left:


          15          22         28
           ------------------------
15  |     XX      .246      .187
22  |    .150       XX      .289
28  |   .169       .162        XX

    The value of 150 mJy for 15 x 22 in LCP is very close to the 
expected theoretical.  As in our previous tests, RCP is significantly 
worse, *particularly* for antenna 22. 
    The histograms have about 20,000 visibilities each, and are 
fabulously gaussian.  (They are of 'textbook quality').