[evlatests] Wideband X-band test results

[Rick Perley]

    We ran a short (50 minute) test of the basic 'wideband' mode at 
X-band: 

    * One IF (AC)
    * 8 subbands of 128 MHz each
    * full polarization
    * center frequency 8.400 GHz (so the frequency spanned is 7.888 
through 8.912 GHz)
    * maximum 3 minutes duration for each observation. 

    The sources were 3C286, Cygnus A, and its nearby calibrator, 
J2007+4029.  I followed Michael's prescription in setting up the OPT 
script:  start with four successive 'OSRO' observations, and limit 
durations to 3 minutes maximum.  Referenced pointing was not utilized. 

    The data are of absolutely, astonishingly, good quality! 

    A quick perusal of the data (it is a holiday, after all) reveals:

    1) Only 0.43% integer zero visibilities.
    2) 27 antennas fringed successfully on both IFs
    3) *NO* 'failures to tune' -- all antennas and both polarizations 
fringed every time. 
    4) The only flagging required was for antenna slewing. 
    5) No phase jumps.
    6) No amplitude jumps. 

    Data quality from all antennas is simply superb.  I haven't seen 
anything like this for many months.   CALIB reports a mean closure error 
of 0.043% which sounds rather good...
   
    But there is one annoying issue, which is no surprise:  The subbands 
are mislabeled -- they are both 'out of order', and have their frequency 
axis reversed.  I corrected the latter problem using 'TABPUT'.  What is 
now needed is to interchange subbands 1 with 8, 2 with 7, 3 with 6, and 
4 with 5.  I don't know if AIPS can do this without actually copying out 
each subband separately, and gluing them back together in the right order.