[evlatests] Alternating 'Extended OSRO' and 'RSRO' modes

[Rick Perley]

    I discovered recently that the OPT now gives us the ability to 
configure the correlator in ways not accessible to me in the past (so 
far as I knew).  I tried a test this afternoon where I set up the 
correlator in a 'RSRO' mode with 1 MHz channel width, and 64 MHz subband 
width -- this allows us to cover all of L-band, with full polarization, 
and no 'overlaps' between A/C and B/D.  The A/C half was set to cover 
the lower half of L-band, the B/D half covered the higher half.  I then 
inserted this special L-band mode into my standard 'all bands as fast as 
possible' sequence, which uses 'Extended OSRO' mode:  2 MHz wide 
channels, covering a total of 2 GHz of bandwidth.  The durations for all 
eight bands (L-band in the 'RSRO' mode described above, and all other 
bands in the standard 'Ext. OSRO' mode) were kept at  shortest possible 
values, giving 30 seconds of good on-source data.    The cycle was run 
three times.

    The experiment worked perfectly.  The 'RSRO' mode setup seemlessly 
covers 1024 MHz of bandwidth, from 988 through 2012 MHz.  All L-band 
phases and amplitudes join seemlessly, except of course at the halfway 
point (where A/C and B/D IFs are adjacent).  The time lost between 
recofigurating between these modes was negligible. 

    I'll next try dialing down the subband width, to get to 125 kHz 
channelwidth -- the agreed-upon resolution for the 'full spectral sweep' 
experiment.