[evlatests] Initialization problems

[Rick Perley]

    I've been running numerous tests over the past 3 days, to prepare 
for the upcoming 'flux densities run'.

    All of these tests involve short observations of multiple sources at 
multiple frequencies and bands.  For all these tests -- with a single 
exception--  the first observation is at P-band, in continuum, with 
requested frequencies of 327.5 and 307.5 MHz, and a bandwidth of 3.125 
MHz. 

    But for all of these tests, the actual frequency observed *on the 
initial* P-band observation, was *not* at the requested frequencies.   
For four of the five test runs which started with a P-band source, the 
actual frequencies observed on the initial observation was 4.6874 MHz 
too high.  For the other test, the observed frequencies were 23.4375 MHz 
too high.  These two numbers are half the bandwidth below 6.25 and 25 
MHz respectively -- seemingly indicating that the system is 
'remembering' a setup from an earlier observation?

    But it's a bit more curious than this, however.  In all five of 
these tests, the first two records -- both flagged were at the correct 
frequency.  The system then jumped to the incorrect frequencies, at 
which it stayed for the entire length of the initial scan. All 
subsequent observations at this band were at the correct frequencies. 

    Finally -- the one test that did not start at P-band ran correctly 
at all bands (including the initial observation, which was at K-band). 

    Until this issue is sorted out, we must emphasize to our users to 
continue to utilize a 'dummy scan' at the beginning of their observe 
files.