[evlatests] Very strange baseline error

[Rick Perley]

    The 30-hour flux densities run demonstrated a possibly unique 
problem with baselines:  The first source observed (J0217+7349) had 
baselines scaled over 1000 times too large!  The source was observed 
multiple times during the 30 hours -- for each scan, the same baseline 
scaling error occurred.  No other source suffered this problem.  (And 
UVFIX repaired the problem).  I dismissed this as a 'one-of' issue. 

    But it isn't.  The follow-up 'Flux Redux' run suffers exactly the 
same problem, with the same source.  It was also the first object in the 
file. 

    Any ideas here?