[evlatests] Modcomp-free Mode mixup

[Rick Perley]

    A very minor problem has been noted on numerous tests.  This seems 
as good a time as any to report it.

    In my normal test sequence, I change from P-band (at 3.125 Mhz) to 
L-band, at 50 MHz.  Each time this has happened, the P-band file  has a 
single record at the end with the wrong frequency and bandwidth.  
Evidently, some data are identified with the following scan's bandwidth 
(50 Mhz), but the preceding scan's LOs. 
    This effect does not happen when changing bands and keeping the BW 
the same. 

    AIPS fills the data with a different 'FQID', so it is trivial to 
remove.