[evlatests] New P-Band troubles

[Rick Perley]

    My standard stress test observes P-band, at a BW of 3 MHz, and the 
other bands at a BW of 50 MHz. 
    Until today, the data which arrives has been correctly labeled. 

    Today, however, I find the AIPS thinks the BW at P-band is 50 MHz.  
The observe script *absolutely* asks for 3 MHz bandwide, and the data 
look fine (other than some caveats I'll report on later). 
   
    I had Ken look at the on-line displays during the observation, and 
he noted the BW was indeed set to 3 MHz. 

    Something, which influences what AIPS thinks is going on, has 
changed since Wednesday afternoon.