[evlatests] Modcomp-Free Spectral Line Stress Test

[Rick Perley]

    An hour of dynamic time was put to good use this afternoon to 
stress-test the modcomp-free system. 

    Eighteen observations were made of a strong source at X-band, in 18 
different line modes:

    1A at every available bandwidth (codes 0 through 9, except 7)

    4 at every available bandwidth (same codes as above). 

    Result:

    Every mode worked, all frequencies were correct (or, the headers 
were correct, but as this was a continuum source, no reference features 
were available for confirmation), and all data (with the few exceptions 
noted below) were of fabulous quality. 

    Special Notes:

    Antenna 17 was out until the last observation -- stuck in a CCW 
limit, according to the log. 

    Antenna 13 has some peculiarities:

                13A has a very bad bandpass -- either mostly noisy, or 
an undulatory shape with period 2.5 MHz.
                13B has a significant delay error -- 16 nsec
                13C seems o.k.
                13D has a huge delay error -- 90 nsec. 

    Antenna 18, IFs A and C, did not give fringes on two of the scans 
(but was fine on the remaining 16). 
 
    Curiously -- the odd effect noted yesterday, wherein the VLA and 
EVLA were incoherent in IF B, when BWs not equal to 50 were used, did 
not show up today!  I have no explanation for this improvement (but do 
swear that the problem noted yesterday really did happen!)

    Otherwise, the quality of the data were simply excellent.  (Caveat 
-- the data were filled with 10-second averaging, which hides nicely all 
the short-drops and other irregularities noted in the past). 

    It's nice to end the week with a very successful test ...