[evlatests] L-Band data -- Origin of yesterday's reported closure errors identified ...

[Rick Perley]

    Looking at yesterday's L-band data now ...

    Frequencies appear to be correct. 

    25R and 26L give perfect (integer) zero amplitudes.
    6, 15, and 20 give only noise on both polarizations
    26R and 27L give only noise. 

    Delays are HUGE = +/- 200 nsec.  Indeed, with the relatively coarse 
frequency resolution (1 MHz), the phase gradient across the channel 
width (0.45 turns in the worse case!) is sufficient to explain the 
'closure' errors I reported on yesterday. 

    No initial flagging was required. 

    More to come ...