[evlatests] Peculiar failure to fringe on ea03, on BD side.

[Rick Perley]

    ea03, on the BD IFs only, failed to give us fringes, but the pattern 
here is unlike any other I've seen. 

    This occurred only at K and Ka bands.  At K-band, only three of the 
39 individual scans were affected.  However, at Ka band, 36 of the 39 
scans were affected.  Most curiously, these 36 were all of one source.  
The other three sources (used for calibration and noise) give good 
fringes.  Weird, eh?  (None of these failures were flagged, so it's 
likely to have something to do with the LO).