[evlatests] 'Bleeding' with rapid switching mode

[Rick Perley]

    Greg Taylor called this evening to report that using 'fast 
switching' mode causes a faint, false image of the calibrator to appear 
in the target image.  He reports that a mild 'QUACK' removes the problem. 

    This should reminiscent of a problem we used to have, but also of a 
known problem we have now -- most source observations have an initial 
record which clearly belongs to the preceding source.   With 'fast 
switching', every source's preceding observation is a calibrator, so I 
wonder if this 'misplaced record' problem is responsible for the 
reported ghost image. 

    It's important that our current users know about this -- the current 
excitement over the SGR source has much time being used in 
fast-switching mode ..., and we wouldn't want a reported detection to be 
only the ghost of the preceding calibration.