[daip] CALIB

[Eric Greisen]

Crystal Brogan writes:

 > When you observe the same source consecutively, but at slightly different 
 > frequencies (within the same band), solint=0 in CALIB does not *appear* to 
 > *always* recognize the true scan boundaries (i.e. despite change in 
 > frequency it is considering it to be a single scan). Since small frequency 
 > changes are sometimes causing large phase jumps these days, this can be 
 > very detrimental - in fact the solution often "fails" completely. It 
 > doesn't seem to happen in every case, possibily depending on whether two 
 > or more integrations are flagged between the two scans.
 > 
 > I had hoped to document this better before reporting it but I haven't 
 > found time (and won't for another week or two) and it is potentially a 
 > fairly serious problem.
 > 

In FILLM, these frequency changes will *probably* start a new scan in
the index table which will then be separated by CALIB.  But if the
index table is created later, when the slight freq difference has been
forgotten, the the index table creator will simply see consecutive
times at the same source and cannot separate the scans because it has
no information with which to do so.

Eric Greisen