[daip] Re: problems with CALIB

[Eric Greisen]

Matt Lister writes:
 > Hi Eric,
 > 
 > There seems to be some problem with the 31DEC04 version of CALIB, 
 > if I run it on a single source VLBA file, it completely flags half of the data 
 > (IFs 3 and 4).  Using the exact same inputs on the same data in the 31DEC03 version, 
 > there is no such flagging.  Also, it is giving these type of errors below, which don't
 > appear in the 31DEC03 version. Was CALIB modified somehow since the last release?
 > I just ran the MNJ and the problems are still there. 
 > 
 > - Matt
 > 
 > BLAZAR> CALIB1: Adjusting solutions to a common reference antenna
 > BLAZAR> CALIB1: CALREF: CANNOT RELATE ANTENNA  4 TO  8 IN SUBARRAY  1
 > BLAZAR> CALIB1: CALREF: CANNOT RELATE ANTENNA  4 TO  8 IN SUBARRAY  1
 > BLAZAR> CALIB1: CALREF: CANNOT RELATE ANTENNA  5 TO  8 IN SUBARRAY  1
 > BLAZAR> CALIB1: CALREF: CANNOT RELATE ANTENNA  5 TO  8 IN SUBARRAY  1
 > BLAZAR> CALIB1: Some solutions were not adjusted to the common
 > BLAZAR> CALIB1: reference antenna. This may be because one or more
 > BLAZAR> CALIB1: antennas were not used. Check your data before
 > BLAZAR> CALIB1: proceeding.

Yes calib was changed to do "correct" weighting of data.  This does
cause some solutions to fail.  The adverb WEIGHTIT was added to allow
you to change the weighting.  Try WEIGHTIT=1.  The CALREF above comes
from there being no time at which there are both solutions for 4 and 8
and for 5 and 8.

A failure to get solutions with the correct weighting may well
indicate pretty bad data....

ERic Greisen