[daip] INDXR bug?

[Eric Greisen]

Michael Bietenholz wrote:
>>From what I can tell, INDXR (31 DEC 07; MNJ 30 may 2008) gets
> ridiculous values for the CL table amplitude gains, ie. the
> opacity/gain correction, for P-band data only (w/ BPARM=0).  It gets
> gains of ~0.02 for most antennas and 1.000 for a few.
> 
> I seems to happen on either the "CH 0" or the "LINE" data sets.
> Doesn't happen at other frequencies.  FILLM gets reasonable values for
> the same thing.  My copy of /home/aips/TEXT/IONS/VLA.GAINS seems fine,
> all the P-band entries are just 1.000.
> 
>                      michael b
> 
> _______________________________________________
> Daip mailing list
> Daip at listmgr.cv.nrao.edu
> http://listmgr.cv.nrao.edu/mailman/listinfo/daip

I wish I had paid more attention to this baffling one at the time.

It is now solved.  The array inside INDXR for generating the CL table 
had 8 bands, whilst the low level routines knew about all 10 in the new 
way of counting.  So Q band data went in P band locations with the 
constant part (~1) ending up in the 1st derivative (*ZN) of P.  It also 
messed up Q and KA due to this error.  FILLM was correct but not INDXR.

Eric Greisen