[daip] Resolved: WSRT Tsys correction

[Eric Greisen]

M.A.W. Verheijen wrote:
> Hi Eric,
> 
>> But SELN should work.
> 
> I've verified that the combination 
> 
>  OPCODE   'CALP'
>  INTERPOL 'SELN'
> 
> in CLCAL produces a CL-2 table in which only the calibrator scans have
> Tsys corrections.  There are no Tsys corrections for the target source
> in the CL-2 table.  The full list of CLCAL inputs in included below. 
> 
>> Is it possible that the SN table does not
>> have the target source number correct in the source column?
> 
> With PRTAB I had a look at the SN-1 table as produced by APCAL.  I found
> the following:
> 
>       TIME RANGE         SOURCE ID
> 
> 0/22:12:00 - 0/22:41:40      1        (3C48)
> 0/22:46:30 - 1/10:46:30      2        (49 target source scans)
> 1/10:50:00 - 1/11:19:20     51        (3C147)
> 
> So there is something unconventional with the source numbers but I would
> have thought that QUAL=-1 (-1=>all) in CLCAL would ignore that. 
> 
> For completeness I have attached PP08.FITS.gz which also contains the
> SN-1 table produced with APCAL using TY-11 and GC-11, as well as the
> CL-2 table produced with CLCAL using OPCODE='CALP' and INTERPOL=''. 
> This CL-2 table will then be used by SPLAT to apply the time-dependent
> Tsys corrections to all sources. 
> 
> I hope that the attached SN-1 and CL-2 tables will be useful in figuring
> out the CLCAL behaviour.  For the moment, we are very happy that we can
> produce a complete CL-2 table so we can move on with the next steps. 

Thanks for sending the TASAV file - it makes diagnosis easy:

Your data set has source numbers 1 through 51, where 1 and 51 are the 
calibrators and 2 through 50 are different sources by name although some 
seem tp have the same coordinates as others.  The CL table honors these 
source numbers with entries for each one appropriately.  The TY table 
only contains source numbers 1, 2, and 51 so the SN table from APCAL 
only contains those numbers.  Poor old CLCAL, told to interpolate each 
source only to itself should produce a result containing only 1, 2, and 
51.

The adverb QUAL is in the modern world probably confusing - the VLA 
would allow observations of a source name, e.g. 3C286 to be 
differentiated by attaching a numeric "qualifier" - a separate column in 
the SU table.  Note that 3C286 qual 1 and 3C286 qual 2 would have two 
separate source numbers and would be treated as if they were totally 
different sources.  People did observe interleaved programs of sources 
with qual 1 and qual 2 and then use the adverb to select all sources 
with qual 1 or all with 2 separately.  The EVLA does not offer this 
option amd messes with CALCODE instead.

The oddity in your data is the choice of naming each scan with a 
different name - when aips comes to imaging it will only image 1 scan at 
a time because of this.  The current issue is in the TY tables that only 
have one target source number so that CLCAL with SELN or SELF cannot do 
the other target source numbers.

Eric Greisen