[daip] ANTAB problem reading WX table with 2001 data.

[Eric Greisen]

On 07/25/2015 11:45 PM, R. Craig Walker wrote:
>
> I tried running APCAL on 43 GHz VLBA data taken in 2001.  I was trying
> to fit the opacity.  For every station, I got a message of the form:
> APCAL1: REAWX: *** NO TEMPS FOUND IN WX TABLE FOR ANT 8, LEFT AT 0 ***
> INVERS is set to 1 and there is a WX table with valid looking
> temperatures for all antennas.  So something is wrong, presumably in
> reading that table.
>
> I compared the table to a modern one from data taken last year.  I see
> a couple of differences, but I don't know if they will confuse the
> reader.  The old table has 10 columns while the new one has 11.  The
> missing one in the old table is the subarray (not sure why the weather
> data should depend on subarray, but oh well).  The other difference is
> the column heading for temperature.  The old file had "TEMPE" while
> the new one has "TEMPERA".
>
> I plotted the data with WETHR, and that found the temperature data,
> although it took some considerable staring at the help file to figure
> out how to get it not to edit data.  I suggest a more obvious switch
> than just OPTYPE='MULT'.  Using display programs that can alter your
> data if you put in inappropriate parameters is always a bit scary to
> me.  VPLOT is another example.
>
> WETHR worked and there are reasonable looking temperatures (~63
> points) for most stations.  Unfortunately there are some suspicious
> data where various parameters step a lot (eg pressure from 650 to 980
> mbar at NL) That's in the WX table according to PRTAB, so it is not
> WETHR's fault.  Sometimes I think I shouldn't look at my data!  The
> issue is not there in the weather data plots from tsm (outside AIPS),
> so it appears to be a calibration transfer issue.
>
> To salvage that, I ran VLOG and and tried using the resulting weather
> file.  I see that there is a warning in the VLOG.HLP that the cal.vlba
> file is based on the legacy system.  That will be changing in the
> rather near future so we should remember to update the HLP file.
>
> VLOG worked, after some fumbling, and produced an external WX table.
> APCAL is happy with that and does give answers that are somewhat
> different from what it got guessing values when it couldn't use the WX
> table.
>
> So the main thing I'm reporting here is that APCAL does not seem to be
> able to read the old (2001) WX table even though PRTAB and WETHR can,
> and APCAL can run properly with the same inputs when given an external
> file.  If the cause is not obvious, I could produce a test file, but
> the data are at home so that will take jumping through a few hoops.
>

I will need to see what happens  - the WXINI routine should complain if 
it does not find a column it expects.  And TABWX is very careful to 
honor whatever WXINI set up.

Eric