[daip] Re: AIPS UV Sort order parameter in FITS file?

[Eric Greisen]

Glen Langston writes:
 > Hi Eric,
 > 
 > I'm using the AIPS single-dish functions to grid some calibrated
 > GBT data.  The data I write is sorted in TB order.  I'd like to
 > put that info in the FITS file so that I can skip the AIPS UVSRT
 > step for TVFLG and SPFLG.
 > 
 > Which of these lines is used by AIPS to determine the
 > sort order of UV data in a FITS File?
 > 
 > HISTORY SORTORD = 'TB      '           / Data time sorted for efficiency
 > ...
 > HISTORY UVSRT  SORT = 'TB' / New sort order
 > ...
 > HISTORY AIPS   SORT ORDER = 'TB'
 >               / Where T means TIME (IAT)
 >               / Where B means BASELINE NUM

      I suspect that this third one is what is used.

 > 
 > Thanks
 > 
 > Glen
 > 
 > PS I've written a somewhat general purpose function to write AIPS SDFITS
 > files.  I find that is hard to use TVFLG or SPFLG to find and flag
 > the bad visibilities.   Since my data have one spectra for each
 > spot in the map, it is easiest to make a map with a very small convolution
 > function, then use maxfit to find the offending spectra then
 > PRTSD to print the visiblities until I find the location, and finally
 > use UVFLG with the appropriate TIMER to excise the visibility.
 > 
       TVFLG is not appropriate to spectral data, SPFLG is what one
would use.  You have N channels and TVFLG can only look at one at a
time.  45*270 = 12150 not 36,400 so I do not see the problem.  AIPS
can now go to 32768 in an image.

 > When using TVFLG, my data get an error message, Too many times
 > since the uv data set has 36,400 visibilities organized in 45 rows
 > of 270 spectra.   Each visibility has is own unique time tag.
 > Is there a better way of organizing the random groups parameter to
 > get TVFLG to make a square grid of all data?

Eric