[daip] Re: X-band data

[Eric Greisen]

Carlos De Breuck writes:

 > AIPS is running fine on our DECs now. This time, I've got a problem with
 > IMAGR. It concerns some X-band snapshot observations. For some reason,
 > the primary flux calibrator did not get observed during this 4h run, while
 > it was in the observe-file. But I have circumvented this problem by using
 > an observation from the day after.
 > 
 > However, there now seems to be a (probably unrelated) problem with
 > cleaning the data. I've split the data, and IMAGR makes the dirty maps
 > without problems. But it won't clean them, crashing during the gridding
 > stage. I've tried re-sorting the data with UVSORT or MSORT, tapering,
 > changing the cellsize (I'm using 0.05 for a 1024^2 image).
 > 
 > Can you have a look if I'm still doing something wrong, or if something is
 > wrong with the data itself (or with IMAGR clean on a DEC running OSF V5.1)?
 > Below are my inputs and outputs of IMAGR, and the header of the UV file (I
 > get the same problem for all 6 sources in this run).
 > 
 > 
 > 
 > FLASH > IMAGR1: Task IMAGR  (release of 31DEC00) begins
 > FLASH > IMAGR1: Create WNJ0305+3525.IMAGR .   1 (UV)  on disk  3  cno   21
 > FLASH > IMAGR1: Beginning channel    1 through    1 with  2 IFs
 > FLASH > IMAGR1: IMACPY: Copied   111873 visibilities to be imaged
 > FLASH > IMAGR1: Create WNJ0305+3525.IBM001.   4 (MA)  on disk  3  cno   28
 > FLASH > IMAGR1: Create WNJ0305+3525.ICL001.   4 (MA)  on disk  3  cno   29
 > FLASH > IMAGR1: GRDFLT: X and Y convolution type = SPHEROIDAL
 > FLASH > IMAGR1: GRDFLT: X and Y parms =  3.0000  1.0000
 > FLASH > IMAGR1: GRDFLT: convolution function sampled every 1/100 of a cell
 > FLASH > IMAGR1: OUVIMG: Sorting data to make them fit
 > FLASH > IMAGR1: GRIDUV: Only   314 Rows in  1280k AP
 > FLASH > IMAGR1: GRIDUV: Max U Baseline     1029679. lambda (=  513 cells)
 > FLASH > IMAGR1: GRIDUV: Ave    2 Channels: 8.435100E+09 to 8.485100E+09 Hz
 > FLASH > IMAGR1: GRIDUV ERROR   4 WRITING GRIDDED DATA
 > FLASH > IMAGR1: OUVIMG: MAKING BEAM BEAM image for CLEAN
 > FLASH > IMAGR1: OUVIMG: FROM UVdata work object
 > FLASH > IMAGR1: CLNINI: ERROR OPENING CLEAN process object
 > FLASH > IMAGR1: CLNUV : ERROR CLEANING CLEAN process object
 > FLASH > IMAGR1: Deleting UV work file:
 > FLASH > IMAGR1: Destroyed  1 extension files of type AN
 > FLASH > IMAGR1: Destroyed  1 extension files of type FQ
 > FLASH > IMAGR1: Destroyed UV image file: catno=     21 disk= 3
 > FLASH > IMAGR1: Purports to die of UNNATURAL causes
 > FLASH > IMAGR1: flash 31DEC00 TST: Cpu=     1.2  Real=    62  IO=8172
 > 
 > 

I have copied your data and tried it here on Linux in versions 02 and
00 and in Charlottesville on our last Dec machine in version 01 (the
only one installed there on Dec).  Unfortunately, it all works just
fine.

The error is an odd one - it claims end of file on write but there was
a test for requested position in the file wrt end of file made before
then which would have generated other messages than the one you get.

By default, the Clean beam is made twice as large as the image when
cleaning but only the same size when not cleaning.  Try IMAGRPRM(10)=1
which may be a work around...

Eric Greisen