[daip] Problems with CALIB

[Eric Greisen]

Wing, Nickolaus writes:
 > We are working with VLBI data in both AIPS and difmap.  First we do all the
 > initial corrections in AIPS, everything from ACCOR to FRING and LPCAL.  Then
 > we take the data out with FITTP and use Difmap to create clean maps.
 > 
 > We are now trying to take the uvfits and mapfits files back into AIPS and
 > run CALIB, but it seems to be choking on one or the other.  When one saves
 > in Difmap it creates, among others, a .fits file, the clean map, and a .uvf
 > file, the uv data.  We take these files back into AIPS using FITLD, naming
 > them Source.UICLN.1 and Source.UUVF.1 respectively (our data is U-band). We
 > then calibrate the data in AIPS using CALIB with inc 'UUVF' and in2c
 > 'UICLN', solm 'p' soli 0.5.  Output from CALIB follows this message.
 > 
 > We are using the latest frozen release of AIPS, 31DEC02.
 > 
 > The odd note here is that on our old Sun Solaris machine running AIPS
 > 15Apr99, with the exact same procedure and inputs, and the exact same difmap
 > files, we have no problem whatsoever.
 > 
 > We also have no problem if we clr2n and use smodel 1 0, but we'd rather not
 > run it that way.
 > 
 > I know Difmap is outside your area of responsibility, but any assistance
 > would be greatly appreciated.
 > 
 > Nickolaus Wing
 > Trinity University
 > 
 > ----------------------------------------
 > qso.ph> CALIB1: Task CALIB  (release of 31DEC02) begins
 > qso.ph> CALIB1: CALIB USING 3C212        . UUVF   .   1 DISK=  1 USID=   1
 > qso.ph> CALIB1: Create 3C212       .UUVF  .   2 (UV)  on disk  1  cno   13
 > qso.ph> CALIB1: Selecting the data
 > qso.ph> CALIB1: Doing self-cal mode with CC model
 > qso.ph> CALIB1: Divide data by model - first compute model by summing
 > qso.ph> CALIB1: CC FILE NOT ALL ONE TYPE/SIZE

      This is your problem.  The modern CALIBs read the CC file and
have detected that the file contains Gaussians of different widths
and/or position angles.  The old one does not know about Gaussians and
ignores the bad CC file.  PRTCC may be able to detect this problem.

TABED may let you edit the CC file to fix the problem also.

This is a difmap problem.

Eric Greisen


 > qso.ph> CALIB1: Field   1 used       0 CCs
 > qso.ph> CALIB1: Determining solutions
 > qso.ph> CALIB1: Writing SN table    4
 > qso.ph> CALIB1: Found        17848 good solutions
 > qso.ph> CALIB1: Failed on      556 solutions
 > qso.ph> CALIB1:                532 solutions had insufficient data
 > qso.ph> CALIB1: Adjusting solutions to a common reference antenna
 > qso.ph> CALIB1: ZOPEN: LUN = 28 ALREADY OPENED IN FTAB
 > qso.ph> CALIB1: OTBSRT: ERROR    1 OPENING SCR  SN FILE VERS   4 TO   4
 > qso.ph> CALIB1: TABSRT: OTBSRT ERROR   1
 > qso.ph> CALIB1: TABSRT ERROR     1 SORTING SN TABLE
 > qso.ph> CALIB1: ERROR OCCURED IN CLBADJ
 > qso.ph> CALIB1: Destroyed UV image file: catno=     13 disk= 1
 > qso.ph> CALIB1: Purports to die of UNNATURAL causes
 > qso.ph> CALIB1: qso          31DEC02 TST: Cpu=       0.8  Real=       4
 > _______________________________________________
 > Daip mailing list
 > Daip at listmgr.cv.nrao.edu
 > http://listmgr.cv.nrao.edu/mailman/listinfo/daip
 >