[daip] Two curiosities from 1 Software correlator FITS file
Eric Greisen
egreisen at nrao.edu
Wed May 7 13:07:08 EDT 2008
Walter Brisken wrote:
> I have produced a FITS file
>
> /home/swc/difx/projects/bl156b/BAD-IM.FITS
>
> that shows two weird effects. I am not sure if the problems are with my
> data or within AIPS.
>
>
> FIRST ISSUE: I believe is a minor printing issue in LISTR with
> OPTYP='SCAN':
>
> -------------------------------------------------------------
> LISTR1: Task LISTR (release of 31DEC08) begins
> cosmos LISTR(31DEC08) 2179 07-MAY-2008 11:13:38 Page 1
> File = BAD-IM .UVDATA. 3 Vol = 3 Userid = 2179
> Freq = 8.102490000 GHz Ncor = 1 No. vis = 70737
> Scan summary listing
> .
> .
> .
>
> ID Source Freq(GHz) Velocity(Km/s) Rest freq (GHz)
> 1 All Sources 8.1025 0.0000 8.1025
> IF( 2) 8.1165 0.0000 8.1025
> IF( 3) 8.1585 0.0000 8.1025
> IF( 4) 8.2285 0.0000 8.1025
> IF( 5) ********* 0.0000 8.1025
> IF( 6) 8.4105 0.0000 8.1025
> IF( 7) 8.5505 0.0000 8.1025
> IF( 8) 8.5785 0.0000 8.1025
>
> Frequency Table summary
> cosmos LISTR(31DEC08) 2179 07-MAY-2008 11:13:39 Page 4
> File = Y .UVDATA. 3 Vol = 3 Userid = 2179
> FQID IF# Freq(GHz) BW(kHz) Ch.Sep(kHz) Sideband
> 1 1 8.10249000 16000.0010 2000.0001 1
> 2 8.11649000 16000.0010 2000.0001 1
> 3 8.15849000 16000.0010 2000.0001 1
> 4 8.22849000 16000.0010 2000.0001 1
> 5 8.31249000 16000.0010 2000.0001 1
> 6 8.41049000 16000.0010 2000.0001 1
> 7 8.55049000 16000.0010 2000.0001 1
> 8 8.57849000 16000.0010 2000.0001 1
> --------------------------------------------------------------
>
> The frequency for IF(5) is set to "stars". Given that the second table
> doesn't show anything odd, I don't think there is a problem with the
> actual numbers but instead think there is a printing problem. A quick
> PRTAB on the FQ table shows:
>
> --------------------------------------------------------------
> PRTAB1: Task PRTAB (release of 31DEC08) begins
> cosmos PRTAB(31DEC08) 2179 07-MAY-2008 11:13:24 Page 1
> BAD-IM .UVDATA. 4 Disk= 3 FQ Table version 1
> Title: AIPS UV DATA FILE FREQUENCY TABLE
> Created by FITLD on 07-MAY-2008 10:57:07
> Last written by FITLD on 07-MAY-2008 10:57:07
> Ncol 5 Nrow 1 Sort cols:
> Table has 1 keyword-value pairs:
> NO_IF = 8
> Table format incompatable with FITS ASCII tables
>
> COL. NO. 1 2 3 4 5
> ROW FRQSEL IF FREQ CH WIDTH TOTAL BANDWI
> SIDEBAND
> NUMBER HZ HZ HZ
> 1 1 0.0000000000D+00 2.000000E+06 1.600000E+07 1
> 1 1.4000000000D+07 2.000000E+06 1.600000E+07 1
> 1 5.6000000000D+07 2.000000E+06 1.600000E+07 1
> 1 1.2600000000D+08 2.000000E+06 1.600000E+07 1
> 1 2.1000000000D+08 2.000000E+06 1.600000E+07 1
> 1 3.0800000000D+08 2.000000E+06 1.600000E+07 1
> 1 4.4800000000D+08 2.000000E+06 1.600000E+07 1
> 1 4.7600000000D+08 2.000000E+06 1.600000E+07 1
> PRTAB1: Appears to have ended successfully
> PRTAB1: cosmos 31DEC08 TST: Cpu= 0.0 Real= 0
> ---------------------------------------------------------------
>
> Again, nothing odd in the table. The version of AIPS I'm running here
> correctly shows IF(5) for a different test FITS file with 8 IFs, so I am
> suspecting that there might be a compiler or fortran library bug that is
> preventing that certain number from printing properly.
>
>
>
> SECOND ISSUE: FITLD complains that:
>
> -----------------------------------------------------------------------
> FITLD1: GETDEL: No IM entry stn 11 src 4 sub 1 FQ 1 at 0/19:08:15
> FITLD1: GETDEL: No IM entry stn 11 src 4 sub 1 FQ 1 at 0/19:08:16
> FITLD1: GETDEL: No IM entry stn 11 src 4 sub 1 FQ 1 at 0/19:08:17
> .
> .
> .
> FITLD1: GETDEL: No IM entry stn 11 src 4 sub 1 FQ 1 at 0/19:09:24
> -----------------------------------------------------------------------
>
> The messages only refer to one scan (the 4th one) at one antenna (11).
> This is odd for a couple reasons. First, there is no visibility data for
> source 4/antenna 11. The second is that antenna 11 was not in the array
> for scans 1 through 4, so I would have expected warnings for scans 1
> through 3 as well.
>
> This example is of a slightly non-standard case -- I've concattenated
> (outside AIPS) two software correlator passes into 1 FITS file and the
> first only had 10 stations, the second had 11. I expect that this
> concattenation will become a popular way to minimize the total number of
> FITS files that are produced by software correlation.
>
> What happens when no IM table is found? Are zeros just filled in? Will
> this cause trouble for anything using the IM table later?
>
> Thanks!
>
> Walter
>
> _______________________________________________
> Daip mailing list
> Daip at listmgr.cv.nrao.edu
> http://listmgr.cv.nrao.edu/mailman/listinfo/daip
The SO table received by has not set to freq offset column. The 2nd,
5th, and 6th values have junk in them. This leads to 2 vanishingly
small values and 1 one vanishingly large value. I see about the GETDEL
message that it refers to CL table entry times, explaining the default
list I got and the list you got. GETDEL is used to put values in the
CL table GEODLY column. I guess that 0.0d0 is okay at times that the
antenna was not observed. A missing IM table might be serious - but
your data have it to the tune of almost 35000 rows.
I would worry about simple concatenation of correlator output files.
The file breaks allow source ID numbers, antenna ID numbers, etc. to
change. A plain concatenation assumes that they have not.
Eric
More information about the Daip
mailing list