[daip] Problem with loading EVLA data into AIPS with BDF2AIPS... (fwd)

[Eric Greisen]

Nissim Kanekar wrote:
> 
> On Wed, 15 May 2013, Eric Greisen wrote:
> 
>> Nissim Kanekar wrote:
>>>
>>> Dear Eric,
>>>
>>> We've been having some weird problems with mails today, so I wasn't 
>>> sure whether a mail that I sent to you earlier today got through and 
>>> thought I should forward it again (below). Apologies if you got the 
>>> earlier one!
>>>
>>> Cheers,
>>> Nissim
>>>
>>>
>>>
>>> ---------- Forwarded message ----------
>>> Date: Wed, 15 May 2013 20:14:31 +0530 (IST)
>>> From: Nissim Kanekar <nkanekar at ncra.tifr.res.in>
>>> To: egreisen at nrao.edu
>>> Subject: Problem with loading EVLA data into AIPS with BDF2AIPS...
>>>
>>>
>>> Dear Eric,
>>>
>>> I've run into a bit of a strange problem with loading some recent 
>>> EVLA data into AIPS with BDF2AIPS in 31DEC13. The task seems to run 
>>> fine and produces an output file with the correct tables and the 
>>> correct source information; however, the number of visibilities in 
>>> each scan is shown by LISTR to be zero (see the attached file, with 
>>> the LISTR output). Was wondering if I'm doing something obviously 
>>> wrong here; could you let me know ?
>>>
>>> Thanks!
>>> Nissim
>>>
>>
>>
>> I did not get your earlier e-mail.
>>
>> Check your adverbs to LISTR - a number of them apply to the SCAN 
>> listing when you would think that they would not (SOURCE among others).
>>
>> Try PRTAB on the NX file (DOCRT=1).
>>
>> Eric
>>
> 
> Hi Eric,
> 
> I tried the same LISTR inputs (inext 'SU', opty 'SCAN', source '') on other
> EVLA dataset and they do list the right number of visibilities, so I 
> suspect that this should be fine.
> 
> I also tried PRTAB on the NX file with DOCRT=1 and got blank output:
> 
> 
> socrates  PRTAB(31DEC12)    187     16-MAY-2013  03:16:49    Page    1
> BDF2AIPS    .UVEVLA.   1  Disk= 1    NX Table version   1
> Title: AIPS  NX
> Created by      BDFIn on 15-MAY-2013 18:35:03
> Last written by BDFIn on 15-MAY-2013 18:35:03
> Ncol 7  Nrow 0    Sort cols:
>    Table format incompatable with FITS ASCII tables
> 
> COL. NO.         1             2             3             4             
> 5     6 7
>      ROW    TIME          TIME INTER    SOURCE ID     SUBARRAY      
> START VIS END VIS       FREQ ID  NUMBER    D/HMS         D/HMS
> 
> 
> 
> I think I've found the cause of the problem though. There's overlap 
> between the frequencies of the AC and BD IFs, so ORDER needs to be set 
> to <= 0, while it was set to the default of 1. Have changed this now and 
> BDF2AIPS is now running much slower (suggesting that it's actually doing 
> something!) and the log file indicates that individual scans are being 
> written. So I think it's all fine now.
> 
> Sorry for not checking more carefully before writing to you! The 
> information is all there in the help file to BDF2AIPS, but perhaps there 
> could be more emphasis on setting ORDER correctly.
> 
> As ever, many thanks!
> 
> Cheers,
> Nissim

Unfortunately, the time you want ORDER to be false is when the AC and BD 
overlap.  It is helpful to have the IFs of AC be a consecutively 
numbered set separate from those of BD (FRING has options that allow you 
to solve for an AC and BD delay but only when IFs 1-N are AC and N+1-2N 
are BD.  Changing the order prevents this option from working. 
Something else went wrong with the first BDFIn job I think.

I am CC'ing Bill Cotton - he may want to look at this.

Eric