[daip] Problem with CRSFRING

[Amy Mioduszewski]

Hi Tuomas,

Thank you so much for diagnosing the problem.  The behavior of DOCAL was changed 
because in all cases we could think of one should calibrate weights.  Since 
using calibrated weights should work, I will try to fix BLAVG.

Amy

Tuomas Savolainen wrote:
> Hi Amy,
> 
> sorry for the delayed answer; I've been traveling and I just got back to Bonn. I cc this also to Matt Lister, since he has noticed
> the same problem with CRSFRING.
> 
> I've been playing a little bit with BLAVG, and I think I found the cause of the problem. What has changed in BLAVG between versions
> 31DEC05 and 31DEC06 is the behavior of the DOCALIB adverb. The diff output of the fortran code looks like this:
> 
> diff /homes/AIPS/31DEC05/APL/PGM/NOTST/BLAVG.FOR /homes/AIPS/31DEC06/APL/PGM/NOTST/BLAVG.FOR
> 32c32
> < C;  Copyright (C) 1995-1996, 1999-2000, 2004
> ---
>> C;  Copyright (C) 1995-1996, 1999-2000, 2004, 2006
> 206c206
> <       DOWTCL = XDOCAL.GE.1.5
> ---
>>       DOWTCL = DOCAL .AND. (XDOCAL.LE.99.0)
> 
> 
> Now, in the old version weights are calibrated if DOCAL > 1.5, while in the new version this occurs if DOCAL is positive and smaller
> than 99. (Do you know why this change has been made?) CRSFRING puts DOCAL = 1, which in the old version means no calibration of the
> weights, but in the new version it tries to calibrate them. This seems to cause the problem. When I put DOCAL = 100 in the new
> version, I get the correct output from BLAVG. So, changing CRSFRING so that DOCAL = 100 should take care of the problem.
> 
> Cheers,
> Tuomas
> 
> Amy Mioduszewski wrote:
>> Hi Tuomas,
>>
>> I am back from vacation and looking at CRSFRING.  So when I take the
>> data you split out from me and run FRING, I get the "ERROR: NO VALID
>> SOLUTIONS FOUND", for 07, 06 and 05.  I then looked at the data you sent
>> me, it is only 6 visibilities and has all zeros for the amplitudes and
>> phases (and U, V, W) the only thing that seems not to be zero is the
>> weights.  Is that what you see? Since I do not see a difference between
>> 05 and the others then maybe the bug is in BLAVG not FRING.
>>
>> Cheers,
>>
>> Amy
>>
> 
>