[daip] FRING oddity.

[Craig Walker]

I have run into a situation where FRING generates CL entries for the
rates that are excessively small.  I have not been able to understand
it.  But here are some details.

The file is a 327 MHz VLBI observation of CYGA where I have shifted the
phase center in two different directions using CLCOR to look at the two
lobes.  Yes, I am aware of the SU table implications.  When I did the
second shift, I deleted the current SU table and replaced it with one
from another version of the file that had an unmodified version.  I
don't know if this hoop-jumping is related to the problem.

Now when I run FRING on this file, with a variety of parameters, I get
rates in the message print that look reasonable, but that show up in the
CL table at around E-20 rather than near the usual E-12.  Earlier
fring's had work normally.  For a while I thought I could get a normal
result by going back to a CL table from before the second CLCOR shift.
But that doesn't seem to be true now.

Note that I am fitting for one baseline only - the only one that showed
any evidence for detections on more general fits.

I tried to make a stripped down version of the data set on one of my
normal disks for testing.  The the problem did not recurr with that data
set (Murphy strikes again).

So the data set that produced the offending result is eolus disk 6 user
262 catno 2:
AIPS 1:    2  262 CYGA 9NOV03 .AVS16 .    2 UV 07-APR-2004 17:29:47
You have to be logged on to eolus to see it because it is my external
firewire disk and is not exported.

A sample fring is in VNUMBER 10.

There is sort of a real time log of a variety of tests:

I just reran the FRING.  SN 7 was with GAINUSE 4 and a small time range.
It has the E-20 problem in the rates.  SN 8 is with GAINUSE 3 and the
same time range.  The E-20 problem does exist for this one (not expected
from what I thought I saw yesterday).
So I tried with GAINUSE 2 (result of "manual pcal").  It happens there
too (SN 9).  So I turned off the RR/LL average (GAINUSE 4, SN 10).
Still bad.  I turned off combine IF.  Still bad.
Allow all antennas.  Aborted with ZABORS: signal 11 received.
Try DOCALIB -1.  Still aborts.
vget fring - try to return to one that used to work.
Indeed it does run.  Made SN 13 - bad rates.
Try antennas 0; aparm 4,0,0,0,0,2,4,0,1
Aborted.
Antenna 5,9.  Runs ok despite having min no. antennas aparm(1)= 4!
Bad rates.
Basically everything I'm trying with this data set is giving bad
rates.  Maybe the earlier test supposedly with CL 3 for gain was not
done right somehow.  Could it be the SU table?

I'm not pretty confused.  Tomorrow, I will probably reload the data and
start over.  I'll leave the offending data set in place.  I may have
messed up a table, but what is happening (excessively low fringe rate in
CL table) should not be happening regardless (not to mention the aborts).


Cheers,

Craig



---------------------------------------------------------------------
     R. Craig Walker            Array Operations Center
     cwalker at nrao.edu           National Radio Astronomy Observatory
     Phone  505 835 7247        P. O. Box O
     Fax    505 835 7027        Socorro NM 87801   USA
---------------------------------------------------------------------