[daip] FITLD problem

[Chris Flatters]

> I am processing a 4 epoch VLBA experiment, correlated in 1995 and recorded
> in a standard mode (128-4-2).  No subarrays were used during observation.
> 
> In NEW & TST, both on Alpha and Solaris systems, we receive warnings from
> FITLD about out-of-order data or detection of a subarray condition. Jim
> Ulvestad has confirmed that this occurs.  (He has used a copy of our
> correlator tapes and loaded the data in Socorro.  His MSGs are included
> below.)
> 
> I believe strongly that the problem is a FITLD problem and not a
> correlator problem.  I obtain the FITLD warning messages when I read data
> for an epoch that I loaded witout warning messages in spring 1999.
> The processing used 15OCT98 or 15APR99 aips.  I cannot be sure.
> 
> I have run USUBA on affected  data and find no subarrays present.

The information given suggests that some of the data in the FITS file is not
in strict time order. Although FITLD does not distinguish between out-of-order
data and subarray conditions in its summary message it does issue additional
messages if it thinks that it has found a subarraying condition.

It is almost certain that the disordering is localized to a short time
interval in the data so MSORT should be very effective in fixing it. Try

TASK = 'MSORT'
GETN n
GETO n			(sort in place)
SORT = 'TB'
GO

MSORT will whine about tables (usual when sorting in place and harmless) but 
should run very quickly. Run INDXR after MSORT. USUBA isn't needed in this
case.

	Chris Flatters
	cflatter at nrao.edu