[daip] CL ver 1 geodelay calculation problem

[Eric Greisen]

Walter Brisken wrote:
> For VLBI files, the CL tables have a column called GEODELAY which contains 
> a tabulated listing of the delay model at each CL interval.  I have an 
> example FITS-IDI file, straight off the VLBA hardware correlator, where 
> FITLD results in this column containing nothing but zeros until row 11149, 
> where reasonable values start.  I have confirmed that the two model 
> tables, IM and MC, have valid values for the duration of the file. 
> Removing NX1 and CL1 and running INDXR results in a CL1 table with zero 
> GEODELAY for the entire length of the file.  I have tested FITLD with 
> values of CLINT ranging from 0.016667 to 1 minute, always with similar 
> results.
> 
> This problem exists both in NEW and TST versions and is consistently 
> reproducible.  The file I found that demostrates this problem is:
> 
> /home/parallax/7200001.FITS
> 
> Thanks for any help on this matter!

This was a long-standing bug in FITLD having to do with more than one UV 
table in one input file.  The time range for the final table processing 
was set by the last UV table.  In testing I also found another bug that 
could cause the IM and MC table to fail to provide delays when there are 
more than one input files and they are not being concatenated.  The 
routine depended on the IM file changing the number of rows, which it 
does not do between segments of VLBA data.  This forced new call 
sequences and numerous changes.

INDXR will fill the values only if you set CPARM to request them.

Eric Greisen