[daip] SPLAT file, new problems

[Eric Greisen]

Lynn D. Matthews writes:
 > I now think that CVEL is actually shifting the VLA antenna, just not
 > always reporting it.
 > 
 > Originally I did not tell splat to pass the autocorrelation data (APARM(5)
 > = 0). To better answer your question, I re-ran splat with APARM(5)=1.
 > When I run CVEL on this new splat file, the task reports average shifts
 > for all 11 antennas. In contrast, when I run CVEL on the splat file
 > created with APARM(5)=0, CVEL reports shifts only for the first 10
 > antennas (VLA being antenna 11).
 > 
 > I plotted some x-correlation spectra with antenna 11 from both CVEL
 > outputs, and they both show identical shifts compared with the original
 > data. Below I append a sample of the messages from CVEL for the two cases.

Solved.

The VLBA correlator shift is of the first antenna in an antenna pair.
 With no cross-correlations the highest antenna pair was 10-11 and so
it was 10 shifting and never 11.  With auto-correlations there is also
11-11 and some shifts of 11.

N.B. with EVN it is the 2nd antenna in the pair.

If you look at the numbers you sent - the low antenna numbers have
almost identical average shifts in both runs, but antennas 9 and 10
have different averages (slightly) because they are counted more.

Thanks,

Eric Greisen