[daip] CVEL

[Eric Greisen]

Anna Bartkiewicz wrote:
> Dear All,
> 
>   Please, let me know if there are known problems with CVEL after Change of 
> Terrestrial Reference Frame.
> I am using Midnight Job AIPS and the spectra are 
> decreasing and changing theis shapes after CVEL (viewing by POSSM). It is 
> also difficult to get an image.
> That happens for VLBA data (12 GHz) and EVN data (6.7 GHz) in my case.
> 
> Is the solution to reduce all steps for spectral line data using AIPS 2010 
> and run cvel only using AIPS 2008?
> 
> Sincerely yours,
> 
> Anna
> 
> ----------------------------------------------------------------
> Dr Anna Bartkiewicz (Niezurawska)
> Torun Centre for Astronomy
> Nicolaus Copernicus University
> Gagarina 11
> 87-100 Torun
> Poland         annan at astro.uni.torun.pl    phone: +48 56 6113040
>

We do not distribute code with known errors.

The change to CVEL for antenna coordinates was minimal.  Previously the 
VLA was right handed and - inside aips - vlbi was left handed.  The old 
code said IF (VLA) ANTY(I) = -ANTY(I).  Now all are right handed and so 
the IF (VLA) portion of that line was removed.  All essential 
subroutines that compute and perform the doppler shift are unchanged 
from 31DEC08.  I would be concerned that the coordinates of some 
antennas may not have been converted to a proper right handed system - 
our code to force that depends on knowing where each antenna - by its 
station code in the AN file - is located.  It might not do the right 
thing for a bunch of stations that are not well known to us.  It is very 
clear that, if you take data touched by 31DEC09 or 31DEC10, and use any 
31DEC08 task that cares where antennas are, that you should not get 
correct results.

If you would like to help debug this, you could load your data from raw 
input files into 31DEC08 and process them only in 31DEC08 including 
CVEL.  If that works more to your satisfaction, lets us know.  I am 
going to be on vacation starting in a few days but perhaps we can look 
at things before then.  CVEL had many changes having to do with making 
it record what it did so that a second CVEL on the output of the first 
with the same parameters would do nothing more.  I did do quite a bit of 
debugging - but possibly not enough.

Eric Greisen