[daip] A dbcon question

[Eric Greisen]

Mark Reid wrote:
> Eric,
> 
>   I was combining 2 EVLA data sets taken 4 days apart and DBCON reports 
> the following:
> 
> DBCON1: Task DBCON  (release of 31DEC11) begins
> DBCON1: New reference day is 20110619
> DBCON1: Scaling u,v,w of data set 2 by  9.9925826E-01
> DBCON1: Scaling u,v,w of data set 1 by  9.9925827E-01
> 
> I don't know why it is scaling the (u,v)'s by 0.999 and I am concerned 
> with the effects on source positions.
> 
> I'm trying to measure the positions of red giant stars (via their SiO 
> masers) within about 20" of Sgr A*.   I do this by calibrating on Sgr A* 
> and making a big image.  If this scaling decreases the angular offset 
> between a star and Sgr A* by 0.999 that would be 20 mas for a 20" 
> offset.  That is a very large amount, since we can measure relative 
> positions to ~1 mas.
> 
> Can you fill me in on what's going on?

DBCON wants to guarantee that it has put the data sets together 
correctly so it shifts each to reference channel 1.0 - that requires a 
rescale of u,v,w.

The need is obvious (sort of).  Consider an old VLA data set with 7 
channels.  Run UVCOP to split each channel out.  What do you get - U,V,W 
in each data set are identical point by point and the header shows 
reference channels like -2, -3, etc on the 1-pixel freq axis.  After you 
DBCON these you have only one ref pixel so you must scale the u,v,w top 
match the actual freq of the data sets you are combining.

Trust me - this part of the problem is straightforward and okay.  Other 
parts I am less certain of.

Eric