[daip] A question about UVPLT

[Eric Greisen]

Cheng-Yu Kuo wrote:
>              To whom it may concern:
> 
>           My name is Chengyu Kuo, a graduate student at University of 
> Virginia. I have a question about UVPLT.  I am trying to use UVPLT to 
> plot the visibility amplitude as a function of UV distance at a certain 
> parallactic angle. My model UV data only has a symmetric Gaussian point 
> source at the center, and the UVPLT (with bparm(1)=17 and rotate=45) 
> should only show a simplgaussian distribution. However, no matter what 
> P.A. I chose, it always looks a lot more complicated than a Gaussian
> (see the attached figure). On the other hand, if I use bparm=0, then 
> everything just looks fine. Could anyone help check what the problem is 

I am afraid that you have not described your problem correctly.  A point 
source cannot be a Gaussian since a point source has no width in any 
direction.  Its visibility function is a constant = the flux of the 
source.  The function 17 looks at the projection of the data along the 
specified position angle computed as

                      XZY(I,IP,LC) = ABS(RANDP(1+ILOCU)*SIN(RADPA) +
      *                  RANDP(1+ILOCV)*COS(RADPA)) * FRQMUL

         or to translate

                      X = abs (u * sin(pa) + v * cos(ps))

scaled appropriately.  As a baseline rotates arround at a constant 
radius and hence a constant visibility for a circular Gaussian, the 
value of X varies from that radius to zero.  Thus each value of 
visibility appears at multiple X up to that circular radius - making a 
plot just like you have sent us.

Eric Greisen