[alma-config] Near-in sidelobes

[John Conway]

On Thu, 8 Jun 2000, [Leonia Kogan] wrote:

> 
> John,
> 
> I have played a little bit with the spiral configuration you gave me a
> year ago.  I would like to say since the beginning that the following
> conclusions are valid only for this spiral.
> 
> So I built the two dimension beam using UVCON=>IMAGR=>KNTR.
> The sinthesyse beam has elliptical shape which is rather far from circle.
> The ration of the ellipse ases is ~2.
> 
> The sinthesize beam has wide wings. 
> 
> 

 Oops something is wrong there - even my spiral of a year ago 
had a circularly symmetric uv coverage - what  declination 
did you use for the simulation, maybe you mistyped +23 instead
of 23. 

I think the wide wings in that version was due to the particular 
spiral parmeters I used  (radius range 8 azimuth, 1.5turns),
the resulting radial density distribution was not gaussian
but mmore centrally condesnsed. Thats not true of the version
I refered to on my website. 


> 
> 
> Now some comments about near side lobes.
> 
> You know of cource that the level of the near side lobes depends on the 
> element density as the function of the distance from the array center.
> 
> The smoother this function goes down the lower near side lobes.
> 
> So your spiral configuration may have low near side lobes because of 
> the tapering of the element density. 
> 

Right, for a two ring configuration, you get a wedding cake
radial density function, with the supperposition of the 
lambda beam for each layer.

> It is simple to get the same result for donut type of cofiguration if the 
> center of the donut will be allowed for the element placement.
> 

I don't know if understand this. But I think are philosphies are 
not so different tou have 2 rings - a spiral can be thought of 
as about 10 concentric triangles, about 4 occupied in any
configuration  I start with a continuous pad
distribution and then form discrete arrays, you optimise each 
array semi-seperately.
  


> But a configuration with the strong condensation of the elements at the 
> center may have problems:
> 
> 1. The size of the array should be bigger. 

Granted the maximum baseline is longer, however the median and
mean baseline length are about the same (the have to be if 
each array gives the same resolution). For the largest 
array the zoom spiral evolves into a configuration with
antennas around the perimeter to maximise the resolution from
a given amount of real estate. 


> 2. The fitting to the topography can be a problem
> 

I don't think it is that much if a problem,  I will post a topography 
fitted version next week I hope. 

> 
> Leonia
>