[evlatests] interband collimation corrections

[Vivek Dhawan]

Sorry - this initially went to only a subset of evlatests.
------------------------------

I was talking about the subreflectors all being in slightly
different places, variations of 3-5mm. The corresponding
equivalent displacements of feeds in the focal plane is proportional to
magnification M, so about 1-2 inches on the VLA. (The axial
or focus position shift is  ~M*M/2)

The interband colimations are not affected - a shift (i.e. across the
optic axis) of the subreflector makes all the bands change colimation by
the same amount. Similarly, the sag under gravity affects all bands the
same, but is accounted for as belonging to the antenna, I believe.
The different subreflector positions can be seen in the different
antennas (Q-band) peaking in gain at different elevations.

However, a change in subreflector tilt moves the colimations by the
same magnitude at all bands, but the vector points around a circle as
the subreflector is rotated around the bands. This does affect the
interband colimations. So also a change in FRM rotation encoder offset.
These are easiest to see if pointing offsets changes from many bands are
plotted on a circle each at the position of the corresponding feed,
right after a change in the subreflector position.

I guess the question I have is, was holography used to set the
subreflectors of antenna 14 and 16 (and others) when they were VLA 
antennas, before conversion to evla?


V.


Peter Napier wrote:
> I thought Vivek made the point that he expects subreflector alignment to 
> cause collimation errors. If that is the case this should be seen on VLA 
> antennas as well as EVLA.
> Peter
> 
> Walter Brisken wrote:
> 
>>
>> Hi Ken,
>>
>> Antenna 16 below looks more like I'd expect than 14 does.  Based on my 
>> C-band simulations I would expect more like a 2' collimation term for 
>> L-band and undetectable terms for C and K.  Does anyone understand the 
>> cause of large collimation at bands where we expect zero?  What 
>> scatter in values is typical for VLA antennas?  Putting this into the 
>> context of the feed position debate of last week, these 
>> antenna-to-antenna differences would probably dominate over any 
>> effects caused by systemmatic feed position offsets.
>>
>>> RSS Error in arc-minutes relative to X band for:
>>>     14     16
>>> L    2.4    1.1
>>> C    1.9    0.3
>>> K    1.0    0.5
>>>
>>> These errors have probably been the state of the antennas since,
>>> at least, the last pointing run two weeks or so ago.
>>>
>>
> 
>