[evlatests] pointing run, Sep 9 2005

[Ken Sowinski]

                        X Band EVLA Pointing
                         September 9, 2005

This was a good pointing run.  The weather was good and the VLA antennas
behaved well.  The poor elevation solutions seen in the C band pointing
run two days ago were not a problem last night.  The EVLA antennas were 
an hour late in getting started but what data there was is quite good.

1.  clunk for last 45 minutes
The changes we have seen in azimuth pointing after going OTT remain with 
us.  This time it occured only once about 45 minutes before the end of 
the pointing run.  Both antennas were affected and the magnitude of the
change was about 0.8 arc-min; perhaps a little more in 16 and a little 
less in 14.  This is the first clear example of this occurring simultaneously
to both antennas.  It is suggestive that the size of the clunk scales with
wavelength: at C band it was ~2 arc-min.  It certainly suggests software.  
If this were a Fortran program I would start hunting around for an 
uninitialized variable.

2.  low elevation effect, refraction?
There is a clear deterioration of the elevation solution at low elevations
which does not occur for VLA antennas.  This is approximately fit by
sag (cos(El)), but it is clearly not the correct model.  tan(z) does a
much better job in fitting the elevation pointing offsets and suggests
that the refraction correction is in error by about 15 percent; the antennas
pointing too low.  We should investigate feeding current weather into
the refraction calculation done in the calc wrapper.

3.  methodology
The last 45 minutes of data were excluded.  All OTT data was excluded.
For the azimuth parameters data from elevation as low as 8 degrees were
included in the solution.  In solving for the elevation parameters data
below 30 degrees were excluded.

4.  results of restricted solution
Az  A1    A2      A6   A7
14  0.0  -0.1    -0.2  0.2
16 -0.6   0.0    -0.9 -0.6

El  E1    E2      E5
14  0.2   0.2    -0.1
16  0.3   0.3     0.3

The data was good enough to attempt a solution for all the parameters.
The only believable result is a value for the perpendicularity term
(A5) of -0.5.  This is formally consistent with the historical value 
for this antenna of about -0.33.  I do not propose to include this
term until we have the other problems better under control.

5.  OTT
It is clear that when observing OTT the antennas do not point the same
was as VLA antennas.  Both the az and el solutions go awry when OTT 
measurements are included.

6.  Pointing model
I will update the X band collimation, and az offset, and tilt for both
antennas later today.  The elevation tilt will be provided to the
EVLA pointing model.