[evlatests] No Referenced pointing for VLA antennas

[Ken Sowinski]

On Wed, 29 Aug 2007, Rick Perley wrote:

>    Major Result:
>
>    ***  The solutions are not being applied, or not being applied
> correctly, to the VLA antennas.  In most cases, VLA antenna amplitudes
> were the same with and without the referenced pointing solutions being
> applied.  Furthermore, for those VLA antennas with large pointing
> errors, the amplitude solutions showed the types of temporal modulation
> expected when the source is on the side of the team.
>    One can make a case that the VLA antennas are being corrected when
> the offset to be applied is small.  But when large (0.5 arcminutes or
> so), it is clear that the applied correction (if any) is not right.
>          By contrast, the EVLA antennas showed in every case the
> amplitude gain expected with application of the pointing corrections.
>
>    This behavior was seen both at K and Q bands.

I would like to argue that telcal rejected the solutions when the
errors were too large and never multicast them out.  This is inspired
by the fact that it was K band pointing and some of the errors were
quite large.  However I can find no real evidence that this really
happened.  Another test is in progress with pointing done at X band
to test this.

>    Minor Results:
>
>    1)  Antenna 14 clearly has something wrong with its pointing model.
> K-band referenced pointing worked well when applied to K-band
> observation, but actually worsened the amplitude at Q-band, when applied.

Antenna 14 (and 20) seem to have been bumped and the collimations
have not all been straightened out yet.

>    2) Antenna 18, in IF#1, is showing significant phase jumps within a
> scan.  The jump is 50 degrees or so, and is the same in the two
> polarizations (A and C).  IF#2 is not affected.  The behavior is the
> same at both K and Q bands.

This sounds like the problem that is usually fixed by changing the
length of the cable which carries one of the timing signals to the L302.

>    3) Most of the VLA antennas showed a significant visibility drop for
> the last 5 seconds of the run (at K-band).  None of the EVLA antennas
> showed this.
>
>    4) Antenna 14 was slow to stabilize its phase at the beginning of
> the run.  Only this antenna showed such an effect.  I had to remove
> about two minutes' worth of data.
>
>    5) A 'dummy' initial scan was put in, to see if the initialization
> troubles occasionally reported showed up.  The first scan was in fact a
> good one (other than antenna 14, as noted above).

Another 'first scan problem' data point.