[evlatests] Pointing Troubles

[Ken Sowinski]

On Mon, 15 May 2006, Rick Perley wrote:

>   I reviewed carefully the results at C and X bands only.  These were 
> sufficient to allow me
> to conclude that there is no discernible pattern to the results.

Rick provided a long list of curious behavior.

There is more than one problem here.  The easy one to explain
is that telcal pointing results for antenna 13 are not always
correct.  IFs BD were dead and the pointing solver did not
reliably reject the BD data in deriving pointing offsets.  This
is easily fixed.

Setting aside the antenna 13 results, I looked at the time series
of measured pointing offsets for antenna 14.  I was pleased to
see that the Modcomp and telcal pointing solvers gave the same
results in all cases.  There is some suggestion that pointing
corrections are applied using measurements from an earlier scan
than intended.  I don't see a clear. long term  pattern of this,
but there are short sequences that can be explained this way.
This can happen if the python script is wrong, if telcal is slow
in sending results, or if the executor takes old pointing results
for current results.  I suppose there are many other ways as well
as other patterns that might be discerned.  It is probably worth
a little investigation to be sure that we are applying the pointing
results that we think we are.

It is not likely that the EVLA antennas are moving at the wrong
time.  Both pointing solvers give essentially the same answers
and a careful examination of amplitudes on a VLA-EVLA baseline
shows no himt of errant motion.