[evlatests] T304 attenuator changes

Michael Rupen mrupen at nrao.edu
Fri Jun 19 17:15:15 EDT 2009


> I reported earlier on the amplitude changes seen in the C-band data, on
> most antennas.  All of these changes occurred between scans, and most
> corresponded to a 1 dB change in power, strongly suggesting that the
> origin is the T304 attenuator chanaging.
>
> I later reported on apparent unstable C-band bandpass responses -- unlike
> at X-band where perfect stability is seen.  Closer inspection of these
> apparently unstable bandpasses shows that they changed shape between
> scans, and were always stable within any one 5-minute scan.  Indeed, most
> of the bandpass solutions look as good as the X-band ones -- it's a subset
> of the scans which show non-standard shapes.
>
> Ken suggested that I look for a correlation between the two phenomena
> reported above -- and it is certainly there.  Most (but not all) of the
> scans showing a change in bandpass shape are the ones with a change in
> power level, likely corresponding to a change in T304 attenuator.  For
> each power (attenuator) setting, there is a different bandpass shape.  If
> we take one of these as the 'right one', the other differs by a few
> percent, and has a distinctly sinusoidal shape with a long period -- over
> 100 MHz.  Not all antennas show this -- of the ten, perhaps one or two are
> much less affected.

This is very interesting, but seems to imply a very different behavior than
we saw in the June 2 & June 8 C band data.  I tried short timescale BPASSes
for the June 8 data for instance (1 minute I think), and didn't see any
significant improvement in the images.  Perhaps Rick could look at the
June 8 C band data and see whether the rest of us have been doing something
foolish?

Is there any sign that the bandpass moves between a small number of "states"? 
or is each scan different from all others?

As Ken says, each of the scans in a given band was taken at the same
frequency; I would hope that means, with the same LO settings, but the
Executor folks would have to confirm that.

Cheers,

      Michael



More information about the evlatests mailing list