[evlatests] WideBand Data Works (mostly)
Rick Perley
rperley at nrao.edu
Tue May 4 14:07:42 EDT 2010
Just prior to the system shutdown for the 'swapathon', Ken took some
wideband data at C-band.
The piece I've looked at has four subbands from each of IF pairs AC
and BD. The tuning of these were arranged so that the eight total
subbands are all adjacent, from 4680 through 5576 MHz (center
frequencies). The subbands were each 128 MHz wide, with full
polarization, and 2 MHz spectral resolution.
AIPS doesn't know about subbands, so the data show up as eight
adjacent spectral windows (IFs) with a single FQID. (In other words,
the discrimination between the AC and BD analog systems is lost). This
can be important, as (for example), one cannot run a global delay
solution in FRING: in this case, the BD delays are very different than
the AC delays. Thus, when plotting the full spectrum via POSSM (the
eight subbands in a row), the amplitudes and phases connect very nicely
in the first four windows, and in the last four windows, but there is no
connection between subbands 4 and 5, since these originate from
different electronics.
Antennas 3, 6, 10, 14, 17 and 20 were not in the array.
Antennas 2, 4, 11 and 27 were completely flagged in all products.
I'll later download the data without flags attached to get a better idea
of what happened here.
Antenna 13 has no receiver.
All others fringed well, most of the time. However:
Antenna 12 gave only noise in subbands 1 through 4 (i.e. IFs A and
C), but fringed nicely in subbands 5 through 8 (IFs B and D).
Antenna 18 was exactly opposite: Fringes in IFs B and D, but not in
A and C. This antenna has a history of not tuning properly.
Antenna 26 gave only noise in LCP for subbands 2 and 8. (4808 MHz
in A/C, and 5576 in B/D).
Antennas 15 and 16 was flagged out, in RCP only, in subband 4 (5064
MHz in A/C). I presume this was a 'pure zero' problem, but need the
unflagged data to know for sure.
Delays looked very good (few nsec), except for: antenna 22 in LCP
in IFs C, and antenna 12 in LCP, in IF D.
Other than these, the data look very good. More details, after lunch.
More information about the evlatests
mailing list