[evlatests] Monday stress test results
Rick Perley
rperley at nrao.edu
Mon Nov 5 22:48:59 EST 2007
I ran a stress test Monday afternoon. This one was set up in the
following way, all in continuum at 50 MHz (except P-band, at 3.125 Mhz):
A) Two consecutive X-band observations, to check for bad first scan
B) A reference pointing, followed by another X-band observation, to
check reference pointing application.
C) K, then Q bands, with reference pointing applied.
D) Three consecutive L-band observations, to check for peculiar
phase jumps upon repeating observations.
E) One P-band observation
F) One C-band observation
G) Alternating between L and C bands, in hopes of inducing the 'lost
header' problem.
General Results:
0) Overall, extremely good data. Phases were excellent at all
bands, except for VLA antenna 12A and C, where strange 'wanderings' are
still being noted.
1) The initial pair of X-band observations were identical -- no
changes in phase or amplitude between them. No 'dummy' scan was needed
here.
2) Referenced pointing appears to work as expected. Amplitudes went
up after application of the corrections, especially for antennas 13 and
25 (more on these later).
3) All is well at K and Q bands.
4) The total of 5 L-band scans did show a 'global' phase jump for
the middle one , but on IFs 'B' and 'D' only -- this represents a change
in behavior! However, the two L-band observations at the end (which had
C-band observations in between) showed to phase jump.
Also -- EVLA antenna 25 had a -185 degree phase jump on *all* IFs,
for this same scan!
5) Each time we change band, the next observation has an unflagged
single record of noise.
6) The last 8 seconds of the last scan was rubbish, with an abrupt
change in phase and amplitude. This must have something to do with the
next observation's parameters getting mixed in with the current
observation?
7) The usual 'short shallow drops' were seen at all bands on most
EVLA antennas. The usual characteristics are repeated.
Specific Issues Noted:
1) Antennas 13 and 25 are *still* not pointing correctly at high
frequencies. The referenced pointing gave a large correction for them
at X-band, but the subsequent K and Q band observations gave zero
amplitudes for both antennas.
2) At P-Band:
a) 26 still gives no fringes whatever in IFs B and D.
b) Antenna 11 sensitivity is low by a factor of ~4 (in
efficiency/Tsys).
c) Antenna 21 sensitivity is low by a factor of ~4.
Presumably this is due to the filter on that antenna.
d) Antenna 23, IFs A and C, is low in sensitivity by about a
factor of 1.5
e) Antenna 24, IF 'A' only, is low in sensitivity by nearly
an order of magnitude
f) Antenna 26, IF 'C', is low in sensitivity by nearly two
orders of magnitude. 26B and D gives no fringes, but 26A is normal.
3) At L-band:
a) 16B sensitivity low by about a factor of 2.
b) No fringes from antenna 12. No idea why.
4) At C-Band:
a) There were large changes in correlation coefficient in IF
'D' for evla antennas 16, 17, and 24. Also in 14C. No idea why. All
others were stable, and normal.
5) At X-Band:
a) EVLA antenna 14C and D showed large drops in correlation
coefficient, which were reflected by measured increases in Tsys. What
can make the receiver do this? The changes were very abrupt (1 second
or less).
b) EVLA antenna 25 low in sensitivity by about a factor of 3
-- all of which is likely due to Tsys being too high by about the same
factor.
6) At K-Band:
a) Antenna 11 low in sensitivity by about 20%
b) Antenna 19 low in sensitivity by about a factor of 2.
7) At Q-Band:
a) Antnna 18 low in sensitivity by about a factor of 50%
b) VLA antenna 20 low in sensitivity by about a factor of 2.5.
c) VLA antenna 5 showed a monotonic drop in sensitivity in all
IFs, by about a factor of 2, over the 1-minute observation.
More information about the evlatests
mailing list