[evlatests] EVLA Sensitivity, from 1 to 12 GHz
Rick Perley
rperley at nrao.edu
Wed Mar 9 20:50:43 EST 2011
The 'receiver rotation trick' test allows us to get an honest
estimate of our true array sensitivity. I'm observing 3C147 -- a strong
source of know flux density, and some blank sky nearby.
The data from last night has been calibrated. (There is an
interesting problem in that the 3C147 models are slightly wrong for this
configuration, and caused a miscalibration of the long spacings by a few
percent ... This is an argument for getting better models soon).
Noise histograms were made for all subbands that I observed (a total
of 96!), using only channels apparently free of RFI, and antennas that
are fully EVLA compliant. These histograms were converted to SEFD, with
some of the results given below.
L-Band:
The lowest 4 subbands (982 through 1244 MHz) are greatly affected by
RFI. The histograms were strikingly non-gaussian, so the results are
not meaningful until we understand the origin. Certainly a higher
spectral resolution is needed (I used 1 MHz). Emmanuel will soon be
doing this.
Freq SEFD
------------------------
1270 633
1350 483
1400 450
1470 454
1510 478
1600 553
1660 551
1720 572
1790 502
1850 415
1920 415
1970 553
--------------------
S-Band
Like C-band, the sensitivity in the lower half (2 -- 3 GHz) is worse
than the upper half (3 -- 4).
SEFDs are near 400 from 2.05 through 2780 MHz, then drop to 340 at
3160 MHz, and to 300 Jy at 3540 MHz, rising a bit to 340 Jy by 3800 MHz.
C-Band
SEFDs decline from ~350 Jy at the low frequency end (4.1 GHz) to about
290 Jy through to 6 GHz, above which it drops to 230 -- 250 Jy right up
to 8 GHz.
X-Band
We have now 5 new-style receivers in, allowing full band sensitivity
measurements for the first time. The results are good.
SEFDs are remarkably uniform throughout the entire band: between
235 and 260 Jy, except at the extreme ends. If we assume the efficiency
is 60%, this sensitivity corresponds to an average Tsys of about 26K.
A more report will follow when the rest of the data are in.
More information about the evlatests
mailing list