[evlatests] EVLA Sensitivity and L, S, C, and X bands

[Rick Perley]

    I've utilized the 'receiver rotation' data to determine the cold-sky 
EVLA sensitivity.  In this experiment, we observed both a known 
calibrator, 3C147, and nearby blank sky.  The former is used to 
calibrate the data (and determine the polarization), the latter gives us 
the noise. 

    The attached plot shows the *antenna* SEFD:  2kTsys/Aeff, with the 
units in Jy.  The array sensitivity is computed by the following, for 
Stokes I (natural weighting):

    rms = SEFD/(2.correff.sqrt(BT))   (Jy)

    where B is the BW in Hz, and T is the integration time in seconds.  
'correff' is the correlator efficiency. 

    The SEFD is related to the system temperature and antenna 
illumination efficiency by (for our 25-meter antennas):

    SEFD = 5.62*Tsys/AntEff


    Of course, this is the theoretical thermal noise limit, and neglects 
a host of other factors... 

    The plotted values were determined by the histogram method:  
plotting the imaginary part of the visibility for all the baselines.  
For all bands, I used only those antennas which are fully EVLA compliant. 

    Comments:

    1) The L-band values are poorer than expected.  At the low frequency 
end (below 1300 MHz), it is clear that we have RFI issues for which the 
selected resolution (1 MHz) is insufficient to provide sufficient clear 
space.  It also seems clear that there is a significant difference in 
sensitivity between the antennas.  A higher spectral resolution survey 
is needed -- Emmanuel has the data in hand.  But even in the 'clear' 
band near 1450 MHz, the performance is ~25% worse than expected.  
Another look may be warranted -- I may be missing something. 
    2) The well-known difference in sensitivity between the upper and 
lower halves of C and S band is clearly present.  We suspect this is due 
to efficiency differences (illumination) (this was measured by Brigette 
last year, but needs to be confirmed with future work), and perhaps some 
difference in spillover (needs new absolute measurements we are 
preparing to make). 
    3) The X-band performance is excellent, except for the highest 250 
MHz -- I suspect these last two points have been skewed by the satellite 
downlinks which operate there. 

    Not shown is the comparison between the old X-band and new X-band in 
the frequency region of overlap.  The new systems are better by 15% in 
the center of the band (8.5 GHz), and much more at the edges. 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: LSCXSens.ps
Type: application/postscript
Size: 20785 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20110323/3628a6b9/attachment.ps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: LSCXSens.jpg
Type: image/jpeg
Size: 34872 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20110323/3628a6b9/attachment.jpg>