[evlatests] C-Band Sensitivity Issues

[Rick Perley]

    The data shown by Emmanuel yesterday morning indicated that many or 
most EVLA antennas are performing far short of requirements at C-band, 
and there is some evidence that the more recent antennas are less 
sensitivity that the older ones.
    I took some data yesterday afternoon at L, C and X bands to 
investigate this issue a bit.  The primary goals were to see if the 
spread in sensitivity could be correlated with system temperature, and 
if any correlations could be seen in performance with band. 

    A)  X-Band.  This band is useful as a diagnostic, as the receivers 
are unchanged between EVLA and VLA, and there is minimal electronics. 
    The EVLA sensitivity requirement is for the antenna SEFD to be 300 
Jy (or, Tsys/effic. = 53K). 
    The data from yesterday show that eleven antennas (9 EVLA, 2 VLA) 
antennas are performing near, or better than, this level.  Recent EVLA 
antennas numbered 1, 2, and 4 are in this group.  Eight antennas (5 VLA, 
3 EVLA) are high by about 10% (not bad) -- recent EVLA antenna #1 is in 
this group.  Five antennas are high by 20 to 30% (3 VLA, 2 EVLA). 
    Two antennas are particularly poor:
    17 is high by a factor of about two on all four IFs.   The Tsys is 
claimed to be 130K, so this is the likely cause.
    25 is high by a factor of nearly three.  Tsys is claimed to be 25K 
-- so the receiver is likely o.k. 
   
    Overall -- EVLA performance is better than VLA, and although the 
mean sensitivity is a bit worse than spec (320 Jy, cf. requirement of 
300 Jy), there's no reason for alarm. 

    B) C-Band.  Here the EVLA requirement for SEFD is 245 Jy (this is a 
tough spec!), equivalent to Tsys/effic = 44K. 

    As noted yesterday, *none* of the EVLA antennas meet this spec -- 
not even close.  The best four antennas are all 'old' EVLA antennas (13, 
14, 16 and 18) -- their mean SEFD is 310 Jy.    The average SEFD over 
all the EVLA antennas (excluding two which are clearly discrepant) is 
about 380 Jy.  Pretty poor. 
    The mean SEFD amongst VLA antennas is definitely worse than that of 
the EVLA -- about 440 Jy, with a remarkable spread, from about 350 Jy 
(for 27, 20, and 22) to 540 Jy (6, 8, 10, and 15).  There is good 
evidence that most of the variation amongst the VLA antennas is due to 
system temperature, as there is a good correlation between the claimed 
Tsys and the SEFD values. 

    I have searched for a similar correlation between SEFD and Tsys for 
EVLA antennas -- and although I believe there is one, it's crystal clear 
that the recorded values of Tsys are only approximately correct:
    a) The four best EVLA antennas all have  believable Tsys values -- 
about 30K. 
    b) All other EVLA antennas have recorded Tsys values which are far 
too high -- typically 80 to 100 K.  But these are not believable numbers 
-- (the SEFDs are not THAT bad), and there is little correlation between 
the Tsys values and the SEFDs. 

    It is probable that the existing Tcal values are incorrect, leading 
to incorrect Tsys.    It is difficult to diagnose problems in 
sensitivity without believable Tsys values, so I think it important that 
we get modern, correct values of the Tcals into the EVLA parameters 
database. 

    Three antennas are particularly poor, and could use quick attention:

    12:  SEFD = 700,  with a (believable) Tsys = 75.  (
     19:  SEFD = 900 (four times too high!), with a claimed Tsys of 190 
-- this could be close to correct.
    24:  SEFD = 600, with a claimed Tsys of 50 (better than most!).