[evlatests] Tests of 'Expanded OSRO' Modes

[Rick Perley]

    A short test was made yesterday to check the robustness of the EVLA 
when observing in the 'Expanded OSRO' mode -- 8 subbands from each IF, 
spanning a total of 2048 MHz.  Nine different frequency tunings, 
spanning the lower 5 frequency bands were employed.  Each observation 
was for one minute.  X-band referenced pointing (using normal OSRO1 
mode) was utilized in the middle of the run.  The cycle was executed 
twice.  The source observed was 3C345, which was high in the NW at the 
time. 

    In more detail, the sequence of observations was as follows:

    1) A Dummy scan in OSRO1 mode
    2) L-band, with AC centered at 1412 MHz, and BD at 1690 MHz.  This 
was to check the sensitivities beyond the nominal 1 -- 2 GHz bandpass.
    3) S-band:  centered at 2488 and 3512 MHz
    4) C-band Lo:  centered at 4464 and 5488 MHz
    5) C-band Hi:  centered at 6512 and 7536 MHz
    6) X-band Referenced Pointing, using 8100 and 8700 MHz in OSRO1 mode.
    7) X-band Lo:  centered at 8464 and 9488 MHz
    8) X-band Hi:  centered at 10512 and 11536 MHz
    9) Ku-band Lo:  centered at 12440 and 13464 MHz
    10) Ku-band Mid:  centered at 14488 and 15512 Mhz
    11)  Ku-band Hi:  centered at 16536 and 17560 MHz.

    Each wide-band observation was for 1 minute, except L-band, where I 
inadvertently set it to 2 minutes. 
    The cycle (except the initial dummy scan) was repeated twice. 

    Basic Result:

    A spectacular success!  With only a very few exceptions (noted 
below) all antennas properly equipped gave strong fringes at every band 
and every frequency.  Setup times are typically 10 seconds or less, 
except at L and S bands, which are more typically 30 seconds. 

    Details:
  
    1) As expected, the scan preceding the X-band pointing, and the last 
X-band observation (which precedes the first Ku-band) both lost 18 
seconds of data at the end. 
    2) Very little RFI is seen (except at L-band) -- the 'RFI Report' 
will be given separately.

    The data were partially calibrated, using the following method:

    a) Initial flagging, to determine the 'setup time', and identify 
which antennas did not fringe. 
    b) FRING was run to find delays.  These were applied with CLCOR.  
(There is a problem here with Eric's new capability to solve 
independently for AC and BD IFs, when both are present -- I'll report on 
this separately to the relevant people). 
    c) BPASS was run -- *** without normalization*** to check on 
bandpass shapes and power levels.   A solution for each scan was made.  
These have been printed out for later perusal. 

    Results, by band:

    A)  L-Band.  All antennas with receivers fringed on both IFs for 
both scans.  The settle time was about 25 seconds. 
        The low-frequency rolloff is quite obvious in the 'AC' setting 
(which spans 900 to 1924 MHz) -- but is largely absent for the BD tuning 
(spanning 1178 through 2202 MHz).  I did not expect this result (and 
must consult with others to understand it). 
        The 6 antennas equipped with the final polarizers are obviously 
better (flatter spectra) in the 950 to 1100 MHz frequency range. 
        There is a clear cutoff of the sky signal below 950 MHz (for 
both old and new polarizers) -- this is then the lowest frequency for 
the EVLA's L-band system. 
         At the high frequency side (BD IFs), the 'new polarizer' 
antennas show a much strong rolloff in spectral sensitivity, starting at 
2100 MHz.  (The old polarizer antennas are good to 2175 MHz). 
        Antenna 15 has a bizarre bandpass shape below 1200 MHz, and its 
power is low by a 6 dB or so. 

    B) S-Band.  Nine antennas fringed, with no tuning failures.  
(Antenna 22 still has its receiver out). The settle time was about 25 
seconds. 
         Antenna 14 is low in power on the AC side by about 5 dB.

    C) C-Band:  All antennas with receivers fringed.  On the second 
scan, antenna 2, in IF 'AC', failed to fringe -- presumably a failure to 
tune.  This failure occurred for both frequency setups within this 
band.  The settle time is about 20 seconds for switched from S to C, and 
15 seconds switching frequency within the band (C-Lo to C-Hi).
         Antenna 27 is low in fringe power by about 10 dB, in IFs 'AC' 
only.  (This could mean either the total power is low, or the system 
temperature is very high). 

    D)  X-Band.  All antennas fringed at the low tuning.  Setup time was 
less than 10 seconds for both tunings. 
          At the high frequency tuning (10.000 to 12.048 GHz), antennas 
7, 12, 13, and 14 gave useable fringes.  This is remarkable, as only two 
antennas (7 and 14) have new wideband receivers!  Antennas 12 and 13 do 
fringe, although weakly, (especially at the high frequency end), and 
have ugly bandpasses. 
           Most of the old-style antennas give useable fringes up to 
about 9.5 GHz. 
           Antenna 16 gives low fringe power, by a factor of ~10 dB in 
RCP, and 20 dB in LCP. 
           Antenna 22 is low in fringe power by about 5 dB.  
           Antenna 27 is low in fringe power in IFs A and C, by about 10 
dB. 

    E)  Ku-Band.  Six antennas are now equipped, and all fringe very 
nicely, except for a tuning failure noted below.  Setup time was 15 
seconds for the first tuning (the previous scan was at Xband), and less 
than 10 seconds for the subsequent two tunings. 
            For all three Ku-band tunings, antenna 20 failed to tune in 
IFs 'BD' for the first scan, and antenna 10 failed to tune in IFs 'BD' 
for the second scan. 
            For all three tunings:  antenna 14 was low in fringe power 
(and in total power) by 7 dB in 'AC', and 12 dB in 'BD'.  Antenna 20 is 
high in power by 5 dB in 'BD'.  Antenna 22 is low in power by 15 dB in 
IFs 'BD'.  Antenna 24 is low in fringe power by ~5dB in IF 'BD'.