[evlatests] C-K test results from yesterday evening ...

[Rick Perley]

    A 30 minute test was done last night, apparently alternating between 
C and K bands on the good calibrator 3C286.  Scans at each band were 
apparently 3 minutes long, and it looks like there was a 'reset' in the 
middle of each of these. 

    The major goal for the test was to see if we still have big phase 
jumps on any antennas.  We do -- but only on antenna 21, in IFs BD 
only.  Two 'jumps' were seen at each band, in each case in the middle of 
a 3-minute scan.  I note also that 21 had an 8 degree change in phase in 
R-L between two scan, on the BD IFs only. 

    Beyond this, perhaps the most interesting result from the test is 
the clear evidence that the data were taken in the middle of a  huge 
thunderstorm.  (It would be useful to confirm this from an actual 
sighting).  But the evidence is clear -- and gives us very useful 
information on the absorption resulting from such an event. 
    All antenna show a notable drop in amplitude, at both bands, with 
K-band being much much worse.  In our new era of real cross-power (there 
being no division by total power to obscure what is really happening), 
and presuming that 'set and remember' is both working and is not 
influenced by external weather, the loss of cross-power amplitude can 
only be due to external absorption, and antenna pointing.  The amplitude 
drops are quite amazing:  20% at Cband, by by a factor of up to *three* 
at K-band.  That means a 40% absorption at C-band, and almost an order 
of magnitude at K-band -- an opacity of about 2.3.   Such an opacity 
must be accompanied by a spectacular increase in system temperature -- 
unfortunately, we are still not able to monitor this, but the effect on 
the SNR is easily seen, with the scatter in amplitude and phases hugely 
(and I do mean hugely) increased at K-band, and significantly at C-band. 
    The absorption is time and space variant -- I can easily track the 
path of the storm over the array by plotting out the opacity.  The storm 
maximum was on the west arm, early in the run, and headed north and 
east.  The east arm was least affected.  Most of the storm was gone by 
the end of the run, where reasonable stability and sensitivity was 
observed. 

    The event was of course accompanied by enormous phase instability, 
at both bands.  The phase jumps noted on antenna 21 are easily seen in 
raw phase at C-band, but could only be discerned at K-band by 
differencing the two IFs. 

    Some other interesting things are noted:

    A) Integer Zeros:  FITLD reported 1.1 and 2.4 %, at C and K band 
respectively.  Most of these are very curiously distributed -- 13 
seconds *before* the end of each 3-minute block, many antennas were 
'blank', if all IFs, for a single record.  Thereafter, to the end of the 
block, the RCP correlations were all present, but many of the LCP were 
integer zeros. 

    B) Non-functioning antenna-IFs -- 'failure to tune' issues: 

    At C-band, antenna 5 gave only noise, on both IF pairs for one 
scan.  Antennas 16 and 19 were slow to begin fringing (a minute or 
less), again on both IFs, and on one scan each (not the same scan). 
    At K-band, things are a bit messier:  antenna 23 failed to fringe on 
all IFs for one scan.  antenna 8 was slow to fringe, again on both IFs, 
for one scan.  Antenna 12 and 20 failed to fringe for an entire scan (3 
minutes) on the BD IFs, for one scan each, while 4, 15, and 26 all 
failed for one scan on the AC IFs.  For antenna 4, the failure was 
interesting in that 4A was identically zero (blanked) for the entire 
scan, while 4C gave noise. 

    C) As noted separately, ea6D is very weak, with a large delay.  Ken 
has reported that this is being addressed today. 

    That's about all.