[evlatests] Antenna 14 Measurements from 20 Sept 2005

[Rick Perley]

    Bob Hayward and I spent most of the day making fundamental sensitivity
and stability measurements in Antenna 14.  The original goal was to track
the sensitivity (via the 'Y-factor' between hot load and cold sky) 
through the
signal chain.   As always, the tests uncovered unanticipated problems that
may, or may not, be related to our original goal.

    1) Up-Front Tests.  We did these in LCP (as the power meter is
single-channel). 

    We did a hot/cold power ratio at the RF, using a 4750/600 MHz
filter.  The Y-factor was found to be 9.3 dB -- a value 1.7 dB below
what we measured in February of this year (and the discrepancy will
increase when we account for the differing temperature of the hot
absorber between February and September).   This represents a
significant apparent loss in sensitivity!

    We next did an efficiency test, observing Cygnus A and Virgo A
in an on-off mode.  These data have not been reduced -- I'll get this
done today.  Visual inspection of the traces however revealed a
curious power instability while on Cygnus A -- but not seen in the adjacent
sky observation, nor on Virgo A, nor on the hot load calibration. 

    As a side issue, we note that we had considerable difficult in aborting
the scripts to end these tests.  Neither we, nor the operator, could get
the abort command to be accepted.  Approximately 30 attempts were
made for both tests before the tests were in fact aborted.   

    2) Test before and after the T304.   We elected to use IF `D', as 
this was
the best behaved channel in the astronomical data I took last week. 

    We elected to check the Y-factor at the output of the T304 first, rather
than at the input.  LOs were set to put the desired 4722 MHz center 
frequency
in the 1750/150 MHz filter we attached to the output of the T304. 
    We immediately noted the output power was very unsteady -- sporadic,
and fast (~100 msec) power fluctuations of ~0.3 dB were noted, as well 
as erratic drifts
of similar size over time scales of tens of seconds.  Yet at other 
times, the power
was semi-stable.  Weird behavior. 

    We measured the Y-factor as 9.4 dB -- the same as at the RF. 

    We moved the power meter to the output of the LSC converter, using a 
11000/100
MHz filter.  These data are perfectly stable (as they were in February 
when we did
out earlier tests).  Crisp and clean. 

    We then spent considerable time trying to track down the origin of the
instabilities which appear to arise within the T304. 

    - We noted that the 8 -- 12 GHz filter which the block diagram 
located on the
wall of the vertex room shows should precede the T304 was not present, 
so we
added one to see if any different in behavior could be found.  No 
changes were
seen.  *** I now can see why -- the updated block diagram on the wall 
opposite
my office shows this filter to be within the T304 module -- We need updated
block diagrams in the antennas!!!

    We noted that even with the power input to the T304 limited to 100 
MHz BW,
the output power instability remained. 

    The instabilities go in both directions (i.e., the `pops' can be 
down as well as
up).   Changes in power level as much as 1.2 dB were seen -- with the power
fairly stable both before and after the jump.  The jumps can be 
extremely fast --
faster than our time resolution of 100 msec. 

    All these data were recorded, and I can show them to anybody 
interested. 

    At 6:10 PM we quit for the day.  We were unable to get to the backend
measurements, so these must wait until Thursday. 

    Rick