[evlatests] Results from Solar Tests of 11 Oct.

[Rick Perley]

    Two tests for 'solar mode' observing were made yesterday.  These were:

    1) A test to determine if the power levels are being set correctly:  
Observations were taken of a calibrator with the T302's Switchable Solar 
Attenuator (SSA) off, then of the sun, with the SSA turned on.  For the 
first two observations, Ken arranged the executor to adjust the T304's 
input and output attenuators to attain the desired power levels, to 
remember these settings, and apply them to subsequent observations of 
calibrator and sun. 

    2) A test to measure the attenuation, phase shift, and delay of the 
T302's SSA:  Observations of a calibrator were made at L, S, and C 
bands, alternating between having the SSA on and off.   The attenuator 
levels in the T304s were held fixed in order to ensure any differences 
are due to the SSAs.  A total of four short scans with the attenuators 
off, and three scans with the attenuator on were made. 

    Basic Results:

    Test 1: 
    In general, the algorithm to set the power levels works well -- 
although as before, the route taken to obtain the correct attenuator 
settings for the on-sun observations is difficult to understand. 
    Ideally, the T304 output power (to the samplers) should be the same 
when on the calibrator and on the sun.  A difference of a couple of dB 
would not be a concern.  For most antenna-IFs, this expectation is met.  
There are some notable discrepancies however:  ea02, IF 'B', for 
example, has a difference of nearly 6 dB -- far more than it should be.  
About six antenna-IFs have power differences more than 3dB between 
on-cal and on-sun observations.   Presumably, this difference is related 
to the algorithm which tries to find the appropriate settings. 

    Test 2: 
    Most antennas switched successfully at all three bands.   We looked 
at three characteristics:

    a)  Amplitude
    The attenuation was measured to be 15 dB (not 20 dB, as I had 
somehow expected), and is the same (to within a dB or so) for all three 
bands for each antenna. 

    b) Phase
    For most antennas, the phase change between 'in' and 'out' states is 
small (10 or 20 degrees or less) -- but for some, the phase change is 
large.  The phase change is proportional to frequency -- as expected as 
these attenuators are in the RF path. 

    c) Delay
    For all antennas and at all bands, the delay difference is very 
small -- less than 0.4 nsec. 

    Perhaps more interesting are some notable exceptions to these results:
    a) Antenna 10 showed *no change* in amplitude between the 'in' and 
'out' states, at all three bands.  However, the situation is not one of 
failed commands, or a failed attenuator, since the *phase changed* for 
every change of state.  So ... something is changing, but that change 
doesn't affect the signal strength.   This behavior was the same for 
both polarizations. 
    b) There were four antennas which shared the behavior of antenna 10, 
but only for one of the transitions:  At L-band, antennas 14 and 17 each 
failed to change from the 'in' to the 'out' state once (at a different 
time for each).   At C-band, antenna 12 failed (once) to change from 
'out' to 'in' state, and antenna 7 failed (once) to change from 'in' to 
'out'.  For all of these, the phases did change, and the behavior is the 
same for both polarizations.  In other words, this failure mode is the 
same as for antenna 10, but occurred much less frequently.