[evlatests] PDif compression -- more tests

[Rick Perley]

    A pair of tests were done this week to see if we can better 
understand the 'PDif compression' phenomenon. 

    1) The on/off moon test was done at X-band with two different power 
levels set by the input attenuator in the T304.  The first on/off was 
with the powers set to their normal level (-20 dBm).  This was followed 
with another on/off with the set points reduced by 6 dB -- to -26 dBm.  
The purpose was to see if this reduced power changed the apparent 
compression.  If so, the internal working of the T304 would be 
implicated.  (The output attenuator was allowed to restore the 
appropriate power to the samplers, so only the internal bits of the T304 
could see any change).

    Result:  No change in PDif compression. 

    2) Keith padded down the output power from the X-band receiver in 
antennas 7, 16, and 17 by typically 6 dB.  These three have the largest 
'PDif compression' ratios of any.   This was prior to the tests noted 
above.  If the compression was due to the input amplifier in the T304, 
we would see a difference. 

    Result:  No change in PDif compression.  These three antennas are 
still as bad as they were before. 

    Comment 1:  We must be barking up the wrong tree.  All the tests I 
can recall doing in the past couple years have exonerated the T304.  
Since the problem is manifest in the X-band system, and since the old 
style X-band receivers did not show the compression, while the new style 
ones do, is the problem somehow connected to the receivers, or the 
switched power injection mechanism? 

    Comment 2:  A few of these new X-band receivers have minimal, or no, 
compression.  Somehow, a few of them work as desired.  In particularly, 
ea25 is, and always has been, perfectly linear. 

    Comment 3:  Four antennas at X-band have dysfunctional switched 
power in one IF or more:

    a) ea02:  zero mean PDif.  (only noise).  Has been this way for some 
time. 
    b) ea20:  zero mean PDif, and bizarre PSum (three different 
levels).  (this is a new receiver). 
    c) ea16A:  PDif is greatly negative (!), and has been for many 
months.  Only this IF shows this. 
    d) ea28, LCP:  PDif is is very small (IF C) or negative (IF D).