[evlatests] Latest X-band RFI tests

[Rick Perley]

    I repeated Friday's RFI sniffing tests, this time with antenna 18 
included, and in both RCP (which has the absorber/shielding in antennas 
13 and 18), and in LCP (which has no shielding or absorption). 

    The channel resolution is about 3 kHz.  The two frequencies observed 
were 8192 MHz, to see the 2nd harmonic of 4096 MHz, and 8799.8 MHz, to 
see the beat between the L302 frequency and the 5th harmonic of 4096 MHz. 

    Results:

    Channel A, 8192 MHz. 

    Total power plots show that both 13A and 18A are visible at low 
level -- perhaps 100 Jy.  Antenna 14A is at a few thousand Jy, antenna 
16 A has grossly saturated the correlator (uncounted thousands of Jy).  
The cross-spectrum between 13 and 18 shows a response of about 10 Jy -- 
this after considerable fringe winding in the 3.3 second averaging that 
was used. 

    Channel A, 8799 MHz. 

    Antennas 13 and 18 have no total power detection of this harmonic at 
all.  For antennas 14 and 16, the response at this frequency has 
saturated the autocorrelation spectrum -- especially grossly with 
antenna 16.  In the cross-spectrum, there is no detection of the 
harmonic at all on baseline 13-18.  All the other EVLA-EVLA baselines 
show the line, but with moderate intensity.  Again, considerable fringe 
winding has helped bring this amplitude down. 

    Channel C, 8192 MHz.

    The line is extremely ('saturated') strong on antenna 14 and 16, and 
at the level of a few thousand Jy on 13 and 18 (which are unabsorbed on 
this polarization), as seen in total power (autocorrelation).  In 
cross-power, the line is about equally strong on all baselines, at a 
level of a few thousand Jy (after fringe winding). 

    Channel C, 8799 MHz. 

    The harmonic line here is greatly stronger than any other seen in 
this investigation, especially on antennas 13 and 18.  In addition, 
antenna 18C shows what appear to be sidebands on each side of the 
central line -- the spacing is approximately 40 kHz.  At least 9 
satellite lines can be seen.  The cross-power spectra show the line as 
expected (at thousands of Jy).  I can also see the 'satellite' lines on 
the baseline 13-18. 

    In summary:

    The absorption/shielding appears to be effective for the 8799 MHz 
line -- presumably because the origin involves radiative coupling 
between the two mixers/LOs.  However, at 8192 MHz, there is still quite 
a strong response, despite the shielding and absorption.  Another order 
of magnitude (10 dB) is likely needed, if we are to reduce this to 
negligible values.