[evlatests] The X-band birdie

[Walter Brisken]

David Sevilla has been pointing his software at the X-band birdie that 
Rick reported in the ~8800 MHz region.  David's program captures samples 
from a deformatter and makes power spectra.  We listened at X-band on 
antenna 16 while changing the X-band tuning.  This is what we (well, 
mostly David) found:

1.	Rick's ~8800 MHz birdie.  The combinations of L302 settings and 
birdie frequencies (with +-0.1 MHz uncertainties) are:

L302		Birdie		Sum
11619.9		8860.0		20479.9
11640.1		8839.9		20480.0
11669.9		8810.0		20479.9
11690.1		8789.9		20480.0
11719.9		8760.1		20480.0
11740.1		8739.9		20480.0

Barry mention something about a 5th harmonic at the EVLA meeting Monday. 
20480 is indeed the 5th harmonic of the 4096.  So I think that the signal 
that is getting through the first stage mixing in the T304 must be
2*L302 - 5*4096, which for the frequencies investigated above lies within 
the 2048 and 4096 MHz bandpass filters.  In a 62.5 kHz channel, this 
birdie doubles the system temperature.  We should check the 
better-shielded T304 that has been deployed to test the 8192 MHz birdie 
supression.

2.	David noticed something quite peculiar.  In every spectrum he made 
at X-band (no matter what the tuning) there is a faint birdie in the exact 
center of the 1024 MHz digitized bandwidth.  This is present at C-band as 
well.  This signal is a bit weaker than the birdie described above.  This 
won't be visible in the VLA correlator as the portion of the band the VLA 
correlator can see does not include this.

3.	RFI -- various over the band.  David or I will report on this more 
later.


-Walter