[evlatests] S-band and those nasty satellites

[Rick Perley]

    The 25 S-band databases that I've reduced recently have also allowed 
a more comprehensive look into the effects of those satellite downlinks 
on our S-band receivers.  Mostly, the news is pretty good.   The objects 
observed were between declinations -22 and +24 -- including one at dec = 
-4.5 (the significance of which will become clear, below). 

    The effect satellite RFI is most conveniently seen in the 'PSum' 
provided with each subband -- notably, those subbands which do not 
contain RFI.   Five subbands are generally directly affected -- numbers 
2, 3, and 14 through 16 (counting up from one).  The others are mostly 
free. 

    For the 25 'sample' databases taken, I looked at the fractional 
variation in PSum.  It is immediately evident that there is a strong 
correlation with declination:  Objects  north of +20 degrees declination 
have essentially no change in power.  The closer one gets to -5 
declination, the more disruption there is.  All sources with typical 
variations greater than 15% were within 5 degrees of -5 -- that is, 
between 0 and -10 declination.  There was no clear dependence on hour 
angle, nor is there any particular hour angle at which the induced gain 
variations are particularly bad.  Objects source of -15 declination , or 
north of +10, generally have variations less than 5%. 

    From this, it seems clear that the two (geostationary) XM satellites 
are not the prime cause of the gain variations.  The more likely origin 
is the large number of satellites broadcasting between 3.6 and 4.2 GHz 
that lie along the 'Clark' belt -- corresponding to a declination of 
about -4.5 degrees. 
    Attached are two useful plots, showing the motion of the three 
Sirius satellites in both HA-Dec, and Az-El coordinates.  I thank Todd 
Hunter for generating these.  To test the idea that the Sirius 
satellites are not the major issue, I reviewed the switched power data 
from one of the 'ARCADE' observations -- this of a source near +57 
degrees.  Referring to the 'HA-Dec' plot,  you'll note that the motion 
of any source of declination between 50 and 60 degrees must take it 
within a few degrees of one of more Sirius satellites -- which have 
orbits set up to make the Satellites 'lurk' in the vicinity for about 6 
hours at a time.  No effect from the Sirius-radiated RFI was seen in 
that particular observation. 
    Also on the plots is the 'Clark belt'.  The danger zone is in a band 
about 5 to 10 degrees above/below the green lines. 

    For the observer, the remarkably good news is that application of 
the switched power very nicely removes the gain variations -- even those 
as large as ~50%.  The additional power from the satellites is evidently 
causing gain compression in the (probably analog) electronics, but 
despite this, the performance of the array remains very good.  Even the 
target source at dec = -4.5 (which went essentially right through each 
satellite in the belt) provided a decent (but not great) image, 
following clipping of the obviously affected data (whole scans, and all 
subbands).  The rms level was about 3 times higher than the others, but 
the map still useable. 

    For reference, I note the bands over which the strongest RFI is present:

    1) 2180 to 2200 MHz.  This band is quite strong, and nearly always 
visible.  I've forgotten the origin, but it looks like geosynchronous 
satellites. 

    2) 2310 -- 2350.  These are generally the strongest, and are from 
the DSR:   Sirius and XM radio.   Nearly always, the subband containing 
these (#3in the standard setup) is rendered quite useless. 

    3)  3600 -- 4200 MHz.  A solid mass of satellite RFI (which extends 
into C-band). 

    Note that for A, and the longer B-configuration spacings, fringe 
winding  with a 4 second integration time is sufficient to essentially 
eliminate the RFI above 3600 MHz, providing the observations are more 
than 15 degrees away from the geosynchronous belt. 


-------------- next part --------------
A non-text attachment was scrubbed...
Name: sirius-xm_vla.hadec.png
Type: image/png
Size: 95439 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140203/c6673322/attachment.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: sirius-xm_vla.azelpolar.png
Type: image/png
Size: 110300 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20140203/c6673322/attachment-0001.png>