Links:

Dynamic spectra (172MHz about 1310) for 85 seconds and .1 second resolution (.gif):
Dynamic spectra blowup about (1385-1396)  (.gif):
Plotting 1390 and aerostat birdies with .1 second resolution (all 85 seconds)(.ps) (.pdf):
Plotting 15 seconds of 1390/aerostat birdies at 330 usecond resolution (.ps)  (.pdf):

Intro:

     Project a2010 (alfalfa) reported a radar like rfi near 1390 MHz on 11sep09 in the early morning. The a2010 setup was:

• alfa centered at 1385 MHz taking 4096 channels across 100 MHz using the wapps and 3 level sampling.
• 300 MHz about 1385 is converted to 250 MHz.
• 100 MHz about IF=250 is then digitally down converted to 0-100 MHz.
  

• The mock spectrometer with two 172 MHz bands centered at rf frequencies of 1310 and 1460 MHz.
• The 300 MHz rf band is 1st mixed to 250 MHz.
• A complex mixing brings IF=175 (1460) -> baseband, and IF=325 (1310) ->baseband.
• Each band had 512 channels.
• 128 Usecond sampling is used.
  
• polA and polB combined.

The 1390 rfi.

• p2473 File 00166 was used. This started at 3:28 AST 11sep09 and lasted for about 4 minutes.
• 85 seconds of the file was first averaged to .1 second spectra
• Dynamic spectra were made of the 1310 MHz band and a blowup around 1390 MHz.
• Power vs time was plotted for the 1390 rfi (at .1 second resolution) as well as the aerostat channels.
• 15 seconds of the file were input and averaged to 330 useconds (the pulse length of the aerostat radar).
  
• The 1390 and aerostat freq were over plotted showing that the 1390 rfi is associated with the rfi.

• Note the frequency increases right to left.
  
• colored vertical lines select various radars:
• red: 
• 1246 and 1261 are the aerostat.
  
• 1359 and 1373 are the bbmixer images (about 1310) of 1246 and 1261
• The dark blank stretches (about 1.4 seconds and .8 seconds) are when the aerostat blanks its transmitter toward AO and the san juan airport.
  
• purple:
• 1330, 1350: FAA radar
• 1270,1290 : bbmixer images (about 1310) of 1330,1350. 1290 also includes the remy radar.
• green:
  
• 1257,1248,1242,1233 : punta salinas mode A. some of these fall on top of the aerostat frequencies.
• 1390 rfi:
• this looks like it is jumping between 2 frequencies separated by about 340 KHz.
  
• This could also be a drift and then a jump back every 12 seconds.
  
• It as a period of 12 seconds (same as radar rotation).
• The peaks line up with the aerostat as well as the blanking region.

• This is a blowup of the previous image about the 1390 rfi.
• You can see the moving back and forth of the power in frequency.
• The jump from left to right (high to low freq) occurs during the aerostat blanking toward AO.

• Black is the aerostat aliases at 1358 and 1373 with a .1  second resolution
• Red is the    1390 rfi over plotted (and scaled up by a factor of 2).
• You can see that the 1390 and aerostat peaks line up.

• The aerostat has 2*150 usec pulses so the 330 usecond resolution matches the pulse length.
• Top: 1390 rfi vs time (1389.89 to 1390.99). The green lines are when the aerostat blanks.
• Bottom: Black is the aerostat aliases at 1358,1373.
• Red over plots the 1390 rfi.
• The main peaks line up as well as the onset of the blanking.
  
• The 1390 does not come on as quickly after blanking as the aerostat. This could happen if the signal after blanking is not strong enough to saturate the system.
  

Conclusions:

• The reported 1390 rfi is associated with the aerostat radar.
• The peaks and onset of blank line up.
• The 1390 frequency moves by about 340 kHz over the 12 second rotation period. This is not understood.
• It is not obvious the path for the 1390 interference to be created in the system.
  
• Since it showed up at the same frequency in the wapps and the mock spectrometer the problem is occurring before the 2nd IF (in the first mixer or transmitted by the aerostat itself).
• This rfi did not show up on the a2010 run on 12sep09.