734-740 rfi in 800 Mhz rcvr

Intro:

• during their 20,21,22 nov10 800 Mhz runs the rfi was seen
  
• data taken on 17,18,19oct10 did not show the rfi.
• The rfi appeared on 20oct10 and remained present for the following days  (except when the 800 Mhz rcvr had trouble and was turned off).

• I looked at the a2524 data and then the  x111 data taken during oct,nov10
• I also searched the hilltop monitor rfi system where it was also seen.
• An azimuth spin was then done to see if the rfi was time variable and which azimuth direction was the strongest
  

Conclusion:

24nov10: 734.7-739.3 Mhz rfi in band. First report (top)

• Top: average  spectra (300 seconds off source, file .00100) from 20nov10.
  
• Black in polA, red is polB.
• The interference is much stronger in polB.
• 2nd: blowup around the rfi.
• 3rd: rms/mean by channel for the 300  1 second records.
• the expected rms was .0069 . The measured values (away from the rfi was .007) so the noise statistics are probably ok.
• 4th: blowup of the rms/mean by channel around the rfi.
• polB is much stronger, but polA is also affected.

• data taken on 17,18,19oct10 did not show the rfi.
• The rfi appeared on 20oct10 and remained present for the following days  (except when the 800 Mhz rcvr had trouble and was turned off).

• The data has 1 minute peak holds taken every 19 minutes.
• The Tsys is about 800 K.
• Dynamic spectra were made with 8 days per page from 01oct10 thru 24nov10. Darker in stronger intensity.
• not seen
  
• rfi_737_101009.gif
• not seen
  
• rfi_737_101017.gif
• appears weakly 23:00 hours 21,22oct10. maybe 24oct10.
  
• rfi_737_101025.gif
• appears on 28oct10 (stronger). continues thru 01nov10.
  
• rfi_737_101102.gif
• Intermittant. strong on 08/09nov10
  
• rfi_737_101110.gif
• visible most days here.
  
• rfi_737_101118.gif
• weak barely visible
  

• Nothing was seen (around 10am).

SUMMARY:

• 734.7->739.3 rfi first appeared 20oct10.
• it is seen in the x111 telescope data, a2524 dps scans, and from the hilltop monitor.
• It was not seen from the platform using the Tektronix spectrum analyzer and a log periodic antenna.
• The band is close to channel 55  tv frequency but the spectrum does not look like an analog tv channel.

29nov10: az swing looking at 734.7-739.3 Mhz rfi  (top)

• az -100 to + 260 deg. za=12.
  
• Clockwise then counter clockwise spin at .35 deg/sec
  
• mock spectrometer.
• 11.469 Mhz bandwidth, 8192 channels
• .1 second spectral averages.

• processing for each .1 second spectra:
• compute total power in 734.7-739.3 rfi band for each .1 sec spectra.
  
• compute total power  in two bands adjacent to rfi for bpc. do this for each .1 sec spectra.
  
• subtract total power rfi band - total power bpc bands  (to remove sources)
• normalize total power to median total power value (to put in units of Tsys).
• Page1: total power rfi band clockwise spin.
• Top: rfi band total power  (black polA, red polB) plotted vs azimuth
• polB is much stronger than polA.
• the polB rfi peaks at az=91.6 degrees az.
• bottom: blowup around polB peak.
• The rfi is about 5.5 * Tsys.
  
• The fwhm is about 2 degs za.
  
• Page2: overplot rfiband total power vs az for clockwise and ccw az swings.
• top: PolA rfiband total power vs az.
  
• black is CW az swing, red is CCW az swing.
• Each swings take about 20 minutes.
• both swings overlay each other exactly. This implies there is no longer time variability in the signal
• bottom:PolB rfiband total power vs az for CW and CCW az swings.
• Both az swings overlay each other exactly.

• Page 1: Average spectra (19.6 minutes) CW az swing:
• top: full 11.5 Mhz average spectra:
• black is polA, red is polB.
• the left side looks more dense than the right. This is just a plotting alias. Blowing up the picture, the comb spacing looks more equal.
• Bottom: blowup 734.7 to 739.3
• Page 2: .1 second spectra at az=91.5 (polB peak CW az swing)
• top: 3 .1 second spectra with offset for display.
• Bottom: 1 Mhz blowup around center of rfi : 737.
• the blue dashed lines are spaced every 45 Khz starting at the leftmost peak of the picture (736.53698 Mhz).
• the red dashed lines are spaced every 45 Khz starting at the 1st peak after 737  (737.04336 Mhz)
• There is a discontinuity in the comb at the band center of 737 Mhz.
• The center shows a dip
• The comb phase jumps by about 12.5 Khz when crossing the center.
  

Summary:

• polB is stronger than polA
• the rfi repeats in azimuth for two azimuth swings spaced by about 20 minutes.
  
• Implies the rfi is stable in time.
  
• The spectra shows a comb with 45 Khz spacing 734.7 to 739.3 with very little leakage outside this band.
  
• Peak rfi polB is near 91.5 azimuth (with dome at za=12).
  
• .1 second spectral average near peak value in 1.3 khz channel is  about 20*Tsys.
• The spectra has a discontinuity at the center (737 Mhz). The comb phase phase jumps by about 12.5 Khz