Links to the new CARSR radar.

Links to the old FAA RADAR page (pre carsr). (top)

Old Faa radar specs.
Modeling the FAA radar total power spectrum (for pulsar observations).

history:

• 28Aug13: FAA carsr starts blanking in ao direction,.
  
• 08Aug13: carsr radar starts broadcasting
  

FAA  carsr-3  radar specs: (top)

Notes on the radar (top)

•  The ipp 15 ipp sequence does not divide evenly into the 12 second rotation period. It looks like they let the sequence move beyond the 12 second period and then they jump by 5 ipps to get back to 12 seconds (more info).

12may20: faa voltage dataset

    Complex voltage data was taken on two of the FAA radar frequencies as well as a off radar freq on 12may20. The setup was:

• mock spectrometer was used to record the data
• center freq: 1349.6, 1344.4 and 1405 MHz
• 8 MHz,bandwidth (.125 Usec sampling)
• I unpacked the mock data into files containing complex floating point data (4bytes real , 4 bytes img)
• 2 files for each frequency: polA and polB (linear pols)
• for a total of 6 files
• there are 96000000 complex samples in each file (12 seconds of data .. 1 radar rotation).
• 768000000 bytes each.
  
• The filenames describe the freq and pol in their names.
• a 7th file is a text file with a description of the file format (fileinfo.txt).
  
• This link will take you to the directory with the files
• you can download them via http by clicking on each file.
• The specs of the faa carsr-3 radar are shown at the top of this page.
  
• A listing of the files is:
• -rw-rw-r-- 1 phil computer 768000000 May 28 17:34 cmplx_8mhzbw_cfr1344.4_polA.dat
  -rw-rw-r-- 1 phil computer 768000000 May 28 17:34 cmplx_8mhzbw_cfr1344.4_polB.dat
  -rw-rw-r-- 1 phil computer 768000000 May 28 17:34 cmplx_8mhzbw_cfr1349.6_polA.dat
  -rw-rw-r-- 1 phil computer 768000000 May 28 17:34 cmplx_8mhzbw_cfr1349.6_polB.dat
  -rw-rw-r-- 1 phil computer 768000000 May 28 17:34 cmplx_8mhzbw_cfr1405.0_polA.dat
  -rw-rw-r-- 1 phil computer 768000000 May 28 17:34 cmplx_8mhzbw_cfr1405.0_polB.dat
  -rw-rw-r-- 1 phil computer       840 May 28 17:34 fileinfo.txt
  

Plots were made showing the total power and spectra

total power vs time 12 seconds full scale (.png)

• The total power (over 8MHz) was computed for each .125Usec sample and then plotted vs time.
  
• The 1st two frames are the 1349.6 MHz band polA and polB (linear pols)
• The bottom two frames are the 1344.4 MHz band polA and polB
• The horizontal access  is time in seconds. 12 seconds is 1 rotation period of the radar
• The vertical access is total power. It has been normalized the system temperature of the lband wide receiver (about 30K)
• The radar is pointing at AO around 2.5 seconds.
• the dropout is when the faa radar blanked it's single in the AO direction.

• The is the same plot as above but the vertical scale has been blownup.

Total power vs time blowup in time around peak value (.png)

• this is a blowup in time around when the faa radar points at AO (for the 1349.6 MHz band)
  
• top:  2.7 seconds of data when the radar points at AO.
• The spikes are the individual radar ipps.
• The blank region is where the radar  blanks when pointing at AO.
• 2nd: 10 milliseconds after  faa radar blanking turns off
• the has  the peak value (the blanking was not perfectly centered on the ao direction.
• The dashed green lines are around the peak ipp pulse.
• 3rd,4th frames.: 1405 band during the peak pulse
• This shows the same time span as frame 2. 
• The dashed green lines show the location of the peak pulse
• The data has been smoothed by 51 samples (6.75 usecs) to beat down the noise
• If the  receiver chain had gone into compression then we should see a dip in the 1405 tp around the peak pulse/.
• there is no dip so the receiver chain did not go into compression.

The final plot shows the average  spectra for the 12 seconds of  data (.ps) (.pdf)

• Black is polA, red is polB
• the vertical scale is linear in the spectral density function.
• There 1024 channels over the 8MHz.
• the spectra averages over the entire 12 seconds  so the peak spectra when the pulse is on would be much higher.
• The bandwidth of the chirped is supposed to be about 1MHz.
• The total power plots were computed over the entire 8MHz of the band (so it included more than just the radar power).

processing: x101/220228/genfiles.pro, faa_volts.pro

09may14: FAA  blanking

    Test data was taken on 14may14 to check the FAA radar blanking. The data looks at the blanking region as well as the sidelobe height over the 12 second rotation period. The data taking 07may14 only has  two 12 second rotations, so it couldn't differentiate between sidelobes and echoes from planes. The 09may14 data took 3 sets of 25 second data (2 rotations per set). The data sets were spaced by many minutes so plane echoes would not be stationary.

    Note: the engineers had been working on the antenna focus for the previous week or two. During this work, they shortened the blanking period for some tests. They promised that before they left, they would put the blanking back to 4.25 degrees.

The setup:

• lbw, linear pol
• az=80, za=18 (this was a peak for FAA compression for alfa data taken a few years back).
• mock spectrometer 16 bit timedomain data.
• 3.2 MHz bandwidth
• center frequencies of each 3.2 MHz:  1257.5,1349.5,1274.5, 1232.5 MHz
• the first two are punta del este (FAA) radar frequencies
• the last two are punta borinquen frequencies.
  
• 3 scans were taken each with 25 seconds of data.
• Start time for each 25 second scan
• 15:36:17 ast
• 15:38:17 ast
• 15:41:17
• So each scan was separated by an integral number of 12 second rotations.

Processing:

•  Each 25 sec scan was input and processed
• Power was computed for each 3.2 MHz sample
• power was averaged to 100 usecs. this is close to the 117 usec pulse length.
• polA and polB was averaged.
• The total power time series was normalized to when the radar was blanked (pointing at ao).
  
• This make the vertical scale Tsys units for AO.
• The normalized time series was then converted to db. 0db is not Tsys at AO.
  

•  Page 1: over plot 6 12 second rotations (with offsets for display)
• black red were the first two rotations starting at 15:36:17
• green,blue are the two 12 sec rotations starting at 15:38:17
• purple,dark red are the two 12 sec rotations starting at 15:41:17
• The gap at 6.5 seconds is when the radar blanks in the AO  direction.
• the topmost plot has a signal that pops up at 7.6 seconds. This is probably a plane
• Page 2: average the first 5 12 second rotations.
• the 6th rotation was excluded from the average because of the plane.
• Top: average 12 second period
• red dashed line shows the peak value (with ao blanking enabled)
• green dashed line shows the far out sidelobe peak value
• The far out sidelobe is about 13 db below the peak.
  
• bottom: blowup around AO blanking.
• blue vertical  line: current blanking
• the current blanking is set to 3.27 degrees (.11 seconds)
• purple vertical line: blanking specified in dod/dhs memo: 4.25 deg (.142 seconds)
  
• the radar power versus rotation phase for 1257.6 MHz radar (.ps) (.pdf):.
• The far out sidelobe is 13 db below the blanked peak.
• To bring the near in sidelobe down to the far out sidelobe amplitude:
• blank by .44 seconds (13.2 degrees)
  
• The radar power vs rotation phase for the 1349.6 MHz radar (.pdf) (.pdf)
• the far out sidelobe is 12db below the blanked main beam.
• To bring the near in sidelobe down to the far out sidelobe amplitude:
• blank by .42 secs (12.6 degrees of blanking).

Summary:

• the FAA blanked radar is 25 to 30 db above the AO system temperature
• in a 3.2MHz total power bandwidth.
• the current blanking is  .11 seconds (3.27 degrees)
• the blanking specified in the dod/dhs memo was .142 seconds (4.25 degrees)
• The engineer narrowed the blanking during his testing.
• To bring the close in side lobes down to the far out sidelobe level:
• the blanking would have to be increased to about 13 degrees (.433 seconds)
  

07may14: FAA,Punta borinquen blanking

    The FAA engineers were working on the punta del este radar (FAA radar) in early may14. They asked for us to check what the radar looked like now at AO.

The setup was:

• The setup was similar to the 14jan14 setup:
• lbw in circular polarization mode.
• The 10 ghz upconverter was used.
• The mock spectrometer took time domain data centered at :
• 1257.59,1349.59 (FAA long pulse duration frequencies)
• 1274.59,1332.59 (the punta salinas radar)
• i used 6 mhz offsets  to get rid of the dc spikes.
  
• 16 bit time domain sampling with 3.2 MHz bandwidth was taken for 25 seconds.

   Processing:

• The total power was computed for each sample and then averaged by 350 samples.
•  This  gave a total power integral of 114 Usecs which matched the 117 usec pulse durations.

• I normalized each dataset to the median value. This should be close to Tsys for the 3.2 MHz band (since the radar duty cycle is only 9%).
  
• FAA radar frequencies 1257.5 (black) and 1349.8 (red) MHz.
  
• Top 12 seconds of  data showing 1 FAA and 1 punta borinquen period
• FAA. Black in 1257.6 MHz,  rd is 1349.6 MHz.
• It points toward AO at 7.97 secs on the plot
  
• The FAA is not showing a strong sidelobe about 1.4 secs after the main pulse (unless this is an airplane echo).
• Punta borinquen radar: green=1274.6, blue=1332.6 MHz
• The radar points at Ao at 2.5 secs in the plot
  
• the punta borinquen radar has a sidelobe at .449 secs (13.4 deg) and 2*.449 secs from the radar pointing at AO.
  
• middle: blowup of FAA radar when in points at AO
• the blanking is .126 secs or 3.8 deg.
• This has decreased from the 4.6 degrees of blanking back on 14jan14
  
• bottom: blowup of punta borinquen radar when it points at ao
• The radar is blanking for 1.42 secs or 4.3 degrees.

The 2nd set of plots shows the spectrum of the radar pulse (.ps) (.pdf)

• Page 1: 12 1 second spectral averages
• 4096 length fft's were done on the 3.2 MHz bw data and then averaged to 1 second.,
• Each color is a different 1 second avg spectra
• The large values are when the radar points at AO.
  
• Top: 1257.5 MHz radar  spectrum
• 2nd: 1349.5 MHz FAA radar spectrum
• 3rd: 1274.5 MHz punta borinquen spectrum
• bottom: 1332.5 MHz punta borinquen spectrum
• page 2: .137 second average spectra right after blanking ends
• These average spectra are when the radar is strongest at AO (right after the blanking)

Summary:

• The FAA radar has decreased from 4.6 to 3.8 degrees
• the FAA radar looks to have a strong sidelobe 1.4 secs (43 deg) away from main beam.
  
• I need to verify that this is not an airplane by looking on the scope at a later time.
• The punta borinquen radar is similar to what it was on 14jan14.
• Notes on the strength.
• The strength of the radar in our receiver is a strong function of the azimuth and zenith angle when the data is taken,.

14jan14: FAA moves blanking sector by 1 degree

The setup was:

• lbw in circular polarization mode.
• The 10 ghz upconverter was used.
• The mock spectrometer took time domain data centered at :
• 1257.59,1349.59 (FAA long pulse duration frequencies)
• 1274.59,1332.59 (the punta salinas radar)
• 16 bit time domain sampling with 3 MHz bandwidth was taken for 25 seconds.

   Processing:

• The total power was computed for each sample and then averaged by 350 samples.
•  This  gave a total power integral of 114 Usecs which matched the 117 usec pulse durations.

• I normalized each dataset to the median value. This should be close to Tsys for the 3 MHz band (since the radar duty cycle is only 9%).
  
• FAA radar frequencies 1257.5 (black) and 1349.8 (red) MHz.
  
• The punta salinas radar also appears in the FAA 1257.5 MHz band.
• Page 1:
  
• Top 12 seconds of 1 FAA period.
• The ao blanking occurs from 10.180 to 10.3207 seconds (4.22 degrees for a 12 second rotator).
• 
  
• 2nd: blowup of FAA radar pulse during ao blanking
•  all 4 frequency bands are over plotted. each color is a different radar frequency.
• 3rd: 12 second rotation period of the punta borinquen radar,
• bottom: blowup of punta borinquen blanking in the ao direction.
• You can see the large sidelobe at 5.2 seconds
• Page 2: Plot blowup of blanking in ao direction using radar azimuth.
• Top: nov13 when blanking covered 272 to 276.4
• bottom: jan14 after blanking region moved to 273 to 277.4
• I took the center of the blanked region and mapped that to the center of the blanked region (as reported to me by the FAA).
• The 1 degree move was supposed to center the blanking region. It looks like it went too far.
• The leading pulse is now about 4 times higher than the peak of the trailing side lobes.
• The two data sets were taken at different telescope azimuths. This may explain the different signal strengths.
• The blanking region is currently:
  
• 273 - 277.6   : 4.6 degrees of blanking
  
• To really center the blanking it would be beneficial to advance the start of the blanking sector to give:
• 271 to 277.6 : 6.6 degrees of blanking.

01nov13: FAA blanks in ao direction.    (top)

Plots:

the total power vs time for a 12 second rotation (.ps) (.pdf)
show the spectra over the 3 MHz (.ps) (.pdf)
the average daily spectrum for 25oct13 through 30oct13. (.ps) (.pdf)

• lbw in circular polarization mode.
• The 10 ghz upconverter was used.
• The mock spectrometer took time domain data centered at :
• 1257.59,1349.59 (FAA long pulse duration frequencies)
• 1274.59,1332.59 (the punta salinas radar)
• 16 bit time domain sampling with 3 MHz bandwidth was taken for 25 seconds.

• The total power was computed for each sample and then averaged by 350 samples.
•  This  gave a total power integral of 116 Usecs which matched the 117 usec pulse durations.
• 4096 channel spectra were computed and then averaged to 1 second. This was repeated 12 times to cover an entire 12second cycle.
• spectra were also averaged for .137 seconds after the blanking sector.
  

• I normalized each dataset to the median value. This should be close to Tsys for the 3 MHz band (since the radar duty cycle is only 9%).
  
• Top: all 4 frequency bands are over plotted. each color is a different radar frequency.
• The FAA radar blanks at 1.5 seconds.
• The punta borinquen radar blanks at 8.1 seconds
  
• The dashed purple line is where the punta salinas radar blanks in our direction
• The 1257.6 FAA radar band is noisier because punta salinas also sits in this 3 MHz.
• middle: FAA radar blanking
• black in 1257.6, blue is 1349.6
• the blanking sector used was 272-276.4 degs az.
• The blanking sector is not quite centered on AO.
• They should probably blank 273-277.4
• bottom: punta borinquen radar blanking
• red is 1274.6, green is 1332.6
• The blanking sector is 108.4 to 112.4
• The blanking is centered on ao. Back in 03oct13 i  had them move the blanking sector up by 1 deg (more info)

• Page 1. compute 12  spectra each averaged to 1 sec. This covers and entire radar rotation (12 seconds).
• Each color is a different 1 second average. As the radar points at AO, that 1 second average spectra will get stronger.
  
• I normalized each 12 second dataset to the median spectral value. This should make Tsys =1 on the plots.
  
• Top: FAA radar at 1257.59 MHz.
• the radar at 1256.l8 is the punta salinas radar. This is why the 1257 total power was so noisy.
• It looks like punta salinas is blanking for about 1 second (30 degs). I thought they used to blank for +/- 45 deg (more info)
• 2nd: FAA 1349.6 band. The blue trace is the 1 second average spectra when the radar pointed at ao.
• 3rd: punta borinquen 1274.6 band
  
• 4th: punta borinquen 1332.6 band
• The fwhm is about .8 to 1 MHz.
• Page 2: .137 second average spectra taken after the blanking sector.
• This should be where the signal is still strong (except that punta borinquen has a large sidelobe 15 deg from boresight.
• frames1-4 show the 4 bands.

• The spectrum analyzer does a 1 minute peak hold covering dc to 10ghz in 19 separate bands.
• The FAA band is measured every 19 minutes giving 108 measurements per day.
  
• these plots are the daily average.
• The 4 days are color coded.
• Top: 1349.6 MHz FAA band
• bottom: 1357.6 MHz FAA band
• I t looks like the blanking was enabled on 28oct13.
• The maximum difference before, after is about 10 db
• This data comes from a swept lo spectrum analyzer.
  
• Since the radar is pulsed, the data  probably does not contain the peak value of the radar

Summary:

• The FAA radar is now blanking in the ao direction.
• The blanking sector is currently 272. - 276.4 . It should probably be increased by 1 deg.
• The bandwidth  of the pulses are about 1 MHz.
• It looks like punta salinas is blanking 30 degs. it used to be +/- 45 deg.
• The hilltop monitor showed a change of 5-10db before, after the blanking (but the measurement did not measure the true peak of the radar pulse).
  

processing:x101/131101/faardr.pro

page_up
home_~phil