The FAA airport radar (pre aug13)

The FAA airport radar transmits at 1330 and 1350 MHz. The transmitter is located east of san juan. The radar is used for traffic control around puerto rico (it is not the radar used for landing the planes), The radar on this page was replaced by a carsr radar on 08aug13.

Links:
Main FAA radar page

Pre aug13 FAA RADAR

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

08aug13: faa carsr radar starts broadcasting
may12: radar ipps jump by .3 %. Dip switches for timing probably set incorrectly.
may08: When the radar transmits at 1330 and 1350 MHz.
10apr07: faa radar ipps jump by .3%.
30nov05: compression in the alfa receiver with the 100 MHz filter installed.
17oct05: location of 1350 faa radar aliases in alfa.
01aug05: compression, baseline variation with long recovery in alfa (from faa radar).
27mar05: faa 1350 MHz radar appears at 1380 MHz in alfa survey data.
08may05: aerostat compresses alfa when oh filter bank amps installed.
06oct04: Looking at the detected radar total power vs time
28sep04: az dependence of faa radar compression in alfa receiver.
01aug04: alfa signal compression by the faa radar.
29apr03. detect 1290,1350MHz radars at 1 ms with lbw. Show rotation drift and compression.
29apr03. IFLO (using lbw,lbn) compression by the 1330/1350 FAA radar .
13apr03. azimuth, za dependence of 1350,1290, and aerostat radars.
04apr03. Radar power levels in the downstairs IF (1290,1330/1350,punta salinas.
04oct01. Radar harmonics created at the 2nd IF.
Information on some "other" radars.
1997/1999 measured lband radar ipps (report)

Old FAA radar specs: (top)

The radar specs listed below were measured back in 1997-1999


transmitter	FPS - 66 (i'm not positive about this)
rf frequencies	1330, 1350 MHz
Pulse len	6 usec 1330 MHz followed by 6 usec 1350 pulse. Each freq a different polarization
5 ipp pulse sequence used	[2633.08, 2821.00,2745.96.,2594.84,3309.96] Useconds
bandwidth	1/6 usecs
Rotation period	Nominally 12 seconds (one measurement showed it to be 11.982 secs).
Blanking	The radar does not blank its transmitted signal. We have a blanker that can blank our back ends.
Sidelobes
Location	Punta del este
distance to ao
Misc	az beamwidth 1.3 deg antenna gain 35 db Pk Pwr. 2Mw AvgPwr 4.3Kw receiver sensitivity -114 dBm

Modeling the faa total power. (top)

Pulsar observers have spurious periods introduced into their search data by the faa radar. A model of the radar was used to compute what these periods shoul d be. The model contained;

The pulse widths for the 1330, 1350 radars
The 7 ipps used for the radar.
The 12 second rotation period with a gaussian beam.

The plot shows the spectrum of the total power for the faa radar with some p2030 data overplotted (.ps):

Top: the spectrum of the total power for the model (red) and some p2030 data.

the pulsar data was smoothed to 1 khz.
The dashed blue line is the limit for the reported model periods.

Bottom: A blowup around the strongest period: 354.50 hz.

The black line is the data.
The red line is the full model
the blue line is the model with no beam rotation.

Table of periods and amplitudes:

The table on the right shows the frequency and relative amplitude for the periods generated by the faa radar model.

Table of periods and relative Amplitudes
Frequency order		Amplitude order
Freq Hz	Amplitude(rel)	Amplitude(rel)	Freq Hz
141.83	0.06	1.0	354.50
212.67	0.10	0.68	709.00
283.58	0.11	0.45	850.75
354.50	1.0	0.43	779.83
425.42	0.20	0.25	496.25
496.25	0.25	0.25	921.67
567.17	0.24	0.24	567.17
638.08	0.14	0.20	425.42
709.00	0.68	0.14	638.08
779.83	0.43	0.11	283.58
850.75	0.45	0.10	212.67
921.67	0.25	.07	992.58
992.58		0.06	141.83

processing:usr/p2030/rfirdr/faa/pltfaafit.pro

may12: faa radar ipps jump by .3% (top)

The observatory has a radar blank that phase locks with the 5 ipp sequence transmitted by the FAA 1330/1350 airport radar. It allows the spectrometers to blank their integrators during the radar pulses.

Time line of faa radar blanker problems:

02may12: a2335 reported the radar blanker not blanking the 1350 radar.
03may12:a2335 reported intermittent operation of the radar blanker.
04may12: I checked the blanker during the morning and it was locked.during the morning, the radar blanker was locked. I took some data.
05may12: a2335 reported blanker not working:
07may12: Blanker not working in the morning so i took some data.
08may12:talked to faa radar tech (martin medina). He said he'd go out to the site tomorrow so we could do some tests.
09may12:tested each channel of the faa radar. channel 1 generated correct timing. The radar blanker would not lock when channel 2 was generating the timing.

Data collection to check the faa radar ipps:

To see what was going on, data was collected using the faa radar monitor helix on the platform.
The setup was:

The detected output from the faa radar helix was fed through an op amp and then into the radar interface.
The RI was setup to sample at 1 usec with 8 bits.
No extra time constants were applied to the input signal.
Data was taken for about 100 seconds.

On 04may12 the radar blanker was locked to the signal (radar operating normally).
On 07may12 the radar blanker was not locked (the radar was not operating correctly).

The data was searched for the ipp periods of the radars.

The plots show the measured radar ipps:

Top plot: a histogram of the ipps that were found in the data. This uses 1 usecond bins.
2nd frame: showing the first few sets of ipps at the start of the dataset
3rd frame: Showing set of ipps at the end of the dataset (98 seconds).
04may12 radar working normally (.ps) (.pdf)

the ipps matched the historic values.
to get the folding to work after 98 seconds, the cumulative ipp was 14104.5 usecs

07may12 radar had different ipps (.ps) (.pdf)

The ipps are now off by about .3%
After 98 seconds of folding the cumulative ipp was 14057.72 usecs

Dynamic spectra of the first 16 seconds of data

The images show:

+/- 100 usecs of data was taken about each ipp arrival time
This was repeated for every 14058 usecs cumulative cycle of the 5 ipps of the radar

So the leftmost vertical line is the first ipp, the 2nd vertical line is the 2nd ipp ... etc.

The cumulative ipp was modified to align the data with radar pulses.

04may12 dynamic spectra when radar was working correctly (.gif)

the ipp used to align the pulses was 14104.93
The horizontal pulse near 5 seconds is the multi path scattering when the beam points at AO.
The fuzz at 11 seconds is the back lobe of the radar (pointing away from us)
The diagonal line near 2.5 seconds is about 60 deg from the radar pointing at ao. This may be the radar pointing at the airport (with planes overhead).

07may12 dynamic spectra when radar ipps were off by .3% (.gif)

The cumulative ipp used to align the pulses was 14057.0900. This is off by .3%
the radar back lobes are barely visible around 5.5 secs.

processing: x101/120507/faardr_07may12,faardr_04may12.pro

09may12: Radar tests with radar tech at the radar site:

On 09may12 the faa radar tech (martin medina) went out to the radar site to do some tests with the radar.

He first explained how the timing for the radar pulses worked:

There are two channels (klystrons). One for 1330 and one for 1350.
Each channel has dip switches that set the frequency for the longest ipp of the sequence of 5 (normally 3310 usecs).

This ipp controls the generation of the 5 ipp sequence

When two channels are used, one is the timing master and one is slaved to the timing of the master.
If the channel that is master has a problem (glitch, dropout,..) the radar switches to have the to other channel as master.
He thought the problem we were seeing was that one of the channels had their dip switches set incorrectly after the last maintenance session.

The test we did with the radar were:

Before he started, the radar blanker was locked.
He turned off channel 1. With only channel 2 running, the radar blanker was unlocked.
He turned on channel 1 and turned off channel 2, the radar blanker locked ok.
He then switched back to both channels, with channel 1 generating the timing pulses. The radar blanker remained locked.

Summary:

Radar blanker unlocked somewhere around beginning of may12.
The ipps were off by about .3%
The unlocking was intermittent.
After contacting the faa site, the radar tech went out to the site to do some tests.
Looks like the dip switches to generate the ipps are set correctly on channel 1 but are set incorrectly for channel 2.
If a channel (tube) has trouble (glitch,dropout) the system will switch to use the other channel for timing.
Whenever channel 2 is the timing master, we have trouble with locking.
The tech told me that they need to shut down channel 2 to check the dip switches for the timing.

To shutdown the system he needs to get permission from the DOD
He said he'd notify me by email when things got straightened out.

For the last few days (11may12..) the radar blanker has been locked. There is no guarantee it will stay this way if the radar jumps to channel 2 for timing master.

13apr07: faa radar ipps back to standard values. (top)

The the ipps for the faa 1330/1350 radar jumped by .4% on 07apr07. This caused our radar blanker to lose lock. We reprogrammed the blanker with the new ipps on 11apr07. On 12apr07 the blanker (with the new ipps) lost lock again. On 13apr07 i remeasured the ipps and they were back to the original values. We reprogrammed the radar blanker with the original ipps and it reestablished lock.
The plots show the measured ipps of the faa radar on 13apr07 (.ps) (.pdf):

Top Plot: Histogram of the ipps found. The measured values are now the same as the "old" standard values.
Center Plot: The detected output at the start of the dataset. The colored dashed lines are spaced by the ipp sequence.
Bottom plot: The ipp sequence at the end of the 16 seconds of data. The dashed colored lines are the ipp sequence using the phase from the start of the file.

Dynamic spectra of 16 seconds of data (.gif):

The horizontal axis shows 100 useconds about each of the 5 transmitter pulses in the 14105. usecond sequence (the image would have been too large if i had plotted all 140105 points of each sequence.
The vertical axis in time in seconds.
The ipp sequence length used to make the image was 14104.875. The radar pulses do not drift with this period.
The radar was pointing at the observatory around second 12. The right going signal is the multipath from the ground clutter.

processing: x101/070413/faatst.pro

10apr07: faa radar ipps jump by .4%. (top)

The observatory has a radar blank that phase locks with the 5 ipp sequence transmitted by the FAA 1330/1350 airport radar. Sometime near 07apr07 the radar blanker failed to lock with the radar. To see what was going on, data was collected using lband wide. A 5 MHz filter bout 1350 MHz was square law detected at 2 useconds and then sampled at 1 usec for about 100 seconds. The data was then searched for the ipps of the radar.
The plots show the measured ipps of the faa radar (.ps) (.pdf):

Top Plot: Histogram of the ipps found. The table below shows the new and old ipps. The ipps decreased by about .4%.

NewIpps	2586	3299	2624	2812	2737
OldIpps	2595	3310	2633	2821	2746
ippDif	9	11	9	9	9

The total ipp len for the old sequence was 4105 Usecs. The total ipp len for the new sequence is 14058 usecs. Looks like there reference oscillator has jumped by a bunch. Data was taken on the previous day (09apr07) and we measured the same New ipps. So the oscillator is relatively stable after the jump.

Center Plot: The detected output at the start of the dataset. The colored dashed lines are spaced by the new ipp sequence.
Bottom plot: The ipp sequence at the end of the 100 seconds of data. The dashed colored lines are the new ipp sequence using the phase from the start of the file. The pulses have drifted off by about 1.3 millisecs in 100 seconds. The correct sum of the 5 ippsequence is 14057.8113 useconds.

Dynamic spectra of 16 seconds of data (.gif):

The horizontal axis shows 100 useconds about each of the 5 transmitter pulses in the 14057.8113 usecond sequence (the image would have been too large if i had plotted all 14057 points of each sequence.
The vertical axis in time in seconds.
The ippsequence length used to make the imagege was 14057.8113. The radar pulses do not drift with this period.
The radar was pointing at the observatory around 8 seconds. The right going signal is the multipath from the ground clutter.
The point at 9 seconds is probably a plane

Solution:

We reprogrammed the radar blanker with the new ipp sequence and the blanker relocked with the radar.
on 12apr07 I talked to the radar technician and he told me that they were having problems with an inductor on the 1350 transmitter. He had switched the 1350 transmitter to low power.
The tech also noticed that the a/d converters for the timing all had red lights. He then switched the radar timing to the other transmitter (looks like each had its own set). When he did that, the blanker (with the new ipps) no longer locked up.
When the technician gets things straightened out we'll switch back to the old ipp sequence in the blanker

processing: x101/070410/faatst.pro

apr08: Plots of when the 1330 and 1350 MHz faa radars are transmitting. (top)

The hilltop monitoring data was used to see when the 1330 and 1350 frequencies were used by the faa radar. The plots use black for the 1330 frequency and red for the 1350 frequency. The vertical scale is dBm measured at the hilltop monitor spectrum analyzer. When the value drops below -50, that frequency is not used. When the value drops below -60 , the hilltop monitor system was not recording. There are normally 108 measurements per day of the radar frequency band.

2008 FAA radar (.ps) ( .pdf)
2007 FAA radar (.ps) ( .pdf) (jan thru may only)
2006 FAA radar (.ps) ( .pdf)
2005 FAA radar (.ps) ( .pdf)

processing: x101/070305/chkfaa.pro

page_up
home_~phil