faa radar aliasing in band
A2010 radar harmonics measured 24mar05:
When 100 Mhz 3 level data centered at 1384 Mhz is taken (a2010),
radar birdies appear in the band. The plots show data taken 24mar05:
spectra for scan 508300566 (.gif): The top plot is pixel 0 polA while
the bottom plot is pixel 0 polB. The birdie at 1350 Mhz is the Faa radar.
You can see another birdie at 1380 Mhz. There are weak birdies at 1404.8
and 1410.4 (1420 Mhz is our galaxy).
rms/mean by channel for scan 508300566 (.ps) (.pdf)
: Black is polA while red is polB. The expected rms/mean by channel is
: (1./sqrt(100e6/4096*2))=.005 (*2 since the input data was hanning smoothed).
The measured value is a bit higher since continuum sources will raise the
rms/mean a bit. It is easy to see the birdies at 1350,1380,1404, and 1410.
Test data taken 03mar05:
On 3mar05 i took some data with the wapps and alfa.
The setup was 100 Mhz 3 level sampling, 1 second samples, with cfr at:
shows 60 seconds of data taken at each cfr (.gif).
A plot of the total
power at 1350 and the birdie frequency (.ps) (.pdf)
The weak birdie can be seen every 12 seconds.
the birdie lines up with the maximum of the faa radar pulse.
the large red pulse is when the radar points at ao. The smaller red pulse
6 seconds later is when the radar is pointing away from ao and there is
horn spill over on the faa radar antenna.
Data taken by a2010 when the 1350 radar is off (but the 1330 Mhz
is on) does not have these birdies present so they need the 1350 Mhz signal.
The table below shows where the 1350 Mhz birdies sits in the 1st IF
and at baseband. It also includes the location of the baseband 2nd and
3rd harmonics of the 1350 Mhz birdie.
|1350 birdie at 1st IF
(cfr-1350) + 250
|1350 birdie at baseband
(cfr-1350) + 50
2nd harmonic at baseband
|1350 birdie 3rd harmonic
|sky frequency of birdie
of the 1350 birdie in the 1st If and at baseband is also shown in the figure
(.gif): The plot displays the locations when the center freq is 1380
Mhz (left panel) and 1385 (right panel). The bandwidth was 100 Mhz so the
aliases at baseband starts at 100 Mhz. The blue lines are the normal mixing
locations for the 1350 radar. The Red lines show where the baseband 2nd
harmonic of 1350 occurs. The green lines are the 3rd harmonic.
When the center frequency is 1380 Mhz, both the 2nd
and 3rd harmonics at baseband alias down to 1390 Mhz. Shifting the center
frequency to 1385 and then 1390 rules out the 2nd harmonic. So the alias
of the radar is the 3rd harmonic.
1380 Mhz top: the location of the birdie in the 1st IF is 280 Mhz. The
location at baseband is 80 Mhz.
1380 Mhz middle: The baseband 2nd harmonic is at 80*2=160Mhz. This
mixes back down to 10 Mhz below the center (40Mhz). Since the band is flipped
at baseband this is a sky frequency of 1380+10=1390 Mhz.
1380 Mhz bottom: The baseband 3rd harmonic is 80*3=240. This then
aliases down to 10 Mhz below the center (40Mhz). It also corresponds to
a sky frequency of 1390 Mhz
1385 Mhz top: the location of the birdie at the 1st IF is 285 Mhz. The
location at baseband is 85Mhz.
1385 Mhz middle: The 2nd harmonic of the birdie is 85*2=160 Mhz. This aliases
down to 20 Mhz below the center (30 Mhz). It corresponds to a sky freq
of 1405 Mhz.
1380 Mhz bottom: The baseband 3rd harmonic is 85*3=255. This then
aliases down to 5 Mhz above the center (55 Mhz). It also corresponds to
a sky frequency of 1380 Mhz.
This could also be occurring at the IF1 in the amplifier
after the mixer (and filter). The 3rd harmonic of the IF ends up at the
same location as the 3rd harmonic at baseband.
The 1380 radar birdie comes from the 1350 Mhz birdie.
It is the 3rd harmonic in the A/D (or in the IF amp after the mixer).
If it is in the IF, it is being created after the 200-300 Mhz IF filter
(in the 20 db amp).
When the 1350 radar is off , the 1340 Mhz radar birdie remains.
If you placed another filter after the 20db amp and the birdie went away,
then you'd know it was in the amp and not the a/d. This filter would not
need that many poles since it is trying to filter things that are an octave
This problem will worsen when we move to 300 Mhz continuous bandwidths.
If the amp after the 300 Mhz IF filter creates harmonics, then all of the
harmonics of all of the radars will be in the band.
The idea of using a hi side lo to throw the radar IF harmonics out of the
sampled band relies on the IF filter to remove the harmonics. If there
are active elements after the filter that go none linear then this idea
will not work.
It is important to know if the a/d itself will cause these harmonics. If
so then you're stuck with them (unless you find an a/d converter with more