LNA plus t-bias

As the title say, the LNA comes from ebay (20 dB gain and a 0.6 dB NF up to 3GHz they promise), it gets the 5V via the a t-bias so that I don’t have to pull another cable up into the mast.

LNA board from e-bay on the left, t-bias block from minicircuits on the right, this is also the part where you put the SDR with its 5V t-bias feed.

What you need are PL to SMA adapters for chassis mounting (hard to find), use hot glue to seal the box (the previous owner made already two holes which were too tall of course) and leave the SMA plastic caps on while you seal the remainder with hot glue. You also need a SMA to SMA cable (which I make and test myself), a 100 nf capacitor over the t-bias output, and a short shielded cable from the t-bias block to the LNA board,  and finally you need some shrink tube. Without the t-bias on port 2 you should hear nothing, put it on and you should see some action. The gain of the rspduo should be adjusted because the noise floor will also come up by about 20 dB. Purpose of the LNA is twofold:

  • do something against the cable losses, for ADSB it matters (see previous posts)
  • increase the sensitivity of the receiver (this required for some SDRs like the USRPs).

Last update: 23-April-2019

ADSB receiver differences

Do the same thing on an airspy/ADSBspy setup and a rspduo/dump1090 combination, and yes there are differences in performance:

difference airspy and rspduo

The airspy result is always within the rspduo result, the latter is a newer receiver, I got the airspy in 2015, so it is probably it is an early release, the rspduo is two months old. The airspy software is somewhat easier, it simply always works, and with the rspduo/dump1090 I oftentimes need to restart it until the ADSB results become visible in the virtual radar server.

The difference is range is approximately 50 km in range at 200 km distance, translated into dBs using the path loss equation I guess this is 2.3 dB in the receiver sensitivity favoring the rspduo. Receiver sensitivity is actually not really tested because we are using different decoders.

Last update: 7-Apr-2019 8:32

ADSB antennas

With the new mast I can easily test different ADSB antennas. I already reported a first test with a low gain antenna in this article which looks like this:

lowgain antenna
Figure 1: the low gain antenna with 0 dB gain

With the mast the situation is slightly different, first because there is more coax cable between the receiver and the antenna, it is Aircell 7 cable which has a loss of 0.2 dB per meter at 1090 MHz. So with 50 meter that is 10 dBs going up in thin smoke, which is really a pity but unavoidable if you haven’t got the LNA (yet) to compensate the cable loss. The easier way out is to build a high gain (9 dB) collinear antenna like explained here. I added a 1/4 lambda sleeve balun and a 1/4 lambda whip to the tip, in the end it looks like this:

highgain antenna
Figure 2: High gain antenna with 9 dB gain (to calculate via 10 log10(n) where n is the number of elements in the collinear antenna (here n=8)

The high gain antenna is called high gain because of its expected radiation pattern which is more pancake like than torus or sphere like with the low gain antenna, so it favors more the signals at low elevation.

rspduo lowgain_03apr2019_a
Figure 3: Low gain performance, so you see aircraft data up to 70km
highgain 6-apr-2019 ADSB
Figure 4: High gain performance, with reduced coax cable length (I managed to bypass 20m) You get to see aircraft data up to 250 km

Figure 4 shows the best result that I got so far, namely, with the high gain antenna and 20 meter reduction in length of coax cable for which I had to drill a new hole in the ceiling. We improved the range by almost a factor 3. To summarize:

  • Low gain antenna, 50m cable, max range is about 70 km
  • High gain antenna, 50m cable, max range is roughly 100 km
  • High gain antenna, 30m cable, max range is about 250 km.

Since we know that the loss is caused by Aircell 7 cable we also know what the difference is between the last 2 results, it is 4dB which is the same as a factor 2.5 in the free space loss equation where the loss = (4 pi d / lambda)^2 with d denoting the path length and lambda the wavelength.

For the 30 remaining meters we should get 6 dB of damping, and this would be a factor 4 if it were to improve the path loss. Curvature of the earth and the height of the involved antennas constrain the maximum range to something like 430 km, so the factor 4 will never be obtained.

Maybe another suggestion is to not use coax cable at all and place the SDR receiver very close to the antenna, and then feed it with USB cable or turn it into a wireless device with a raspberry PI and a SDR. This is a personal choice that people have to make, I want to be able to use my SDRs for various experiments, so this is not an option.

There are various things to learn from this experiment:

  • 1 GHz is a frequency where cable losses become a serious factor in the design
  • The antenna gain is relatively easy to improve
  • The antenna pattern shows that we are easier affected by obstruction, for instance, the south west sector has more obstruction not because the antenna is below the rooftop heights (it is above it), but according to me due to the city and the port area in that domain.

Last update: 6-Apr-2019