FlightAware Discussions

Receiver for close proximity traffic - "airport feeder"

Hello all. I have just set up a feeder (RBPi 3, Nooelec Mini2+, GNS Ten-90 ) antenna at the local airport to improve low level traffic coverage. At first the antenna was located indoors but yesterday I moved it outside and up on the roof. A trial location, it will be better placed later. I directly noticed better coverage where aircraft on ground would be reported while taxiing but when I moved the antenna to the roof RSSI ventured up to -2, -1 and the reciever would not pick them up.

I have read some in these forums and come to the conclusion that 2 options might be useful. Since I am in an already well covered area there is no need to pick up planes far away. So either use a simpler antenna such as mag-plate stick antenna, or work with the gain. Since the feeder is remote without web access I’m experimenting with another feeder that has shown the same issues, local aircraft simply are not reported.

To make a long story shorter: I am currently trying with a gain setting of 29.7. Is lower the way to go - will lower gain “automatically” give better coverage on planes close and low? Or would a less refined antenna give the same result? I´ll continue trying but those of you who have already been down this road… any thoughts?

Just reduce the gain and the RSSI should reduce as well alleviating the issue.

The GNS antenna can’t really have the promised gain as it’s too short to have any better gain than a simple quarter wave antenna.
Anyhow i wouldn’t change the antenna, changing the gain should be more than sufficient.

1 Like

So… how low is low? Below 15 or below 30 or…? I changed my dump1090-mutability setting from “max” to “agc” and then to 29.7, it is visible here if you look at todays graphs.(Yes I am still on mutability on that reciever, the others are dump1090-fa.)

So the more negative the RSSI the lower the signal.
As long as the signal isn’t greater than -2 dB it will not oversaturate usually.
There is no fixed gain setting to recommend, you adjust the gain and then check the RSSI you get as a result and adjust the gain so the RSSI you get doesn’t get above -2 dB or so.
Or you adjust it won’t go above -5 dB if you want to keep a good margin.

Looking at the graphs you can probably run 36 and it should be fine.
But the range isn’t bad right now so i’d just leave it at 29 for the time being and check if it’s still cutting out.
That way you should have plenty of head room when aircraft taxi close to the antenna.

1 Like

Lowering the gain greatly improves the odds of decoding strong local signals. SDR dongles are incredibly sensitive, and they also have a fairly broad “front end”.

Another problem on an airfield (and a few other places) is that there may be nearby strong out of band signals that still are inside the front-end pass-band of the receiver. These strong signals can swamp the circuitry and cause enough distortion that the desired signals are not properly decoded.

Three possible solutions:

  1. reduce the gain until you start seeing aircraft on the airport
  2. install a 1090 mhz filter between the antenna and receiver to specifically exclude strong unwanted out of band signals that can cause “desense” from overload or saturation of the front end. The problem is much like trying to hear a conversation in the car at a red light while someone is booming and thumping in the next lane with their high powered stereo system. While much of the racket is not in the same frequency spectrum as the person you are conversing with, it still introduces so much noise that, while you might hear the person speaking, you might not be able to interpret enough syllables to figure out what they said.
  3. put an attenuator in the feedline in addition to the 1090 mhz filter. This externally cuts the signal down to a manageable level, so the SDR has a better chance of decoding strong signals. This has a second effect of bringing distant aircraft signals below the detectability level, reducing clutter in the feed since your goal is to fill in local signals on the airfield.

Good luck with your setup. The same ideas apply for feeders at hospitals to capture helicopters coming and going, or anywhere else that the aircraft will be extremely close to the receiver.

2 Likes

Thanks n5psp. I’ve reduced (or set) the gain to 32. It reduces the area covered but as I wrote the region is well covered for high altitude commercial flights. In the following days I will keep an eye on the result, and adjust accordingly if necessary. The current setup will probably be replaced by a better antenna (Radar-110?) and a FlightAware orange stick, that will require new tests.

Bear in mind however that the orange stick has a preamp built in. This, coupled with a better antenna may have the opposite effect to what you want, making local traffic more likely to overload. Without a filter, you may also be more likely to suffer from overloading from your local ATC transmissions, mobile phones etc. If you only want local traffic “better” equipment may not be what you want.

1 Like

DVB-T’s front end covers a frequency range of about 24 MHz to 1800 MHz.

Strong signals which are even far away from 1090 MHz, but are still within 24 MHz to 1800 MHz range, can overload and de-sence the DVB-T’s front end.

It is wise to determine what and how strong RF signals exist in this band at your location .

 
 

Spektrum - How-to Speedily Scan RF Noise in band 24MHz ~ 1800MHz

 

Scan 1 of 3 - FA Antenna + Generic DVB-T (no internal or external filter)

Thumb-Generic DVB-T
CLICK ON IMAGE TO SEE LARGER SIZE

Scan 2 of 3 - FA Antenna + ProStick Plus (Only Internal filter of ProStick Plus. No External filter)

image
CLICK ON IMAGE TO SEE LARGER SIZE

 

 

Scan 3 of 3 - FA Antenna + ProStick Plus (with internal filter) + External Filter (FA Light Blue)

imageimage
CLICK ON IMAGE TO SEE LARGER SIZE

 

 

1 Like

I’m not sure it’s worth it trying to “improve” that setup.
The dynamic range of the SDR will be the same as it’s the same base ADC chip being used. Thus with the FA sticks you’ll just need to reduce the gain further but really nothing much will have changed.

Adding a filter is the only thing that’s worth considering in my opinion.
Just keep increasing the gain until you have dropouts again then back it off a bit.

If it’s currently at 32 … looking at the graphs … try how 38 works.

I suppose adding an antenna that has higher gain (and therefore higher directionality towards the horizon) might be an improvement.
But i’m not sure it’s worth it.

If you’re going to spend money it you could probably run two SDRs one with higher gain and one with lower gain. But that requires quite a bit of software setup and might not suit you.

You could go to a receiver that has higher dynamic range like an airspy.

But before doing anything i would compare actual coverage to the theoretical coverage: GitHub - wiedehopf/tar1090: Provides an improved webinterface for use with ADS-B decoders readsb / dump1090-fa
tar1090 should even work with dump1090-mutability but you could just change to readsb or dump1090-fa.
An easy way to do that would be the automatic install scripts referenced here: Raspbian Lite: ADS B receiver · wiedehopf/adsb-wiki Wiki · GitHub

1 Like

You can also use an attenuator if you still have signals that seem too strong. It gets kind of involved detailing how to build one, but there might be a 10 dB, 20 dB, or 30dB broad spectrum attenuator out there. It just inserts much like a filter or pre-amplifier. Inside it is just a resistor array, small enough to not produce its own impedance bumps. Since you aren’t transmitting, you don’t need a high power dissipation unit (those are not cheap).

Another gotcha on an airport with a lot of ground clutter, is multi-path signals.
This is where you get a signal directly from the aircraft, but the same signal also hits the side of a hangar or the vertical stabilizer of that parked King-Air at the pumps, and is reflected to your antenna. At UHF frequencies and above, all sorts of flat surfaces act like mirrors. Some act like poor quality mirrors so only a weak signal arrives with speed of light delay, while some act like incredibly good mirrors, and can deliver a delayed signal that is almost as strong (or even sometimes stronger) than the direct signal from the aircraft.

Since ADS-B data is sent at about a megabit, a delay of only 300 meters will smear the signal out more than a single bit interval. The two colliding signals then result in a mix that cannot be decoded by anything affordable for most of us.

The second multipath problem other than data smearing is “picket fencing”. This is where the primary and reflected signals alternately arrive in and out of phase, so they add and cancel out very fast as the aircraft moves.

The best way to initially deal with suspected multipath is to get the antenna as high as possible, ideally several feet above the top of nearby hangars and parked aircraft.

My experiences with other UHF / low microwave signals was to stick the (70 cm band, not ADS-B) antenna atop a 10 foot joint of rigid metallic electrical conduit, then secure that to the edge of the building roof with conduit straps. The one I set up required 2 more joints beneath it extending down the side of the building to the ground. We put a floor flange on the bottom so the weight wouldn’t cause it to try and sink into the dirt. We didn’t use any guy wires, just a couple of conduit straps anchored to the roof edge the same as the gutter, and it survived 85 mph winds that tore the roof off of another building about year later.

1 Like

Thank you all for the input, most interesting. I adjusted the gain at the airport feeder (remote location, no ssh access) and the result is so far very good. This was ADSB-equipped helicopter.


This site still is on 32 (@wiedehopf the graphs I linked to are on one of my other sites that have similar problem, still on 29.7, and you can easily switch to tar1090 on that adress…) and I need to see how MLAT aircraft are reported on the ground. Blue dot in the image is the antenna location and there are several buildings in the way that you can not see here.

We´ve backed down the equipment list and probably will go for Vinnant FE1090 and Nooelec Smart v4 or mini2+. More than enough, and there will be some fine-tuning to do. :slight_smile:

MLAT needs 4 stations receiving the aircraft to work.
Thus you’d need quite a few stations to get MLAT results from aircraft on the ground.

Looks like you nailed the worst of it.

MLAT works by coordinated time difference of arrival from multiple receivers. As wiedehopf said, you need at least 4 that can see the aircraft simultaneously to get sufficient data for MLAT calculations. So you’d need 3 other sites surrounding the airport that have a line of sight to the ground to get MLAT to work.

Ideally you’d want them at least a couple of miles apart, forming corners of a square or diamond pattern on a map with the middle of the main runway near the center of that.

That’s where I’d start researching VPN tunnel servers. One that I set up, although it’s on a ham radio club server that is on a static access and isn’t really designed for public use, is how I keep up with some remote APRS (ham radio) Raspberry Pi systems in remote locations.

There are a few ways to do it, and I’m looking to set one up a bit more cleanly than the way I did in 2019 to access a unit sitting on a mountaintop using solar panels and a battery for power, and a 4G router and yagi antenna to talk to a cell site about 40 miles away.

1 Like

side topic:

I have the GNS antenna here as spare, but it shows significant less performance than the other ones. The 6dBI are obviously a marketing joke.

Probably has to do with the direction that gain points, and whether it is referenced to a theoretical ideal dipole, or if it is referenced to an isotropic radiator.

I tend to look at most antenna claims as 99% hype, unless they include an honest 3D plot of the radiation pattern.

Antennas do not add energy to the radio signal. That is determined by whatever the transmitter feeds it. For receiving, they can not pull in more than what arrives. What they do is modify the shape of the pattern where they are most sensitive for receiving and where most of the transmitted energy is concentrated.

This comes at the expense of energy or sensitivity in a different direction.

Real world examples: A discone antenna behaves pretty close to an isotropic radiator. It also has broad bandwidth. These are often used atop air traffic control towers and other air to ground communication where the target can be anywhere between the horizon and directly overhead, as it hears and radiates almost equally in all directions. They are also used frequently for scanners because they work well across multiple bands.

I built my own 1090 mhz antenna. It is just a simple quarter wave, with downward swept counterpoise radials, partly because mechanically it is easy to make from 3/32" brass rod which solders nicely into the center terminal of an SO-239 or N connector, while the radials easily mount on the chassis flanges of the same connector.

The radiation pattern has its strongest area about 3 degrees above the horizon, and drops off with increasing angle until it has a slight null straight overhead. This can be lived with as most aircraft will be within a degree or two in elevation from the horizon, especially with very distant flights.

Higher gain antennas that are vertical whip types can include phasing coils and other tricks to effectively “stack” elements, and make the pattern look more like a flat washer than a donut or teardrop cross section up-tilted toroid. These perform incredibly well for land mobile two way radio because everything you want to communicate between ground vehicles is going to be fixed to the earth’s surface and thus mostly at the horizon.

Those really don’t perform well at all with satellites, which are only on the horizon for a very brief time. Aircraft communication is going to be similar to satellite, in that they can be high overhead, but because they are about 20x closer to the earth and moving 20x slower, have a lower fraction of the time in view where they are at a high angle overhead.

Thus, a simple quarter wave antenna is probably the best trade-off for most 1090 mhz reception, mounted high enough to see over trees and nearby buildings. Consult the ARRL Handbook as to why the radials should be bent downward at about 30 to 45 degrees (skipping a large chalkboard full of mathematical equations and 90 minute classroom lecture), as this pulls the primary lobe closer to the horizon. Otherwise you’ll miss a lot of aircraft out on the horizon.

That also gets the impedance closer to 50 ohms, which for receiving isn’t anywhere as critical as transmitting, but a better impedance match means more of the signal that does arrive at the antenna makes it down the coax and into the SDR, so weaker signals can be sniffed out of the airwaves.

1 Like

lots of text for my statement that there are better antennas out , even with lower values :slight_smile:

1 Like

Another upside of adjusting gain settings. Not that the RBPi 3 runs very hot but better reception while reducing load is nice.

dump1090-localhost-cpu-24h

Time for an update. Yesterday we - a work party of 4 - set about on getting the Vinnant FE1090-P up on the roof, run a cable and update the Pi with new hardware. It took a while and to make a long story short, when we were done we had a message rate of… 0.0-0.2 per second! Which was a ‘bit lower’ than expected. Well, it turned out to be two problems. The first one was that the crimping was not done right, of course on the antenna side so we had to take it down from the pole 10 ft above the roof. Lesson learned, always check your connection before mounting it. The second one was more difficult to spot but the NooElec Nesdr Smart v4 that was supposed to be used is most likely faulty. We get no errors but no data either, no matter what cable or antenna we use.

So we will get the ProStick Plus, because it is in stock and reasonably priced. There is a cell phone tower quite nearby that might affect reception so the filter will be handy. Now its connected to the Mini2+ and that works but the amount of messages is about 1/3 of what I get at home, just some 15 km away and comparable surroundings. Same antenna also.
It’s all in a days work for flightfeed installer-man!

Put on a filter yesterday. A “NMRF FBP ADS-B” that I bought ages ago on ebay. Traffic went up like a rocket. Now waiting for the Pro-stick and then I’ll start looking on gain when we have some data to analyze. :+1:

1 Like