FlightAware Discussions

ADSB weakest bBFS range

I was looking at some traffic this morning and randomly checking the signal strengths of nearby and distant aircraft. It seemed on my setup that right around -32.2 dBFS is where a/c would reliably stop being received, regardless of distance. The distance where this actually happens varies from day to day as atmospherics change and plane direction of travel varies, etc.

I’m curious as to how weak a signal folks are able to receive and theoretically how sensitive the Flightaware hardware ( blue FA Pro stick ) is in practical terms. (I know the stated capability is “over 300 nm” but I think I read where that is also subject to some software audit/rejection in terms of Flightaware )

1 Like

The dongles are actually quite sensitive receivers, however the problem is not usually sensitivity in this application but dynamic range. Aircraft transmit with powers up to 500W, so a signal from the furthest possible distance allowed by terrain should be plenty strong enough to decode. The complication is that there are usually many aircraft much closer than that, whose received signal strength will be much higher. Being able to receive the very strong signals from close aircraft without over-saturating the dongle while also retaining the ability to decode longer ranged signals is something of a compromise.

The RSSI figure in dump1090 is affected by internal amplification, and is not the absolute received signal strength so it will change when you alter the gain setting. The practical limit for received signal strength does appear to be about -35dB or thereabouts for RTL dongles.

You need someone like Oliver or prog to explain the details of exactly why this is, since RF voodoo is not really my field, but this page has some technical details about the dynamic range of software radios that you might find interesting.

2 Likes

This kind of matches my experience but it would be helpful to know in this particular application where most users are actually seeing signal dropoff.

Regardless of terrain, height, etc. which are physical line of sight clipping elements, there seems to be a level at which the signal is no longer able to be processed. The last known good packet is processed and the timer increments to “n” and is discarded from display.

Well this is typical for my dongle with the rtl-sdr.com LNA:

1 Like

Well the sensitivity of the dongle can actually be much improved by using an LNA.

Depends on the noise floor.
Which is combined from noise picked up by the antenna and the noise of the dongle.

Typical with that gain setting.
If you increase the gain the weakest signal will be stronger.

But there is of course a minimum level which can be found by reducing the gain further and further.

Yes - That setting is pretty much optimal for that combination with the current antenna though. Further reduction harms range.

1 Like

You seem to be running pretty low gain already, do you actually notice tracks missing for aircraft with strong signal when you turn it up?
And what’s your typical >-3dB percentage with those settings?

It’s hard to say for sure - I didn’t notice a huge difference in missing tracks, but increasing the gain much more did result in an increase in single message tracks, which means that there’s likely more bogus messages being generated due to noise or garbling.

Typical strong signals are about 4% over a day with the gain set to 29.7. I disabled the autogain script eventually because there would be the occasional busy day that caused it to drop another step or two, and that did cause a noticeable reduction in range and messages. I do have a pretty long coax at the moment, so with a shorter length I’d expect to drop that down another level or so.

It would most likely only happen for aircraft flying below 5000 ft and closer than 10 km.

Anyway if increasing gain doesn’t give you a numbers improvement, no reason for doing it.
Might be worth setting it one notch higher for a week or so :slight_smile:

I already have a bit of data for that:

There’s very little in it. Marginally more aircraft with the slightly higher gain, but it’s close enough that it could just be variation in traffic. Slightly higher message rate with the lower gain.

It’s kind of moot anyway now since I’ve just replaced that dongle with an airspy mini.

After I rebuilt my station after last years hurricane the graphing scripts were not put back on, so going to add some of the scripts to mine to gather data again. I have no quibble with my range, dropouts, reception in general, but as things age and weather ( sun, salt water air, high wind) I’m curious what the noise and signal stats are now vs at install time. Nothing lasts forever in that climate.

You shouldn’t see much change in performance just from age - one thing that is very important for an outside installation though is making sure all the connections are watertight.

I’ve pulled many cables from boats where the copper is corroded all the way along the length because of improper waterproofing. That will definitely cause a degradation in performance over time. That’s in a saltwater environment though which is much harsher than inland. I make a point of putting silicon grease in any outside connectors now - it keeps water out and inhibits any surface corrosion.

Consider these: https://github.com/wiedehopf/graphs1090#graphs1090

adsb receiver takes over the complete web server directory, the above doesn’t.
Also as you see in caius graphs, i’ve added some more stuff.

2 Likes

Since Im at the site now, there’s no time like the present :slight_smile:

I don’t have much to add here. The minimum (post amplification) power that the dongle’s ADC can distinguish is theoretically somewhere around -50dBFS. The -35dBFS-ish limit is mostly a demodulator limit (there’s no hardcoded limit, but below that there’s just very little useful data to use, <2 bits, you’d need close to perfect reception and a lot of luck with timing to see anything weaker)

2 Likes

So reinstalled the performance graphs as @wiedehopf suggested and ran for a week or so. % >-3Dbfs = .1 Here’s the signal:

dump1090-localhost-signal-14d

But also the noise floor raises with the gain. There is a sweet spot of the gain, where the SNR is best, in my tests with the FA dongles (fitted with internal preamps) it was when the “gain” was around 35-38dB.

Also take into other considerations, your location, obstructions, CELL phone towers, trees, locally generated RF noises. Trees is one of my big problems. I see the difference between the times when the trees have no leaves and when they have the leaves. Locally generated RF noises reeks havoc on my SDR Play reception on HF. If I unplug the power supplies to my raspberry Pi’s the SDR Play quiets up.