I have noticed that in my case these long distance spikes mostly occur during nighttime when there’s less traffic.
As @wiedehopf stated it was a “noise” coming from a plane which never flew on that route i’ve seen it.
Clearly a fault somewhere. Maybe they need some experts 
I have such spikes every few days in my graphs
Yeah, dump1090 has some stuff to try to filter out the worst of this case, but there seem to be a few different ways it goes wrong and some are harder to match. Agreed it’s all garbage position data though.
I am fortunately away overseas on holiday so i can’t dig out the data to show what i found.
I fed the output of the Pi port 30005 to ADSB scope and plot it.
My understanding is that dump1090 listens continuously and if the noise looks reasonable it is passed as a valid data packet.
These occasional packet looks legit but may not be.
Unless you can identify that sighting by its ICAO number it IS just noise regardless of what your graph shows. Once you have an ICAO number you can at least check to see if it is possible that plane was in that neighbourhood at that time.
You could monitor port 30005 and calculate your distance from each location packet and report anything over say 350NM and then test their legitimacy.
S.
It’s not RF noise; it is a bad transponder transmitting a correctly-checksummed message that contains a garbage position.
1 Like
Is this always the case?
Are we talking about the same example?
I was referring to the graph @foxhunter provide a few posts back. How do we know that is a bad transponder?
Thanks,
S.
What else should it be?
In my case the typical range never exceed 160 NM.
So any signal in the range of +400NM as i have posted above must be a kind of hoax
It’s simply not possible due to my geographic environment
As i wrote above i occasionally receive a location packet from over 400NM where i can identify the aircraft AND establish that the aircraft was in that vicinity at that time.
Why do you believe it is a hoax and who would perpetrate such a hoax?
S.
Tropospheric ducting and reflections off aircraft are two prorogation phenomena that can deliver signals from well beyond your normal range.
No, but it’s the case with the BCS aircraft mentioned above that I was talking about. While I was in Cambridge I’d see the same thing and it was not RF noise, it was a bad transponder. IIRC they’d transmit a “position” with longitude close to 0 and a wildly varying latitude. If you happen to be a few hundred NM from the 0 meridian (or if you’ve disabled the max range check) then some of those positions are close enough that dump1090 considers them plausible but they’ll show up as spikes in your apparent reception range.
Here is the test in dump1090: dump1090/mode_s.c at master · flightaware/dump1090 · GitHub
This filters out the worst offenders but there are other similar failure modes that don’t get caught by that (the test is deliberately very conservative to avoid throwing away any real positions - the values being transmitted are possible valid positions!). I considered an aircraft blacklist but populating and maintaining it looked like a lot of work.
Filtering on (lat==0 or lon==0) seems like it wouldn’t discard too many valid positions, as the GPS has quite a fine resolution.
But if you really don’t want to throw away positions, then that’s too broad of a filter obviously.
How likely is it to get a valid position decode from RF noise?
You need a pair of messages for a position, so i’m not sure how it could happen.
compare it with sites like FR24 then you know if the aircraft exist at this position or not.
In my case it did not.
A completely spontaneous valid message is highly unlikely, but a message with sufficient corrupted bits that still passes CRC is definitely possible. I don’t think the error correction is particularly robust. When you consider that you could see well over 100 million messages per day, a few dodgy ones becomes much more likely.
1 Like
I haven’t worked out the details but “pretty unlikely” is the short answer; the raw rate of messages that incorrectly pass CRC is higher than you’d think because the demodulator throws a lot of noise against the wall so eventually some does stick (IIRC my back-of-an-envelope calculation said something like 1/minute), but you need a pair of messages that appear to be from the same aircraft within a short time window with somewhat consistent contents, which is much much less likely.
Do you need the pair of messages to pass it to piaware and then flightaware or do you need a pair before it is output as a location packet on port 30005?
Where is the dats for the graphs obtained?
The discussion based on a report in the graph at > 350 NM. I thought piaware dropped locations that far away.
I’m half way around the world from my home time zone. Forgive me if i have it all wrong. Helsinki is really nice.
S
30005 will be raw messages, not positions.
So whoever receives the raw messages will have to decode the position.
The graphs get the position from the aircraft.json, just like the local map.
Maximum range which is not discarded by dump1090 is by default 360 nmi.
But you can adjust that in the settings.
In regards to the raw messages, those are obviously still output on 30005 and not discarded.
Piaware and other feeders or applications receiving the beast data will have to do the position decoding themselves and decide what is valid and what is not.
1 Like
Another one from tonight, clearly outside my maximum range:
