FlightAware Discussions

HOWTO: Airspy mini and Airspy R2: Piaware / dump1090-fa configuration

The previous lines were two hours earlier, that’s why I didn’t include them

Sep 15 15:14:14 g0pkt-piaware airspy_adsb[31142]: CPU 75.2 %, target 85.0 %, adjusting preamble filter: 5 -> 6

Is the one directly above. I doubt that’s related.

No, it was several hours after the last decoder start.
Pi was booted days ago.

Did I induced any nightmares? Sorry! :rofl:

LE: For anyone there, Widenhopf stayed up late (his time zone) and helped me troubleshoot my lock-ups of airspy_adsb at reboot. I am grateful!

4 Likes

@wiedehopf is our hero :slight_smile:

3 Likes

2.2-RC15

  • -C cpu target algorithm improved to react much quicker and finer, will no longer announce adjustments in the system log. non-integer preamble filter possible again by timesplitting between two preamble filter settings
  • -P is a secondary preamble filter which is useful when encountering
    wrong altitudes or planes showing ground in VRS, it will only apply to
    non-CRC messages, such as most non-position messages. Of course the drawback is that the number of correct messages of that category are slightly reduced as well.
    (try -P 5 or even less if you run high -e settings and have an issue with wrong altitudes)
9 Likes

Have you stopped printing -e changes? After this update Graphs1090 still shows it but the log does not.

Changes are too quick to print them, it would flood your log.

1 Like

Good to know. Seems to be much more solid here. Good change. Thank you.

@wiedehopf Thanks for doing that! I’m anxious to experiment with it.

I pushed -C to 65 and all altitudes in VRS look correct - no GND!

-v -t 60 -f 1 -w 5 -e 4 -E 25 -C 65 -D  -P 5

airspy-localhost-misc-2h

How about impacts of that to message rate?

-e at 15, testing -P 2 and -P 1, not seeing issues. CPU usage down the same with either. And no GND in VRS watching for a few minutes with 45 to 50 aircraft. Normally at -e 15 I would see maybe 3 to 5 GNDs at any given time with 50 or so aircraft. Not a single GND seen yet. Message rate, who knows, but nothing obvious is visible.

I’m assuming -P 1 is less filtering than -P 5, like -W

chrome_tdzunK8g42

Further testing. Looking for bad altitude readings with -P 1. Well there are none!

Usually at -e 15 I get a lot of aircraft with final altitude reading being very bogus as they drop out of range. Like all these dark orange/maroon planes. Hundreds of bad altitudes in this image.

But with -P 1 and -e 15, not a single obvious bogus reading yet!

Edit
One more, you can really see the weakest messages are filtered out with -P 1 in the ADS-B signal graph. I’m not seeing any ill effects from this. Planes changing altitudes live on the map appear to be changing in 25 feet increments up/down every second or two like normal.

Just for reference, the same timeframe Airspy graphs.

1 Like

Please read what exactly it is, i’ve described it.
It works like a secondary -e for messages that don’t have CRC.
So 1 is actually the most restrictive version currently available.

Each message has preamble, so something that starts every message regardless of type.
It’s a certain pattern and the preamble filter looks for that pattern.
With high -e, the preamble filter is less restrictive, so … anything that remotely “looks” like that pattern is let through and decoding is started.
-P uses that same check again with a different factor when a potential message has already passed the main preamble filter and the message type is known.
-P 1 is the same for non-CRC messages as -e 1 would be for all messages.

If you’re really curious about how those options affect the numbers, record a local sample and test on that.

3 Likes

Thanks, I guess I did misinterpret the release notes. Makes sense. Testing -P 5 now. I already see a GND after 2 minutes with under 20 planes. So now I have a reference of -P 1 and higher. :grinning:

With RC14 and below using -C 85 and no -E set, I was getting -e to go up to around 36 or so… With RC15 I had the same… I checked VRS and noticed a few GND planes… So set -P 5, however now -e has shot up to nearly 50…

So I’ve now set -E 39 and will run with that for a bit.

So? :slight_smile:
With -P 5 or whatever setting you choose, increasing -e isn’t an issue usually as the less selective -e only applies to CRC messages.

Thanks for the detailed description.

I don’t think I need to use -P.

2.2-RC16

  • slight change to the json to support more data in graphs
  • minor bugfix: watchdog could falsely trigger on startup

This will hopefully be the last RC if there are no bugs.

5 Likes

some red, orange, yellow to be added and we have rainbows :wink:

2 Likes