After above script finishes installation, it will start “sign-up” wizard and display following message: “Welcome to the FR24 Decoder/Feeder sign up wizard!”
Enter your
email address
sharing key
latitude
longitude
altitude
autoconfig (yes/no)$: y
logfile mode: 0
logfile path: /var/log/fr24feed
sudo systemctl restart fr24feed
In browser, go to address IP-of-PI:8754/settings.html
(1) Change Receiver :
From AVR (TCP) to Mode S Beast (TCP)
(2) Change Host/IP
From 127.0.0.1:30002 to 127.0.0.1:30005
Click “Save” button, then “Restart” button.
Both these buttons are at bottom-right corner of settings page.
See screenshot below
Ok, odroid now up on the pole. Seems to be working ok, only 1 issue, the graph ADSB max range not showing data, all others working. Thanks for above feeder entries
When piaware gets a location, so does dump1090-fa.
That location is required for range calculations.
You might need to reboot for the location to be updated.
It has multiple gain stages, 1 step is meant to be 3 dB approximately.
But if the gain stage settings when combined don’t allow that specific difference it’s just different.
So for 3 steps you’d expect approximately 9 dB difference, so a difference of 10 or 11 is close to 9 in a rough sense
It’s not a exact science anyway with these measurements.
Maybe for some reason i can detect weaker signals than with the higher gain setting.
That sounds counter-intuitive, but it’s possible.
The airspy has several gain controls that you can see in software like SDR# if you set the gain to “Free”. You can adjust LNA gain, Mixer gain and IF gain. It’s reasonable to assume that on something like the adsb decoder with just one gain setting, there is a combination of those settings changing to get the ~3dB steps. Perhaps a lower gain setting results in a lower noise floor due to the combination of these internal settings and that allows better reception of distant signals.
ADS-B should be limited by line of sight rather than signal strength anyway given the transmitter power used (can be up to 500W).
Perhaps a utility to produce a histogram of received RSSIs would be useful for setting gains - should be possible capturing the aircraft.json for a period of time.
Possibly a change to the script used to get the data and put it into collectd would work. You wouldn’t actually need to store a histogram, but could calculate a 5 number summary for each sample and store the values in collectd. That would allow you to plot the signal distribution in a way that would be good enough to show how well adjusted the gain is. Should work for dump1090 decoding as well.
124 signals.sort()
125
126 if length > 0 :
127 minimum = signals[0]
128 quart1 = signals[length/4]
129 median = signals[length/2]
130 quart3 = signals[3*length/4]
131 maximum = signals[-1]
The sort puts the list in ascending order, length is the number of items in the list.
Oh wait already see a problem with a 1 element list.
Nevermind it should do integer math and round down.
Which average is more representative is dependent on the data set. It’s situational, but you usually use the median when representing distribution, since with the mean a few outliers can skew it quite heavily one way or the other.
