Can someone explain to me how a local process on my Raspberry Pi displays MLAT derived targets on tar1090? Is there a white paper that explains how this is done? I thought all the MLAT data went to a server and the positions and timestamps my feeder sends are combined with others to find a position in 3-space based on arrival time. I got that, but it looks like my R-Pi is getting MLAT data and doing the calcs to find positions autonomously. Is that really happening or is the MLAT server sending me back some positions? By the way, got my antenna up on my 45' tower now. <FlightAware 1090 Antenna> <1090 LNA/Filter> <50' LMR400> <Bias-T> <RTL-SDR> --->[ R Pi 3B+]<--<AdaFruit Ultimate GPS>. ADSBx reports that my station has between 46-50 MLAT sync peers. I'm getting positions out to 220 NM. Lots of aircraft around the Pacific NW. I'm kind of hemmed in by the Olympic Mountains on one side and the Cascades on the other, but it's open going North and South. Is there a way to improve the "Health" stat? Mine moves around between 93-95%. There seems to be a few peers with poor time sync that are poisoning the stats. Can I manually deselect peers? Just in time. Looks like it's going to rain for the next week. - Bill
The MLAT server returns positions/results. Note that you'll likely need to reduce gain to 25 with your setup. Nothing to do with the sync stats though. Removing peers with bad sync is being done, but it's tedious. Don't worry about the health number, unless it's your station that's the problem.
When I go to look at stations with bad sync to my station they are out of sync with a lot of stations. It is interesting that I see those stations are out of sync with about half of their peers by a lot, but very closely in sync with the other half of their peers. It's like there is another time source that these other stations have in common with the peer I see as out of sync with me.
wiedehopf: My antenna has an advertised gain of 5.5 dBi. The LNA has an advertised gain of 27 dB. The LMR400 has a calculated loss of 2.4 dB. The bias-t has an advertised loss of <1.2 dB. There might be another 1 dB loss in connectors in this arrangement. In /etc/default/dump1090-fa has the receiver gain set to -10. I see in the ADS-B Message Rate graph it says that 73.6% of the messages are greater than 3 dBFS. That says to me that for a lot of the messages coming in, the wave forms are clipped. The peak signal level is -1.1 (dBFS I assume) and the median is -2.3. If median signal level is the key to optimum detection, what should the target median signal level be? How much dynamic range is there in the algorithms for reliable detection? I'll try plugging in a gain of -25 in /etc/default/dump1090-fa to see how it goes, but if I have targets I can dial it right in. Thanks, Bill
Unless your station is red constantly, it's fine. It's really just some averaging and voodoo math to get a 'heath' number.
In /etc/default/dump1090-fa has the receiver gain set to -10. For maximum sensitivity, ergo range, the gain must set to a level where noise is lowest, mean level is not reaching maximum and peak level is above -3dBFS. possible gain values 0.0 0.9 1.4 2.7 3.7 7.7 8.7 12.5 14.4 15.7 16.6 19.7 20.7 22.9 25.4 28.0 29.7 32.8 33.8 36.4 37.2 38.6 40.2 42.1 43.4 43.9 44.5 48.0 49.6 -10 basically sets it to max gain.
https://www.mictronics.de/2017/03/dump1090-gain-vs-performance/ Eventually once you get comfortable I recommend flashing a non FlightAware image.
The aim is usually 1% to 10% messages above -3dBFS. Are you even using the piaware sd-card? If you are the gain is changed via piaware-config. But i assume you are just using a standard Raspbian? Then your approach to changing gain is correct. If you used my bundle install and want to adjust gain manually, i'd recommend removing the automatic gain adjustment. (https://github.com/wiedehopf/adsb-s...ation-for-dump1090-fa#removal--deinstallation) Or in any case use the gain values used by the dongle. (0.0 0.9 1.4 2.7 3.7 7.7 8.7 12.5 14.4 15.7 16.6 19.7 20.7 22.9 25.4 28.0 29.7 32.8 33.8 36.4 37.2 38.6 40.2 42.1 43.4 43.9 44.5 48.0 49.6 -10) If you manually set another value the automatic gain script will become confused and set a gain of 49.6 i think. So best to just remove it if you are tweaking manually anyway.
I'm well aware, but the automatic gain script i wrote is pretty limited and doesn't choose the closest number. Anyway if you are changing the gain manually it's best to just remove that script.
I'm using an image of the latest PiAware sd-card with the ADSBx feeder grafted on. I tried changing the gain in /etc/default/dump1090-fa and as predicted nothing really changed in my graphs and when I looked at the file after a restart, the gain was set to -10 again. I then changed the gain in /boot/piaware-config.txt to 25.4 and restarted. I went back and looked at the dump1090-fa file and it had it's gain set to -10. I expected to see that changed by the restart to the value in piaware-config.txt to -25.4. Doing a reboot of the RaspPi did set the value in dump1090-fa to -25.4. doing a piaware-config -restart isn't enough. Is there a way to look at the settings in the RTL-SDR directly?
You do: sudo piaware-config rtlsdr-gain 25 sudo systemctl restart dump1090-fa That should be all that is required. And don't use negative values, the only negative value for gain is -10, and that's for AGC.
wiedehopf: Thank you. Sounds like the -10 number is a number with the sign bit in whatever size register saves the gain value and that turns on AGC. And, the gain settings are actually positive numbers. Setting the gain to 25.4 (over the default -10 setting) has increased the number of messages and the number of positions by about 100%. The number of messages with levels greater than 3 dBFS is now below 10% and the CPU consumed by the decoding algorithm is down by about 10%. Seems like there may have been a lot of IMD the decoder was having to deal with. - Bill
