changing -e 10 to 11, 12, 13 does not change the overall picture, just increases CPU. Changing sample rate from 12 to 20 does not change things, but uses more CPU and the mini will run hotter.
I am sure there are changes, but insignificant in the grand scheme of things here which is to test dump1090-fa-airspy. However what I have found is that the airspy mini can do just as well, and better than the Orange FA dongle. I get around the same number of aircraft and same distance as well. I see more messages.
And this annotated set from graphs1090 highlights some areas of difference
Observations: the Mini using airspy_adsb provides similar aircraft numbers and more messages than the reference system. So why donāt I use the Airspy Mini instead? Simple: airspy_adsb currently does not support Mode A/C.
What I cannot tell is if the messages are all valid in either system, however this may help, as it shows a breakdown of messsage types, and potential errors or bad decodes.
If you use airspy_adsbās -r Reduce the IF bandwidth to 4 MHz option, you can get the same effect in dump1090 by setting --freq 1089200000
Questionā¦ If you use --airspy-enable-rf-bias to power an LNA, do you think dump1090 should turn the bias back off again when it exits? Right now, I leave it on.
@gtj0 - Well done sir! Iāve been out of it for awhile, just happened to have a few minutes to browse and found this thread. Itās come a long ways!
Hope to have time again to test within the next couple months and I still have a couple identical rigs up to help with the dirty-work. Just need time. Heck, Iām still on v3.7 and a dev version at that. Thatās how far behind things have gotten on my end.
Looking forward to catching up with everyone again soon. Keep up the good work.
And you have a lower number of aircraft without position on the Airspy.
I assume your message rate > -3dBFS seem to be too high with 11% on the FA dongle
@gtj0 I agree, should always leave it in a known state, leaving it as you found it is correct.
You just know that there would be a ābutā, thoughā¦ But I use the LNA powered via USB socket not bias-t, as I have a passive splitter and need the LNA on regardless of the state of any of the connected receivers. If my only option was to power via dump1090 enabling bias-t, and if I blocked DC at the other splitter ports via a capacitor, I would want dump1090 to leave it on. I suspect this is a corner case !!! If someone has this requirement, I would expect they have the skills to make separate LNA power arrangements. So on reflection (maybe a pun in there), yes always turn off when dump1090 exits: safer option and let the ācorner-caseā experimenters do their own thing, as I have done.
Yes a little high, I have been āpushing the boundariesā of the dongle chasing the last drop of performance. I found the same trend, though when the -3dbfs was lower, compared with airspy-adsb.
I will lower the gain a notch and see what happens. trying to avoid changing too many things at once: makes comparision difficult
KIV I am in a very quiet location, RF-wise at 1090Mhz and that may well make a difference to performance.
Agreed. Iāll try to add it this weekend but Iām pulling my antenna platform down for a quick refurb and that usually ends in more work than expected and certainly more aggravation.
Yeah I got the same result but I never used it with airspy_adsb before. Thanks for testing.
type or past-- A stop job for unit dump1090-fa.service has finished.
--
-- The job identifier is 4401 and the job result is done.
lines 524-546/546 (END)
Aug 03 08:53:42 raspberrypi dump1090-fa[14470]: 0 non-ES altitude messages from ES-equipped aircraft ignore
Aug 03 08:53:42 raspberrypi dump1090-fa[14470]: 2133 unique aircraft tracks
Aug 03 08:53:42 raspberrypi dump1090-fa[14470]: 42 aircraft tracks where only one message was seen
Aug 03 08:53:42 raspberrypi dump1090-fa[14470]: 99 aircraft tracks which were not marked reliable
Aug 03 08:53:42 raspberrypi dump1090-fa[14470]: CPU load: 16.8%
Aug 03 08:53:42 raspberrypi dump1090-fa[14470]: 7606664 ms for demodulation
Aug 03 08:53:42 raspberrypi dump1090-fa[14470]: 10771 ms for reading from USB
Aug 03 08:53:42 raspberrypi dump1090-fa[14470]: 2127750 ms for network input and background tasks
Aug 03 08:53:42 raspberrypi dump1090-fa[14470]: Tue Aug 3 08:53:42 2021 BST Normal exit.
Aug 03 08:53:42 raspberrypi systemd[1]: dump1090-fa.service: Succeeded.
-- Subject: Unit succeeded
-- Defined-By: systemd
-- Support: https://www.debian.org/support
e code here
You do still need libusb but it was probably installed automatically anyway.
Since you already installed airspyone_host, run this command from a Linux terminal and paste the results back here.
!!! If someone has this requirement, I would expect they have the skills to make separate LNA power arrangements. So on reflection (maybe a pun in there), yes always turn off when dump1090 exits: safer option and let the ācorner-caseā experimenters do their own thing, as I have done.
Good luck