Do you have any numbers of the difference i performance between the Pi and the laptop?
I’m seeing improvements on my Pi4 when I increase -e, but if I’m running with 24MHz sample rate I max out -e at approx 9.8, then I start losing samples.
I do get better results with a sample rate of 12 and -e at 12.0 than with a sample rate of 20 or 24 with a -e around 9.8 - 10.5.
There is an old laptop in my pile of computer junk that could be used but it would be nice to get some numbers before hunting down a hdd for it and starting the process of replacing the PI4…
I guess it all depends of the actual traffic in your area. If with your Pi you are getting all the airplanes in your area, further increase of -e will not add anything. My Pi3B was loosing samples ever at lower -e than your Pi4. Looks like there is one guy close to you (joakim77andersson) that does better than your stats, so there is room of improvement. But he is at 83 miles away and that’s where you get most of your flights, plus he might be on higher ground… Don’t know specifics.
Increasing -e adds something only in really busy airspace, with planes both close and far simultaneously.
If you see better results with 12MHz SR and higher -e, a laptop like mine can let you run 20 or 24 SR at the same -e. Will that add any thing? It depends of your location.
For me that was a nice hobby, that’s why I did it. Plus I had that free laptop with a decent i7 CPU (2nd gen indeed, but still OK).
I have the stats below, compared with my neighbors, up to 20 miles around me. This is a flat coastal area, so antenna height is important.
The top position is my Airspy, with Intel CPU laptop. Second performer, from all those, is the FlightFeeder that I got from FA, using a splitter from the same antenna cable. That’s my benchmark radar station.
My next two radars are UAT, that’s why the low numbers.
I’m not getting all the traffic in my area, but that’s due to terrain (trees mostly, but also some lower hills) that blocks most of the small aircraft traffic south of my location. The coverage is spotty but seem to be a bit better with an increased -e value (and a bump in gain).
My Pi is “slightly” overclocked, that and the lower amount of traffic is probably why I can get a bit higher on the -e setting.
joakim77andersson lives pretty close to ARN, the largest airport in Sweden and he sees more of the lower altitude traffic than I do (due to distance and terrain), that’s probably why his numbers are a bit better than mine. I may be located a bit higher than him, but not high enough to have a similar line-of-sight as he does. My antenna is placed on the roof of my house and I can perhaps increase it’s height by 3-6 feet without too much problem. Don’t want to erect a mast since the cost is too high for a hobby.
I get a fairly decent amount of traffic from routes between continental Europe and Asia and the Middle East to north America (120-200+ NM). How much varies with weather and season (snowy trees are quite effective in blocking the signals) and a day with some tropo can double or even tripple the traffic at 200+ NM.
Since I live where I live my main goal is trying to squeeze out every last bit of usable information from the signals I can get. It’s quite tricky since the conditions vary from day to day and the air traffic is what it is during lockdowns and travel restrictions. I currently have about 1/3 of the traffic pre covid in my area. Before, the small aircraft traffic didn’t make a noticeable impact on my daily stats, now they do since that traffic has been virtually unchanged during this period.
I’m working on setting up my old laptop. The CPU is fairly crappy (dual core, no threads) but I’ll try it out and see how it performs. Since airspy_adsb isn’t threaded it should still perform better than the Pi4. The bottle neck will probably be the USB bus. But we’ll see.
Your stats looks really good, the difference between the two systems clearly shows that the airspy is better, but perhaps not worth the extra money unless you’re an enthusiast with nerd ambitions 
Thanks for your answers 
1 Like
Or if you have disposable income because you didn’t go anywhere “fun” in 2020. 
2 Likes
It seems I may have made a mistake here. Time to redo the budget for 2021 
1 Like
Spent it all on fine alcohol, heh?
I have two teenage boys (both 190cm+). Food is a morale-booster they say. To me it looks more like a money-sucking black hole. 
1 Like
Even gradually increasing the -e number, I’m still well down on before at 12 MHz. My -e is currently at 11.1, it’ll go up to 11.5 overnight. Still no lost samples.
A few days of this and I suspect I’ll be going back to 20 MHz and -e 9.9 if it doesn’t improve.
1 Like
How has your CPU utilisation been changing as you stepped up the -e value?
I was able to run the 12Mhz decoder with e set to 12, and I didn’t see any gain. There didn’t seem to be any considerable change either way compared to the previous setting of 20MHz with e at 10. I’ve put it back to 20Mhz and have the pi running at 2GHz. This allowed me to increase e to 10.75 without losing samples.
I have also adjusted the gain down - it was set to 17 and is now set to 14. I’ll let it run for a few days to see how it goes, but comparing to the Martello tower receiver which is normally very similar to mine, there does appear to be a bit of an improvement.
Your figures are definitely down - I have a similar number of aircraft seen as you and more positions over the last couple of days. Normally you are comfortably above me in both stats.
I’ve got bored with slowly increasing this and have set myself back to 20 MHz and 9.9.
Hello, I’m new to this… Have my antenna at hight now outside… WOW what a difference. Running Airspy 2 into Uputronics amp. Been reading learning a lot from these forums.
My current options - -v -f 1 -b -e 9.4 -w4 -t300 -m 20
Gain set to 21
Pi 4 running the show… With a decent fan. Worth overclocking the pi to push -e ? currently around 75% CPU. Temp a steady 32c
Any other suggestions or thoughts would be welcome ?
I would suggest trying running on a bit lower gain. Pergaps 18 or 19. Depending on traffic pattern at your site there can be some positive effect with a litter lower gain. Since you also lower the noise of the system by lowering the gain you can increase the message rate on more distant aircrafts if you set a lower gain.
My site get a little higher message rate from more distant aircrafts when running a gain of 18-19 compared to 21. If I drop too low (16 or lower) I loose the more distant aircrafts but get a higher message rate from aircrafts at a medium distance from the reciever.
I’ve overclocked my RPi4 and can run -e at just a bit over 10 with -m20. I think I see a slight advantage but it’s not that much of a difference.
The big enhancements come from antenna height and placement (as you have noticed). The other enhancements you can do make much less impact on reception and message rate.
thanks… Will drop gain and see what happens.
does the Airspy cap out at x number of messages ?
Thank you
1 Like
Run this and see how many lost samples you’re getting
journalctl -eu airspy_adsb --no-pager
I dont think I’m dropping any
May 02 10:43:16 raspberrypi airspy_adsb[2793]: Caught signal 15
May 02 10:43:16 raspberrypi systemd[1]: Stopping Airspy ADS-B receiver…
May 02 10:43:16 raspberrypi airspy_adsb[2793]: Decoding stopped
May 02 10:43:16 raspberrypi airspy_adsb[2793]: Push client disconnected from localhost:30004 (beast)
May 02 10:43:16 raspberrypi airspy_adsb[2793]: Client disconnected 192.168.50.4:58726 (beast)
May 02 10:43:16 raspberrypi systemd[1]: airspy_adsb.service: Succeeded.
May 02 10:43:16 raspberrypi systemd[1]: Stopped Airspy ADS-B receiver.
May 02 10:43:16 raspberrypi systemd[1]: Started Airspy ADS-B receiver.
May 02 10:43:16 raspberrypi airspy_adsb[2483]: airspy_adsb v1.85
May 02 10:43:16 raspberrypi airspy_adsb[2483]: Listening for beast clients on port 47787
May 02 10:43:16 raspberrypi airspy_adsb[2483]: Acquired Airspy device with serial B58069DC393F4513
May 02 10:43:16 raspberrypi airspy_adsb[2483]: Decoding started at 20 MSPS
May 02 10:43:20 raspberrypi airspy_adsb[2483]: Push client connected to localhost:30004 (beast)
May 02 10:43:22 raspberrypi airspy_adsb[2483]: Client connected from 192.168.50.4:41232 (beast)
May 02 10:48:02 raspberrypi systemd[1]: Stopping Airspy ADS-B receiver…
May 02 10:48:02 raspberrypi airspy_adsb[2483]: Caught signal 15
May 02 10:48:02 raspberrypi airspy_adsb[2483]: Decoding stopped
May 02 10:48:02 raspberrypi airspy_adsb[2483]: Push client disconnected from localhost:30004 (beast)
May 02 10:48:02 raspberrypi airspy_adsb[2483]: Client disconnected 192.168.50.4:41232 (beast)
May 02 10:48:02 raspberrypi systemd[1]: airspy_adsb.service: Succeeded.
May 02 10:48:02 raspberrypi systemd[1]: Stopped Airspy ADS-B receiver.
May 02 10:48:02 raspberrypi systemd[1]: Started Airspy ADS-B receiver.
May 02 10:48:02 raspberrypi airspy_adsb[4232]: airspy_adsb v1.85
May 02 10:48:02 raspberrypi airspy_adsb[4232]: Listening for beast clients on port 47787
May 02 10:48:02 raspberrypi airspy_adsb[4232]: Acquired Airspy device with serial B58069DC393F4513
May 02 10:48:02 raspberrypi airspy_adsb[4232]: Decoding started at 20 MSPS
May 02 10:48:06 raspberrypi airspy_adsb[4232]: Push client connected to localhost:30004 (beast)
May 02 10:48:07 raspberrypi airspy_adsb[4232]: Client connected from 192.168.50.4:41416 (beast)
Well there is a limit on the 1090 Mhz band due to overlapping messages.
The airspy decoder is pretty good though and can sometimes decode partially overlapping messages especially if the messages are a bit out of phase i think … not exactly sure.
I think the record for message number is around 2900 per second by either caius or keithma.
But that was before covid and with very very very good antenna sites.
I don’t think pushing -e is really worth it at those levels but that’s very subjective as doing objective tests is hard.
I think pushing e further is a theoretical extra 1 or 2 percent of messages.
That’s tested by recording a sample and then decoding it non real time with very high -e setting then comparing.
I like to run my gain rather low … but it’s a matter of taste:
Have you done a theoretical terrain outline?
GitHub - wiedehopf/tar1090: Provides an improved webinterface for use with ADS-B decoders readsb / dump1090-fa
Then after the outline and having tar1090 run a couple of ours check ptracks:
GitHub - wiedehopf/tar1090: Provides an improved webinterface for use with ADS-B decoders readsb / dump1090-fa
Oh and in regards to 1090 MHz band limitation there is still lots of ModeAC messages on that band as well.
I have no clue why any radar nowadays would still interrogate that but that’s apparently how it is.
I suppose you can’t interrogate ModeAC without also interrogating planes that can also transmit ModeS.
But airspy_adsb only decodes ModeS, that very good though.
ModeAC messages carry very limited info so not really useful.
Hi, yes running TAR1090 and have run a theoretical range outline. I did see a post on its on my to do list to add that range outline to TAR1090…. Think that might be beyond my capablity… only started with the Pi as something to do in Lockdown and then got into this flight tracking game and am hooked….
Moving the Antenna outside the loft a hight has made a massive difference. For my local airport I was way down the rankings… on the fliightaware stats page… Should move up considerably now…
Thanks for the help…
It’s really just a matter of following the instructions.
If one of them isn’t clear, please elaborate so the instructions can be improved.
You can use /?pTracks without having the outline present, just not really something compare it to.
You have the graphs installed?
GitHub - wiedehopf/graphs1090: Graphs for readsb / dump1090-fa / dump1090 (based on dump1090-tools by mutability)
Then you could show what your signal graph looks like.