Thoughts on optimizing gain

Actually it looks like you updated it to account for that back in March when you changed graphs1090, so the version on github should be ok. If it’s not working, then either there is some other problem or he doesn’t have graphs1090 installed.

Yeah i don’t even know which script exactly is being talked about and which URL he used to get it :slight_smile:

It’s this one:

sudo bash -c "$(wget -q -O - https://raw.githubusercontent.com/wiedehopf/adsb-wiki/master/caius_scatter_plot.sh)"

As long as he uses that command or downloads it freshly from github it should work. If it’s an old copy from last year or whenever it needs to be updated to the current version.

OK, I was using This script but I see it has been modded in the last hour or so, so I’ve just run it again with the same result:
"/dev/stdin" line 42: Can't plot with an empty cb range!

Yes, I did have (and do have) in /etc/default/graphs1090

# enable data collection for caius scatter graphs
enable_scatter=yes

I do have graphs1090 installed.

Some other problem maybe; looking through the logs, I see:

2021-05-16T04:25:23+01:00 daphnis collectd[10997]: rrdtool plugin: rrd_update_r (/run/collectd/localhost/interface-eth0/if_octets.rrd) failed: could not lock RRD
2021-05-16T10:08:25+01:00 daphnis collectd[678]: rrdtool plugin: rrd_update_r (/run/collectd/localhost/interface-eth0/if_octets.rrd) failed: could not lock RRD
2021-05-16T16:23:28+01:00 daphnis collectd[678]: rrdtool plugin: rrd_update_r (/run/collectd/localhost/table-localhost/gauge-cpu_temp.rrd) failed: could not lock RRD

which shows some problem, perhaps relevant.

The update I did just recently shouldn’t affect the problem you are seeing. It just added a second graph output that shows performance history all in one place:

It uses data which is pulled from graphs1090 each night, and stored in /var/lib/graphs1090/scatter, so have a look to see if you have any files in that directory. There should be one file per day - if you do have some, perhaps PM me a sample of one so I can see what is being recorded there.

I’m not sure what would be causing those errors with collectd though.

2 Likes

I have PM’d you a sample from /var/lib/graphs1090/scatter thanks.

Useful looking graphs which is why I have been trying to get the script to work. Your latest graph shows a very impressive range for aircraft; gainy aerial, serious height, lucky terrain ?

I replied to your PM.

I don’t have a particularly high gain antenna, though I am using an LNA. The antenna is mounted on the roof and I have a fairly clear horizon which helps a lot.

After discussion with @caius the solution was simply to do nothing and wait :grinning: Sure enough, this (the following) morning, the script works and graphs were produced.

The problem was that in my re-install of Raspbian Buster and fr24feed, I had not initially set the dump1090-fa location. Although I had corrected this issue prior to my post above, the sole file already produced was full of zeroes for range. This produced the (rather abstruse) gnuplot error. So, all that was required was to wait for the file next day, including range figures.

The collectd error was not relevant and remains to be resolved at some point.

Thanks to @caius for the script and pointing to the solution.

1 Like

Tried installing caius scatterplot on 978-uat only feeder but can’t find database files. which files are equivalent for 978?

sudo bash -c “$(wget -q -O - https://raw.githubusercontent.com/wiedehopf/adsb-wiki/master/c aius_scatter_plot.sh)”

Make sure you have the switch in the graphs1090 config file set to record data for scatter plots. It’s commented to show which one it is. The file is at /etc/default/graphs1090.

It produces output once per day shortly after midnight.

I’ve never tried it with uat only because it it isn’t used here, but I think it should work ok.

Well no the scripts won’t work with 978 because the files are named different, only one file for the number of messages … scatter.sh doesn’t do anything for the 978 data.

1 Like

Was able to get the 1090 scatter graph working on one unit, but a second is giving a 404 error. WIth filezilla can see the file is there just can’t get to it from the weblink. Also noticed that the rrd files for disk i/o are not generating, Have tried to update graphs1090 and scatter.sh. Any suggestions? 404 error http://(pi)/dump1090-fa/data/graph.png - Pastebin.com

The file that needs copying with Filezilla


The file visible from
http://192.168.86.37/dump1090-fa/data/graph.png

Quick question regarding gain settings for an Airspy mini

I know that the devices has a wider tolerance, but which measures should i take care of?
Or is it simply try& error?

Device is currently set to Gain 19 with these graphs, Traffic is around 80 aircraft visible and a few close aircraft passing by with less than 3 miles distance.

image

Yes.
But what you show is fine / normal.
Matter of taste.

1 Like

Hi @foxhunter
I find the Airspy mini is very tollerant of strong signals but you have to remember it is a very wideband receiver not specifically designed for 1090 Mhz. I found better results by using the FA dark blue 1090 bandpass filter in front and this allowed me to increase the gain without amplifying adjacent out-of-band signals.

dump1090-localhost-signal-2h

However I agree with @wiedehopf it’s really a matter of taste.

At my location having a “mean median Level” around -14/-15 appears to be optimal and don’t worry if the Peak Level is 0.0: the Airspy copes really well.

Thanks for the hint, but my Airspy is not connected directly to the antenna.
I am using the Uputronics LNA/Filter combination in front of it exactly for this reason

image

Therefore i could go higher to 21, but then the green band is cut on the top a bit.
I am totally fine with the current results, it’s just a bit system tweaking

In my many many tests and tweaks, I always come back to gain of 19 for Uputronics and 14 for RTL-SDR LNA/filters. Some stations might differ a click or two, but it probably won’t meaningfully change the statistics.

Every time I test/tweak, I watch GA aircraft fly directly over my station at 1000’ to 2500’ and commercial at 5500’+. Airspy just doesn’t care. Message rate might decrease for a few seconds below 2500’ feet above your roof, but that won’t impact anything meaningfully. By comparison, my UAT receiver gain is super low with Orange FA stick and the 978MHz Uputronics LNA/filter. Low flights right over me still hang at 0 messages p/sec for 10 to 15 seconds at a time during low direct flyovers.

I have one direction without trees and in that direction Airspy reports frequent -0.0 db signal strength out to around 50 nm or maybe more. It’s a pretty impressive receiver IMHO.

That was a previous disicussion as i see these as well. But the Airspy did not seem to be impressed by these.

I’ve changed the gain down by one step from 19 to 18.
No significant change except the number of Tracks with single messages dropped:

image

Is this now just by chance or is it related to the gain setting?