Just merged this addition into the master branch.
To add this feature update your local repository and run install.sh. You can skip the installation of everything except the portal.
Upgrade instructions from the first post can be used to update your repository:
cd ~/adsb-feeder
git fetch --all
git reset --hard origin/master
./install.sh
CHANGELOG For January 14th, 2016
•When the portal is installed the argument --measure-noise is added to the dump1090-mutability init script.
•Added “noise” level line and information to the signal level performance graph.
Thanks to @mighty3507 for the example he posted which was used as a reference when modifying the existing signal graph.
Was playing with that trying to adjust the size down somewhat so the resolution of the graph when it pertains to the other information displayed in it was not so “washed out”. Mine did not go down past the -24 mark hence the reason I set it there but things hardware wise are not all that tuned on my end either. Anyways this is easy enough to fix.
Open ~/adsb-feeder/build/portal/graphs/make-collectd.graphs.sh.
Look for the function “signal_graph()”.
Within this function look for the line " --lower-limit -25 "
Change this to a lower number such as -35.
Regenerate your graphs or wait until cron takes care of that for you.
Would love to hear what setting works for others so I can change the make-collectd.graphs.sh script in the repository to one that better suits everyone in general.
My max is currently -33 and min is -40. So I set -40 as the lower limit for the graph. Need to watch this for a day I guess. Thanks for this nice addition!
@sjacket99: jprochazka wanted to see what people are showing with the new “noise” graph added. Like this (amplified with gain setting in dump1090-mutability of 45:
this what I see: http://s12.postimg.org/cnqyed3j1/dump1090_localhost_signal_1h.png
the setup is - freshly built Jessie image with jprochazka scripts attached to the 4 segment coco I was trying to build for UAT with no amplification. Gain set to -10 and --enable-agc added to dump1090 additional parameters. The later section is with --enable-agc removed.
Looking over the graphs everyone posted it looks as though -40 is the way to go. I am still in the process of dialing back my gain but had my graphs lower limit set to -35 and things still dropped off the scale especially in the middle of the night when there was very little traffic in the air. The change has been made to the file make-collectd-graphs.sh to expand the graph to -40 this morning.
Thanks for the feedback everyone.
When you turned AGC off it resulted in a lot less noise so the line fell off the chart. Up until this morning the graph only reading down to -24 was not enough. I changed the file in the file make-collectd-graphs.sh in the repository to show down to -40 which seems to bring things into view better on less noisy set ups. You can use the instructions in this post, post186826.html#p186826, to manually change the one line that needs changed or update your scripts and run install.sh again skipping everything except the portal installation as outlined in the first post on this thread.
how do I undo all the changes install.sh made …meaning uninstalling dump1090 mut piware …plane finder exchange and graphs so that I can run the script again on a clean slate , after doing a upgrade last night reflecting the power graph and editing dump to change -10 to - 40 as suggested the conection between dump and milat broke graphs broke and only seeing the occasional plane as long is it isent more than 50 miles ways , updating wont fix it the option to update the feeders dosent even appear am using ubunu 14.04
The -40 mentioned earlier had to go with a change to the file “~/adsb-feeder/build/portal/graph/make-collectd-grpahs.sh” not dump1090-mutability.
If by changing -10 to -40 you are talking about the gain setting in “/etc/default/dump1090-mutability” try making 40 a positive number instead of a negative number. Keep in mind there is no magic number that works for everyone. This number may need tweaked a bit depending on your setup. On my setup a gain of 28 worked better for me. As far as MLAT goes check that the setting “BEAST_INPUT_PORT” is set to 30104 in the file “/etc/default/dump1090-mutability”. After making anychanges either restart dump1090-mutability or reboot the system to make sure everything comes up fresh.
Uninstall dump1090-mutability:
sudo apt-get remove dump1090-mutability
Uninstall PiAware:
sudo apt-get remove piaware
Uninstall Planefinder ADS-B Client:
sudo apt-get remove pfclient
For ADS-B Exchange remove the line “path/to/adsbexchange-maint.sh &” from /etc/rc.local and “beast,connect,feed.adsbexchange.com:30005” from your piaware-settings -mlatResultsFormat setting.
The web portal, including performance graphs, simply delete the contents of /var/www/html.
A note on the signal level floor:
Still trying to find the bottom… I moved my --lower-limit for signal_graph() in make-collectd-grpahs.sh on my live feeder from -40 down to -45. I have been averaging a max of about -29 and min of about -41 since beginning to dial things in a bit gain wise so my noise line has been falling off the graph just a hair especially during slow times. Curious if this is happening for others and thoughts on lowing it just a tad bit more in the repository. Here is what things look like with the change: 192.168.254.2/graphs/dump1090-lo … al-24h.png
That is with a preamp, and gain set to 48 in dump1090.
One statistic that I’ve found quite useful for assessing performance is the number of messages received per visible aircraft. It’s just calculated using the existing figures gathered by collectd so is simple to do. A graph looks like this:
During the day when it’s busy, then number of messages per aircraft drops quite considerably. I think this is due to garbling and the limited dynamic range of the dongles - a higher quality receiver can sort out the mess of cross talking transmitters better. The minimum messages per aircraft shows how well the receiver is doing that. If the gain is set too high or too low, then this will drop. For my setup, the minimum seems to be typically about 500 msgs/min/ac. When there are fewer aircraft, the number is a lot higher.
It would be better to plot the data as an X-Y plot rather than a time series, but I haven’t found a way to do that from rrd files.
