Huh?
What did you delete? 
I thought that was what you did.
before your new procedure backup stop collectd then delete xml and rr files then restore. Without knowing which files should be deleted it becomes trial and error, with possible deleting of extra files to cause wanted graphs to disappear.
My one station has lost the Messages > -3dBFS on Graphs1090. Any idea how to get it back? It was working perfectly, then I installed an Airspy and had that running to compare if there is anything to be gained, but reverted back to the standard RTL when I saw it made almost no difference in the amount of planes received. I ran wiedehopf’s scripts for the installation and uninstallation. After reverting back to the RTL-SDR, the -3dBFS data never came back. Uninstalling graphs1090 and reinstalling it made no difference. any ideas?
3dBFS on Graphs1090 don’t work with airspy
1 Like
I noticed. I reverted back to the RTL-SDR but it still stayed away. trying to get it back.
On trl again then just reinstall graphs
sudo rm /var/lib/collectd/rrd/localhost/dump1090-localhost/dump1090_cpu-airspy.rrd
1 Like
thanks, Wiedehopf. Now I can sleep again 
dump1090_cpu-airspy.rrd comes back after a reboot. It also has to be removed from /run/collectd/localhost/dump1090-localhost/.
Then everything seems to work as it should after a reboot. Hopefully I didn’t screw things up now.
Oh yeah i totally forgot i now have everything in /run which the collectd service is running.
Means you only get one continuous write every 24h instead of lots of small writes all the time 
Makes it easier for people eliminating sd-card writes.
2 Likes
Now my system has settled down (and before I start tinkering with it 
) I was wondering what the experts out there might make of my graphs?
It would be good if I could solicit your analysis/opinions and then maybe in another week or so after I change the AIrNav FlightStick for the FlightAware ProStick Plus I could be so cheeky as to post my graphs again to compare them both?
My RadarBox station ranking has been pretty steady at 8 in the UK/120 Globally so that should be pretty easy for me to keep my eye on but although I can loosely interpret the graphs it would be better to ask the opinions of those who know!
The Gain on dump1090-fa is set to -10 if that makes any difference and I did try it on what would be effectively the next setting down 49.6 which reduced the messages >3dB from a possibly too high(?) 13% to a possibly too low(?) 0.3%
Thanks & kind regards,
-=Glyn=-
If I look at the noise floor with your gain setting at -10 (AGC on) and amount of signals >-3dBFS, I think you should reduce your gain. Most probably the manual gain of 49.6 was still too high, causing overload. The easiest solution is to install Wiedehopf’s automatic gain script, and let it do the gain tweaking for you. Do you have a pre-amp in the chain?
So … with those numbers the automatic gain script would probably switch back and forth every night.
Manual gain change then until the right gain is found. I would suggest running a scan to see if the system gets overloaded by some out-of-band signals, like a cell-phone tower close by.
Thanks for the reply hellhound604 & wiedehopf,
Well I did reduce the gain to 49.6 and as I said the percentage of messages >3dB went down to 0.3% which as I understand it is much lower than the ‘target’ of 5% and apart from the expected signal level drop didn’t seem to affect things much at all?
But at the same time my ranking dropped although this could have been due to other factors such as weather, number/location of aircraft etc.
And if the signal level is still below zero when the gain is set to -10 would that indicate it is not overloaded? 
There is no external pre-amp in the chain but I believe the FlightStick may have one fitted along with a filter not unlike the ProStick Plus? The antenna is a Watson Radar Extender connected by @7m Westflex 103 and has proved itself over many years although that’s not to say its performance hasn’t deteriorated over time.
How would I run a scan to check for out-of-band signals although judging by my poor mobile reception this might not be an issue!
Thanks & kind regards,
-=Glyn=-
That’s really not so much a target as it is an orientation.
It’s likely not a major difference in your case.
Well you might lose some very close by positions for low flying aircraft with the higher gain (which happens very rarely and thus doesn’t affect stats at all).
Switching to the FA Pro+ is actually very likely gonna tell you if you have strong interference / need a filter for it.
The AirNav flight stick has the filter before the LNA, the FA Pro+ has the LNA before the filter and is thus more sensitive to both signals and overload of the LNA.
Thanks for the replies guys…I have read most threads covering dump1091, Airspy/FlightStick/FlightAware, filters, etc. that I could find with up to 800+ posts per thread and as is usual for me it seems to raise more questions than answers!
I haven’t yet had the courage to swap the FS for the Pro+ but will finally get round to it and will update this thread on my findings however unscientific they may be!
Reading between the lines I’m guessing the FS has the filter before and the FA Pro+ after?
Thanks & kind regards,
-=Glyn=-
This is an amazing tool, thank you @wiedehopf for taking your time to get it completed.
I just have a question or two, how can I interpret this graph? Is my gain too high (currently 40 and when running <awk “$(cat /run/dump1090*/stats.json| grep total | sed ‘s/.accepted":[([0-9]).strong_signals":([0-9]).*/BEGIN {printf “\nPercentage of strong messages: %.3f \n” , \2 * 100 / \1}/’)”> it returns 25).
