UAT signal graph, made from individual plane RSSI, not the same data as the signal level graph for dump1090-fa.
Same approach as for the signal level graph for airspy receivers.
Also for UAT i’ve made it so even if an aircrafts position is 60 seconds old it’s still counted (30 seconds for ADS-B).
Just to make sure everything is caught by the graphs, otherwise they stay so empty.
Oh if you only see one plane, the mean, peak and weakest signal level graphs will all sit on top of each other obviously.
Only file missing after an upgrade is this: File /var/lib/collectd/rrd/localhost/dump1090-localhost/dump1090_dbfs-min_signal.rrd not found! Associated graph will be empty!.
This is the better thread for the signal level graph discussion.
I’ve made it so the Median is used instead of the average signals if defined.
So when the Median starts to be collected, it’s used, the past data is the average signal.
It’s kinda ugly but this uses the existing data for the past.
Edit: Pushed out for testing if someone wants to try.
Ah, I see why the label now - it actually plots the mean and the median - mean before the change and median after. I see a considerable step between them:
Kind of interesting that the noise line is above the median! I recall obj saying that it’s a kind of pseudo-noise number rather than anything particularly meaningful though.
Yeah that wasn’t the reason for the name. (not entirely at least)
The mean of the median is printed as a number after the text
That looks really strange for you.
I suppose it makes sense as you get many more messages from aircraft that are transmitting stronger.
Especially with your noise floor.
The peak is also still taken from the dump1090 metric instead of using the strongest aircrafts signal level at the time.
I suppose i’ll have to change that.
Gotta say the dump1090 statistics don’t seem too useful, i might just override them.
Also plotting the noise seems pointless if you get many messages far below the noise “floor”?
Yeah that makes it very obvious that the gain is set too high, and there is no obvious difference in the peak signal level. This could be quite a useful change.
I get a lot of aircraft very close, as one of Heathrow’s approach stacks is right over my house, and I also get London City departures literally right over the top of me, albeit they are less frequent.
The noise floor isn’t really a noise floor from what obj said. It’s just a measure of how much non-ads-b signals there are. I’m not sure it’s particularly useful though, as it includes stuff like bad decodes and undecodable collisions etc. which you can’t really control. It just seems to be at a certain level depending on your gain setting and how busy it is.
That’s probably a good idea, for consistency if nothing else. At least you know exactly what it represents then.
Except that for the airspy, that’s a perfectly good setting.
Due to ModeS message encoding, overdriving the ADC doesn’t hinder demodulation of messages.
You just don’t want to overdrive the analog stages before the ADC.
Airspy has quite an advantage with its dynamic range compared to an RTL dongle. I wonder if having the gain like that affects the ability to decode overlapping signals. What is interesting is that increasing the gain also increases the minimum signal level observed. That implies that the limiting factor for your reception is not signal strength.
So i can’t get signals weaker than the LNA noise level.
There is also a certain 1090 MHz noise level received by the antenna.
Not sure which limit i’m scraping at.
I’m limited by terrain anyway, those weakest signals are probably already through a tree line that’s far away or some refraction or something like that.
Hmm i wonder if i change the peak signal to not the strongest message received in the last minute but to the signal strength of the currently strongest aircraft, some people with rtl-sdr dongles might increase gain.
Oh well, not my problem i guess.
I’ll make the changes and see how it looks on your end.
OK. I’ve set my gain to maximum to see what it looks like. My system is in a bit of an odd state at the moment because I made some changes for comparison purposes. I’m using a different antenna to normal with about 20m of WF100 between the LNA and the filter. I’m trying to get a baseline for a kind of worst case, since when I move it outside it will have a different pre-amp and likely a bit shorter coax run.
No need, was just trying to guess on the limit of what you can receive.
There might also be a noise floor in the analog front end which is limiting without LNA.
Anyway you can do an update again, just pushed out the peak signal changes.
Oh i also found the limit on reducing gain.
When i change from 12 down to 11, the weakest signal doesn’t get any weaker, so i’m likely losing signals at that point.
I would probably go with the lower gain setting shown in the graph, but that’s more personal preference than actual evidence there is any benefit to doing that.
