Strange noise level behaviour

Hi all,

Finally had the time to now decently mount my receiver antenna on a pole outside.
Before I had my antenna and lna/dongle/RPi next to the window 50cm away from my pc.
Whenever I turned on my PC, I saw the noise level rise in the graphs and when shut down, levels went back “normal”
However, now that I moved it outside and used some good coax cable (Ecoflex 15) my noise levels are moving along the plane activity overhead.
During the night the noise level drops below the weakest signal, steadily rises in the morning to the same level as the “mean median level” of received ADS-B signal, to drop back down again in the evenings.
Also during the day, the noise level fluctuates in a somewhat similar fashion with the amount of received ADS-B messages.

Antenna is mounted on a metal pole connected to a zinc downspout that goes into the ground. So I expect that the grounding should be sufficient.

As it is not a static noise level, I dont really know in what direction to start troubleshooting this behaviour.
Does anyone have any idea what to try first?

Find attached 2 screenshots so you guys can see the different signal and noise levels

dump1090-localhost-signal-7d
dump1090-localhost-signal-48h

I’m sure the more experienced members here will know for sure, but my graphs look very similar to yours, and having run a few scans ( see Do I Need A Filter? ) I had fairly severe 3g and 4g interference from a mast a few hundred yards away.

A flightaware filter has improved things but I still get fairly significant drops in noise overnight and rises again as everyone rises from their slumber :slightly_smiling_face:

Search for the “do I need a filter” topic and run the diagnostics mentioned in it. That will show whether or not you’re picking up interference from, for example, a cell phone tower that was attenuated by your house when the antenna was inside.

Noise includes partial messages, at night you get fewer of those.
In other words, noise is influenced by number of planes.

The weakest signal is a much better measure of effective noise as it shows you how weak of a signal you can receive.

So, I ran the test as described in the “Do I need a filter” topic. However I do get an error message when it comes down to the python part. But I did manage to get the complete scan.csv file but the Pi won’t produce an image for me.
So I tried to do that part on my PC to have the scan image drawn, but the links in the aforementioned topic have been “outdated” and newer versions are available. But I cannot figure out how install the newer Pillow version.
More then happy to share the scan.cvs file to someone who knows how to get it plotted into an image.
(Sorry, not really a software guy, more a hardware guy)

Something I did not mention in the start message, the LNA I am running is the following:
New Product: RTL-SDR Blog 1090 MHz ADS-B LNA.
So somewhere my guess would be that I do not have that much interference from other sources, but it is good to check anyway if the filter is not damaged.

Those partial messages, is that due to the fact that the SDR dongle gets overloaded with the amount of messages and not being able to process them all? Or does that have a different source?
Fyi, only 1.5% of the messages I receive are above -3dBFS

I wasn’t clear, it’s messages that can’t be decoded for whatever reason. As soon as you can’t decode it, it counts as noise.
That can be due to interference or just overlapping messages. After all on 1090 MHz aircraft just send, they don’t check for the frequency to be “free” as the messages are entirely too short to do that.

Honestly it would probably make sense to record noise in approximately message size chunks and then do some statistics to exclude the top 20 % to get a somewhat “true” noise reading.

No, indeed, they just blast it out there on the 1090mhz band and it is up to the receiver to make sense out of it.
I would not be surprised if there are quite some overlapping messages in my surroundings… :sweat_smile:

What is the error message?

This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.