On the problem presented on the FACT Station:
There was a sequence of observations on Dev Network, to see if there is an exogenous factor that produces observations with intense noise. The observations are as continuous as possible and even overlapping (in fact some of them are back to back) so as to determine whether an external source was operating and then shut off.
In the attached PDF you can see all test observations with results.
Green ones the ones without noise (regardless they caught a satellite or not) and with white ones that did not have a waterfall.
My personal point of view is that it is not an exogenous factor. Thoughts?
FACT-Observations.pdf (27.4 KB)
I think it is hard to tell from the time sequence alone. It could also be a correlation with the receiver frequency.
When I hear the terms “UK, interference, VHF” together, the first thing that comes to mind are pagers. Apparently they are still very popular and are right around the 2 meter band. A taxi station could be another possibility.
I would try to find the signal(s) using a wide band analyser and / or look in a frequency database if they have such thing publicly accessible.
Examining the test observations I deduce (Sherlok style ) the following:
- In no observation there is a change in noise floor (within the observation)
- Observations are of nominal length (more than 10 mins avg)
- On overlapping obs (back to back actually) the only change is that we stop and start, still noise floor changes considerably even on aprox the same frequencies
- If it was an external source causing the noise, we would see starts and stops (especially for pagers etc as they are short bursts)
Thus I am pretty confident that there is an issue with the AGC or gain control in general from the rtl-sdr. Pretty sure this specific one is faulty. We should replace it and check.
Now that you mention these things, I have seen similar noise due to bad USB connectors.
This is exactly the behavior I have faced with my station. In my case it was with NooElec rtl-sdr and it looks like randomly chooses to enable or disable AGC, although in gnuradio script is defined to disable AGC.
Trying the same device with gqrx on laptop, I never had any problem like this. So, I’m not 100% sure that this is a hardware issue. I think there is a possibility to be a software one, maybe a bug in GNU Radio when running in rpi.