Optimisng Signal-to-noise Ratio?

Hi all,

I’m pretty happy my station has all the range it’s ever going to get, based on geography, as shown by the HeyWhatsThat range rings.

I’m a bit rusty on SNR issues - how is my current SNR looking after the last 24 hours or so…?


Remind me - what optimisation options and tweaks are available?


At the moment, im using fA’s dump1090 (as part of the standard package install).
I used Mutability in the past, but that seems to have dissapeared - later instructions I found suggested the only way to install that these days is to build it yourself - is that still a thing?

Judging the amount of “> -3 dBFS/Hr”, I think changing the gain might help here.
“dump1090-fa”'s default setting is “-10”, right?

Thanks Hbokh, coming back to me now! :slight_smile:

Ran the gain testing scripts from here:

post197318.html#p197318

Chnaged from the default -10 to 49.6, which showed as “best” on my rig.

Just testing for PPM adjustment now…

Still running dump1090-fa from Romeo Golf’s branch of ADSBReceiver… I’d like to use Obj’s Mutability fork, but none of the install instructions (including build from the repo) seem to work.
Has Mut been discontinued now?

Bake a Pi
Go to OPTION-3
Follow Item 7) Install dump1090-mutability v1.15~dev

Done a test install 15 minutes ago. It is working

Thanks abcd567 - I’ll give that a try later …

OK - found the log file…

I set my gain to 49.6 following the gain test results.

I get:

461865 messages with signal power above -3dBFS
4117658 total usable messages

Which gives 11.21 - 11.21% !

If correct, this suggests I need to drop the gain, but then I ask why the gain tests suggested 49.6 was best?

Basically, the test is highly imperfect since traffic is continually changing. Ideally you’d want to test two otherwise identical setups at the same time and take a sample for a few minutes with one setup at a static gain vs the test rig at your test gain. Although it’s imperfect (as-is), it’s still a decent way to find a rough ballpark - especially if you sample multiple times and at different times of day/night. It’s really tough to compare continually changing data without a watermark, but not everyone has multiple rigs laying around to test so something is better than nothing for the most part. Just keep the shortcomings in the back of your mind when analyzing results.