FlightAware Discussions

Back again

I’ve reduced graphs1090 to write 10 MB once a night.
So it’s not really hard on sd-cards anymore at all.

I’d say it depends how many feed clients you’re already running.
But yeah an independent device is the safer bet.
On the other hand i don’t think reflashing the Radarcape with a fresh image is that complicated either is it?

Ok, read descriptions in that post. Airspy currently at gain 21. I invoked the -e at 8, it raised the weakest from -11.8 to about -7.0. Where is the sweet spot? Is the gain too high. Itis pretty quiet signal wise at the moment.

I already gave the guideline to place the cyan line around -30 dBFS (graphs1090 signal graph).
Just change the airspy gain to make that happen.

If you are going to be rude, don’t bother replying.

No, it is not that complicated. A little bit more on the Radarcape, but not for the Airsquitter. This device is connected as mass storage and can be flashed directly (They use a 16GB sd card).
For the radarcape you have to open and remove the card for flashing, but the rest is the same.
The devices are built for easy usage without having too much IT skills.

Jetvision offers recovery images for both devices. A restore is done in approx 15-20 Minutes including configuration of the services.

I haven’t tried using a different (non Jetvision) image, not sure if this would even work.

Yeah, my statement was more on the Radarcape itself. Jetvision optimized that for SD card usage as a long-term device. Even many of the logs are in RAM, so after a reboot they are gone.

However i have no clue if the installation of graphs1090 would even work beside the fact that modification for rcd is needed.

Lower the gain to about 16 and see where your weakest signals end up. If they’re still above -30, lower the gain again. Play with the gain setting until you see the desired results.

Thanks for your reply. The figure of -30 has been stated a number of times. In regards to ADS -B signal level, is it better to have the greater negative reading or closer to zero? Currently with both PI 4’s at gain 21, the weakest signal level, on the weakest antenna gain, is at -17 signal level ,on a 2 hour reading. The other one I shall have to wait, there was a power supply issue feeding the pre amp on the other antenna, which is the stronger gain. Now I seem to remember Joe and his ADSB project when talking about these graphs, produced a great document about where you needed to tweak to get the best results and at what levels in regards to a aspects that affect your detections. Any one still got access to that gem???

1 Like

Yay !!! Found that link to Joe Prochazka github repository (still available!!) Performance Graphs · jprochazka/adsb-receiver Wiki · GitHub . Are the items discussed still the same for the current versions we run? Anyone who want a bit more info about performance graphs, look at this, from the creator (I think he developed it, but apologies if that not correct, he put a lot of work in supporting people setting up their own systems. :slight_smile:

Interesting in comparing the 2 different signal level states, Joe’s and the current one. With Joe’s the 3 variables are mean level, peak level and noise. Currently we have 1st to 3rd quartile (not sure what data the quartile is drawing in statistics), mean medium level, weakest and peak level. Not sure what happened to the noise level, that showed how much problem you are getting from the garbage about in the signal world, like having a 4G mobile tower about 300 metres from your rig… :wink:

You know about that from user @wiedehopf

GitHub - wiedehopf/graphs1090: Graphs for dump1090 (based on dump1090-tools by mutability)

This is used by most of us helps talking about the same values. Not sure if these are really comparable.

From a cursory look the statements there are simplified and some are just incorrect.
mutability created those graphs.

Weakest antenna gain? Can you explain what you are referring to?
You have 2 antennas, one with less gain?

Messages can be decoded down to -40 dBFS or even lower.
But to make sure you’re not losing anything, it’s a good start to put the weakest signals near -30.
That way you usually don’t have very strong signals that the receiver fails to pick up due to temporary overload (just for that message).

The noise number is problematic in dump1090 and the airspy decoder doesn’t report it.
The weakest signal is a more reliable indicator and available with the airspy decoder.

Also right click copy image, then Ctrl-V here helps a lot to let others help you.
Much better than trying to describe images.

As you see i like to put the weakest signal even lower.

It is a pity the noise value is not there any more. A handy thing to let you know if you have some issues that might require a filter. Well I re imaged my new SD cards after they arrived today. The ones that will stay with the PI4’s. All set up and happy with settings and set up.

From refreshing myself with Joe’s great insights with his graphs, my settings, as is, just from installing the Airspy script and no tweaks, gain at 21 of both PI 4’s, is pretty good. Working off the basis of:

Mean Level - The mean level of signal strength. The higher the number, the better (i.e. -20 is better than -23.2 in this example).

Peak Level - The best signal strength you were able to attain during that defined period of time graphed. The closer to 0 that number is, the better.

A little while ago, I decided, in best interests, to drop the height of my antenna mast. Although well guyed and no previous scary moments in storms, I was willing to maybe possible loose some detection range, for the sake of not worrying about what was happening on the roof, during a storm.

First the both rigs are showing a Peak Level of 0.0 which I find a bit dubious. With the current version, over 8 hours, on my A rig, strongest gain antenna and I am taking the Mean Median Level to be a average of you mean level, -1.3. My B rig, on the weaker gain antenna, the Mean Median Level at -5.0. According to Joe’s instructions, pretty good. But the litmus test, watching the tracking live. I live about 450 kilometres north west of Sydney Australia. I have great North to South West coverage. Tracking flights from Asia and they approach Sydney, over Queensland, all the way to they start their finals, dropping into the Sydney basin. Same for traffic heading North East from Southern Australia towards Brisbane. We are on the Western slopes of Australia’s Great Dividing Range, similar to the US’s Rockies, but not as high. My easterly coverage is a down due the sandstone wall, but I can track up the eastern coast traffic heading for Brisbane from Sydney, when they hit their cruising heights. Here is the clincher with the current setup. I can track US traffic as they approach Sydney from the NE, also traffic coming from New Zealand, heading for destinations west of Sydney, at their cruising altitude. I have never tracked flights from Melbourne heading for the US as they cruise past Sydney. With my strongest antenna and my current settings, I am now tracking these flights. Such is the benefit of the development of the PI’s and the setups supplied from this site. Happy camper. :slight_smile:

Just checked my graphs (FA Stick) and the noise is there (green line). Not sure if this is what you are missing.


Maybe it’s not in because of the Airspy, but that’s something @wiedehopf need to answer.

Correct and also a long lead to the PI 4 that is in the office, the other one is up on the mast. Noted re loosing very strong signals due to overload. My airports main approach direction which is basically east to west, due to our predominate west to north west wind patterns and the fact that the main runway is basically East - West. This approach path is probably about 700 metres south of my place and sometimes I will note a landing aircraft’s track will temporarily drop off as the plane passes by, due to your above reasons, but it is generally only for about 15 seconds. We used to have based at our airport, the Royal Australian Air Force’s basic flight school. They also trained a number of Asian nations pilots as well. Our airport was one of the busiest regional airports in Australia. Unfortunately BAE who also assisted with the flight training, found a better deal else where, so our local traffic dramatically dropped right off. But we still have visits from the Airforce, with RAAF Williamtown (Newcastle) and Richmond (Sydney) only a little journey away from us flight wise, and a broad section of their planes from F35’s to C17’s and everything in between. So basically I suppose my problems from a lot of strong signals due to heavy traffic is not an issue for me, with most of it at 30,000 ft plus above. Happy with what I am seeing so far, out of the box as is, without too much need for tweaking. :slight_smile:

Sorry, I can’t be of more help. I’m new to Airspy myself. My advice to lower gain was about the best I can offer. @wiedehopf gave a more helpful answer. Best of luck.

It is there for rtl-sdrs, it’s not possible with the airspy decoder.
You won’t get it using the other graphs … actually they won’t work properly with the airspy decoder at all feeding into dump1090.

This advice is very very outdated and was probably written when people weren’t using any amplifiers while now most SDRs being used for 1090 have an extra LNA built in.
It really wasn’t the best advice back then but it probably was ok advice for most cases.
It’s advice in regards to building a better antenna, if you keep the gain constant, a better antenna / lower loss coax will result in stronger signals.
With LNAs you don’t want the SDR gain at maximum to distribute the signals in a power band that doesn’t temporary overload the SDR for the messages from the aircraft flying low and close by.

Go to a gain of 16 on the first and gain of 17 on the second rig.
Check if your range reduce at all, i’d be willing to be it will not. (of course subject to changes in traffic and conditions)