today i was tracking a flight heading to my home airport. as the flight was descending and about 30nm away the aircraft image just stopped moving. for another two minutes or so i was still seeing “hits” that were approximately -3dBFS and every second or so, but no motion from the image, then the image disappeared. this would make sense if the amplitude was very low and the time between hits was long but his was not the case.
can someone advise if this is normal for the skyview algorithm.
Things to consider:
- Altitude of the plane and receiver location;
- Possible receiver overload. -3dBFS at 30NM is a lot.
- Decoding errors;
- Not all hits contain position information;
- Was the mapping ADS-B decoded or calculated MLAT?
dxista, thank you. regarding the -3dBFS, i seem to remember staff from FA posting an image showing a waveform but can’t find. -3dBFS is 70.7% of full scale, so is the full scale rating a peak value or RMS (or other)?
also, i did not know that not all hits contain position info…thanks for the education
I’m no expert on the topic, but my understanding is that -3dBFS is the maximum signal before clipping occurs. The ‘room’ between -3 and 0 dBFS account for losses. No digital signal can be on the plus side of dBFS scale.
They may not contain position info at all, if only Mode S. The mapping may have been calculated (MLATed). Once/if you receiver loses sync with other sites, MLAT feedback from Flightaware may stop.
well, from an analog perspective (i’m old) peak level is 3dB above RMS level, so if -3dBFS is a “limit” then a peak would be right at 0dBFS and would be on the edge of clipping. i have no idea how this translates into the digital world. i am running a piaware with SD card and no modifications to gain so from what i’ve read my gain is at -10. i rarely see signal levels lower than -12dBFS to -14dBFS when flightaware “times out” and the image goes grey. i have seen folks discussing much lower signal levels but have never seen them.
i can imagine the FA staff thinking about how much work they could get done if they weren’t responding to our questions. if someone on the discussion board or FA staff could recommend a paper or tutorial for us to read it might help reduce the volume of questions (of course this may have been done and i’m just not aware…maybe a section in the website devoted to ADS-B education…anyway)
Try setting the gain lower to see how the receiver responds. I suggest 48.0 to start.
Not knowing your setup makes it hard to suggest other things, but lower values than you see currently are very possible. My guess is that you receiver is overloading and/or your noise floor is very high.
There is a lot of archived knowledge on this forum, but you need to search past posts.
FA antenna on roof, RTL-SDR LNA mounted right at antenna on roof. 75 ft of LMR400 coax. then lightning arrester mounted on DX engineering ground plate to new ground rod (bonded to main ground, 6 feet from new ground rod). then 20 ft of LMR400 into house. then external Bias-T (mini circuits) for LNA (i’m not much of a Linux driver so used the external Bias-T) and then into RTL-SDR dongle. (various adapters in signal chain introducing loss).
i have read various posts re optimizing gain via software but am more comfortable doing via non software solution so will try that approach.
many thanks for your input
My setup is the same. The only difference is that my FA antenna is indoors, so no coax losses to deal with. Same dongle and LNA/filter combo, including external bias T injector.
I arrived at 48.0 after a lot of tests. In my case, good performance with small variations, ranged from gain levels from 33.9 to 48.0. I’m confident that -10 is likely too high a setting for your configuration.
It does not hurt to try.
this is fun, thank you
I liked better the manual gain changes approach, as opposed to using scripts. See here:
The -3dB value comes from a normal Mode S / ADS-B message having a 50% duty cycle, and the signal level measuring power not voltage.
thanks for the info, so the ratio of average power to peak power is 0.5 (ie 50% or -3dB from full scale) and the concern is that the peaks are 3dB above the average, so if the full scale power level is -3dB then the peak power level is right at clipping (0dBFS) and anything above -3dBFS would have some of the peaks clipped? is this correct?
Yep. It’s just a very rough indicator that you may be over-amplifying the input.
anything above -3dBFS would have some of the peaks clipped?
ADSB messages are amplitude modulated so even clipped messages can be decoded. It is possible to have decoded messages at high as -1dB RSSI power level. ADSB Decoders work by comparing the high (on) level to the low (off) level to get the 1 or 0. There is no clock signal transmitted since it is inferred from the on/off pattern. Here is a link if you are interested in knowing more.
https://mode-s.org/decode/index.html
We have a ADSB FAQ page and a monthly newsletter that go over some topics. If you want a specific topic added you can either email or direct message me on this forums. We do take requests.
One thing I should mention is that a plane transponder send a few Watts of power output (small planes) to 500W of power output (large jet planes). This means that signal level vs distance is not true relationship. Jets hundreds of miles away will have a higher detected signal power than a small GA plane very close.
3 Likes
this is great, thank you very much
hello dxista, after your comments and FA staff comments (i was very please to see david baker’s comments re -1dB RSSI) i purchased two attenuators (a 1dB and 2dB) and just received. i placed the 1dB attenuator right before the dongle and am monitoring. it’s hard to tell (too bad air traffic isn’t constant) but the results seem promising. i am not a LInux driver so the 1dB attenuator (hardware) seemed much easier to me. i will also play with the 2dB unit after some time monitoring the 1dB.