Testing the Airspy with the New And Improved Version of ADSBSpy
Airspy have recently released an update to their ADSBspy decoder, which is an Airspy One/R2 compatible decoder for 1090 MHZ ADS-B signals. According to 'prog', the software developer of ADSBSpy, his setup can see almost double the number of aircraft and with fewer false positives when using the updated software. Prog writes that the secret to the improvement is some reworked DSP code that aims to exploit oversampling in the Airspy to the maximum.
Good detection performance with the new #ADSB decoder for #Airspy. Almost double the number of aircraft with fewer false positives. @flightradar24 pic.twitter.com/keLzi9jOzq— prog' (@lambdaprog) April 16, 2018
We compared the new (126.96.36.199/39) decoder against the old decoder (188.8.131.52) which used to get similar performance to dump1090. The test setup was two Airspy dongles connected to a dipole antenna via a splitter, with our Triple Filtered ADS-B LNA used by the antenna. One Airspy was used to power the LNA via it's bias tee, and both units received the same amplified signal. We found indeed that the new version of ADSBSpy receives a good number more aircraft in our set up, and an increased number of ADS-B messages too.
It seems that most of the additionally received aircraft must be from extremely weak signals, because when looking in Virtual Radar Server the extra aircraft usually only show their ICAO and maybe altitude and speed until they get closer.
So far this software appears to provide the best performance on ADS-B that we've seen so far, so if you are using an Airspy for ADS-B tracking we'd like to hear results from anyone who upgrades.
Yes but not support for hackrf one, is so sad
If i read correctly, one of the two Airspy R2 is used to power the150 ma for the Triple Filtered ADS-B LNA. 150 ma is not to much for the Airspy R2 .
Do you know the maximum bias-tee current for the Airspy R2.
Thanks in advance.
My biggest issue is the 20MHz multi-lateralization (MLAT) timing precision. All the programs seem to be for 12MHz timestamps. Not sure why they are doing this, or why they do not have output options. I do MLAT with ADSB-Exchange and ADSB Spy is still useless to me.
Use the beast mode.
Thx Vince. I loaded it up and am seeing a 67% rejection rate by VRS as well. I was hoping it would work but it appears to still have issues. When I manually start it up it returns airspy_adsb v1.19 with this version.
Since backing it down to FEC 1 and increasing the VRS scrutiny to 2 messages before accepted, I have had no issues whatsoever and it looks like there’s an improvement. Can’t say for certainty since it’s the weekend but for this Sunday so far so good.
Has the Raspberry Pi version been upgraded as well. Can’t tell from the Airspy site as it does not list a version #
prog said he upgraded all builds. There is a difference. Now I’m getting a lot of bad decodes that would correspond to the “increase” in planes seen. They pass parity but the planes are not even in the air, most of them aren’t even registered. I increased the scrutiny in VRS and the bad decodes still get through. This is with FEC 2 though, I don’t think it’s reliable. Turned back to FEC 1.